diff --git a/.github/workflows/create-tag.yml b/.github/workflows/create-tag.yml index 7aef7335..ae8252d7 100644 --- a/.github/workflows/create-tag.yml +++ b/.github/workflows/create-tag.yml @@ -11,6 +11,7 @@ on: - acp - approval-gate - auth-credentials + - context-compaction - harness - hook-fanout - iii-database diff --git a/.github/workflows/release.yml b/.github/workflows/release.yml index a212ac70..83cb8f1f 100644 --- a/.github/workflows/release.yml +++ b/.github/workflows/release.yml @@ -6,6 +6,7 @@ on: - 'acp/v*' - 'approval-gate/v*' - 'auth-credentials/v*' + - 'context-compaction/v*' # harness/v* is handled by release-harness-bundle.yml (single-dispatch fan-out). - 'hook-fanout/v*' - 'iii-database/v*' diff --git a/context-compaction/Cargo.lock b/context-compaction/Cargo.lock new file mode 100644 index 00000000..f7f7e7a9 --- /dev/null +++ b/context-compaction/Cargo.lock @@ -0,0 +1,2689 @@ +# This file is automatically @generated by Cargo. +# It is not intended for manual editing. +version = 4 + +[[package]] +name = "aho-corasick" +version = "1.1.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ddd31a130427c27518df266943a5308ed92d4b226cc639f5a8f1002816174301" +dependencies = [ + "memchr", +] + +[[package]] +name = "android_system_properties" +version = "0.1.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "819e7219dbd41043ac279b19830f2efc897156490d7fd6ea916720117ee66311" +dependencies = [ + "libc", +] + +[[package]] +name = "anstream" +version = "1.0.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "824a212faf96e9acacdbd09febd34438f8f711fb84e09a8916013cd7815ca28d" +dependencies = [ + "anstyle", + "anstyle-parse", + "anstyle-query", + "anstyle-wincon", + "colorchoice", + "is_terminal_polyfill", + "utf8parse", +] + +[[package]] +name = "anstyle" +version = "1.0.14" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "940b3a0ca603d1eade50a4846a2afffd5ef57a9feac2c0e2ec2e14f9ead76000" + +[[package]] +name = "anstyle-parse" +version = "1.0.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "52ce7f38b242319f7cabaa6813055467063ecdc9d355bbb4ce0c68908cd8130e" +dependencies = [ + "utf8parse", +] + +[[package]] +name = "anstyle-query" +version = "1.1.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "40c48f72fd53cd289104fc64099abca73db4166ad86ea0b4341abe65af83dadc" +dependencies = [ + "windows-sys 0.61.2", +] + +[[package]] +name = "anstyle-wincon" +version = "3.0.11" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "291e6a250ff86cd4a820112fb8898808a366d8f9f58ce16d1f538353ad55747d" +dependencies = [ + "anstyle", + "once_cell_polyfill", + "windows-sys 0.61.2", +] + +[[package]] +name = "anyhow" +version = "1.0.102" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "7f202df86484c868dbad7eaa557ef785d5c66295e41b460ef922eca0723b842c" + +[[package]] +name = "async-trait" +version = "0.1.89" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9035ad2d096bed7955a320ee7e2230574d28fd3c3a0f186cbea1ff3c7eed5dbb" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "atomic-waker" +version = "1.1.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "1505bd5d3d116872e7271a6d4e16d81d0c8570876c8de68093a09ac269d8aac0" + +[[package]] +name = "autocfg" +version = "1.5.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "c08606f8c3cbf4ce6ec8e28fb0014a2c086708fe954eaa885384a6165172e7e8" + +[[package]] +name = "base64" +version = "0.22.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "72b3254f16251a8381aa12e40e3c4d2f0199f8c6508fbecb9d91f575e0fbb8c6" + +[[package]] +name = "bitflags" +version = "2.11.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "c4512299f36f043ab09a583e57bceb5a5aab7a73db1805848e8fef3c9e8c78b3" + +[[package]] +name = "block-buffer" +version = "0.10.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "3078c7629b62d3f0439517fa394996acacc5cbc91c5a20d8c658e77abd503a71" +dependencies = [ + "generic-array", +] + +[[package]] +name = "bumpalo" +version = "3.20.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "5d20789868f4b01b2f2caec9f5c4e0213b41e3e5702a50157d699ae31ced2fcb" + +[[package]] +name = "bytes" +version = "1.11.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "1e748733b7cbc798e1434b6ac524f0c1ff2ab456fe201501e6497c8417a4fc33" + +[[package]] +name = "cc" +version = "1.2.62" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "a1dce859f0832a7d088c4f1119888ab94ef4b5d6795d1ce05afb7fe159d79f98" +dependencies = [ + "find-msvc-tools", + "shlex", +] + +[[package]] +name = "cfg-if" +version = "1.0.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9330f8b2ff13f34540b44e946ef35111825727b38d33286ef986142615121801" + +[[package]] +name = "cfg_aliases" +version = "0.2.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "613afe47fcd5fac7ccf1db93babcb082c5994d996f20b8b159f2ad1658eb5724" + +[[package]] +name = "chrono" +version = "0.4.44" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "c673075a2e0e5f4a1dde27ce9dee1ea4558c7ffe648f576438a20ca1d2acc4b0" +dependencies = [ + "iana-time-zone", + "num-traits", + "serde", + "windows-link", +] + +[[package]] +name = "clap" +version = "4.6.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "1ddb117e43bbf7dacf0a4190fef4d345b9bad68dfc649cb349e7d17d28428e51" +dependencies = [ + "clap_builder", + "clap_derive", +] + +[[package]] +name = "clap_builder" +version = "4.6.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "714a53001bf66416adb0e2ef5ac857140e7dc3a0c48fb28b2f10762fc4b5069f" +dependencies = [ + "anstream", + "anstyle", + "clap_lex", + "strsim", +] + +[[package]] +name = "clap_derive" +version = "4.6.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "f2ce8604710f6733aa641a2b3731eaa1e8b3d9973d5e3565da11800813f997a9" +dependencies = [ + "heck", + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "clap_lex" +version = "1.1.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "c8d4a3bb8b1e0c1050499d1815f5ab16d04f0959b233085fb31653fbfc9d98f9" + +[[package]] +name = "colorchoice" +version = "1.0.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "1d07550c9036bf2ae0c684c4297d503f838287c83c53686d05370d0e139ae570" + +[[package]] +name = "const-hex" +version = "1.19.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "20d9a563d167a9cce0f94153382b33cb6eded6dfabff03c69ad65a28ea1514e0" +dependencies = [ + "cfg-if", + "cpufeatures", + "proptest", + "serde_core", +] + +[[package]] +name = "context-compaction" +version = "0.1.0" +dependencies = [ + "anyhow", + "chrono", + "clap", + "iii-sdk", + "serde", + "serde_json", + "serde_yaml", + "tokio", + "tracing", + "tracing-subscriber", +] + +[[package]] +name = "core-foundation" +version = "0.10.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "b2a6cd9ae233e7f62ba4e9353e81a88df7fc8a5987b8d445b4d90c879bd156f6" +dependencies = [ + "core-foundation-sys", + "libc", +] + +[[package]] +name = "core-foundation-sys" +version = "0.8.7" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "773648b94d0e5d620f64f280777445740e61fe701025087ec8b57f45c791888b" + +[[package]] +name = "cpufeatures" +version = "0.2.17" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "59ed5838eebb26a2bb2e58f6d5b5316989ae9d08bab10e0e6d103e656d1b0280" +dependencies = [ + "libc", +] + +[[package]] +name = "crypto-common" +version = "0.1.7" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "78c8292055d1c1df0cce5d180393dc8cce0abec0a7102adb6c7b1eef6016d60a" +dependencies = [ + "generic-array", + "typenum", +] + +[[package]] +name = "data-encoding" +version = "2.11.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "a4ae5f15dda3c708c0ade84bfee31ccab44a3da4f88015ed22f63732abe300c8" + +[[package]] +name = "digest" +version = "0.10.7" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9ed9a281f7bc9b7576e61468ba615a66a5c8cfdff42420a70aa82701a3b1e292" +dependencies = [ + "block-buffer", + "crypto-common", +] + +[[package]] +name = "displaydoc" +version = "0.2.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "97369cbbc041bc366949bc74d34658d6cda5621039731c6310521892a3a20ae0" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "dyn-clone" +version = "1.0.20" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d0881ea181b1df73ff77ffaaf9c7544ecc11e82fba9b5f27b262a3c73a332555" + +[[package]] +name = "either" +version = "1.15.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "48c757948c5ede0e46177b7add2e67155f70e33c07fea8284df6576da70b3719" + +[[package]] +name = "equivalent" +version = "1.0.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "877a4ace8713b0bcf2a4e7eec82529c029f1d0619886d18145fea96c3ffe5c0f" + +[[package]] +name = "errno" +version = "0.3.14" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "39cab71617ae0d63f51a36d69f866391735b51691dbda63cf6f96d042b63efeb" +dependencies = [ + "libc", + "windows-sys 0.61.2", +] + +[[package]] +name = "find-msvc-tools" +version = "0.1.9" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "5baebc0774151f905a1a2cc41989300b1e6fbb29aff0ceffa1064fdd3088d582" + +[[package]] +name = "fnv" +version = "1.0.7" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "3f9eec918d3f24069decb9af1554cad7c880e2da24a9afd88aca000531ab82c1" + +[[package]] +name = "foldhash" +version = "0.1.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d9c4f5dac5e15c24eb999c26181a6ca40b39fe946cbe4c263c7209467bc83af2" + +[[package]] +name = "form_urlencoded" +version = "1.2.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "cb4cb245038516f5f85277875cdaa4f7d2c9a0fa0468de06ed190163b1581fcf" +dependencies = [ + "percent-encoding", +] + +[[package]] +name = "futures-channel" +version = "0.3.32" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "07bbe89c50d7a535e539b8c17bc0b49bdb77747034daa8087407d655f3f7cc1d" +dependencies = [ + "futures-core", +] + +[[package]] +name = "futures-core" +version = "0.3.32" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "7e3450815272ef58cec6d564423f6e755e25379b217b0bc688e295ba24df6b1d" + +[[package]] +name = "futures-executor" +version = "0.3.32" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "baf29c38818342a3b26b5b923639e7b1f4a61fc5e76102d4b1981c6dc7a7579d" +dependencies = [ + "futures-core", + "futures-task", + "futures-util", +] + +[[package]] +name = "futures-macro" +version = "0.3.32" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e835b70203e41293343137df5c0664546da5745f82ec9b84d40be8336958447b" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "futures-sink" +version = "0.3.32" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "c39754e157331b013978ec91992bde1ac089843443c49cbc7f46150b0fad0893" + +[[package]] +name = "futures-task" +version = "0.3.32" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "037711b3d59c33004d3856fbdc83b99d4ff37a24768fa1be9ce3538a1cde4393" + +[[package]] +name = "futures-util" +version = "0.3.32" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "389ca41296e6190b48053de0321d02a77f32f8a5d2461dd38762c0593805c6d6" +dependencies = [ + "futures-core", + "futures-macro", + "futures-sink", + "futures-task", + "pin-project-lite", + "slab", +] + +[[package]] +name = "generic-array" +version = "0.14.7" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "85649ca51fd72272d7821adaf274ad91c288277713d9c18820d8499a7ff69e9a" +dependencies = [ + "typenum", + "version_check", +] + +[[package]] +name = "getrandom" +version = "0.2.17" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ff2abc00be7fca6ebc474524697ae276ad847ad0a6b3faa4bcb027e9a4614ad0" +dependencies = [ + "cfg-if", + "js-sys", + "libc", + "wasi", + "wasm-bindgen", +] + +[[package]] +name = "getrandom" +version = "0.3.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "899def5c37c4fd7b2664648c28120ecec138e4d395b459e5ca34f9cce2dd77fd" +dependencies = [ + "cfg-if", + "js-sys", + "libc", + "r-efi 5.3.0", + "wasip2", + "wasm-bindgen", +] + +[[package]] +name = "getrandom" +version = "0.4.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0de51e6874e94e7bf76d726fc5d13ba782deca734ff60d5bb2fb2607c7406555" +dependencies = [ + "cfg-if", + "libc", + "r-efi 6.0.0", + "wasip2", + "wasip3", +] + +[[package]] +name = "h2" +version = "0.4.14" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "171fefbc92fe4a4de27e0698d6a5b392d6a0e333506bc49133760b3bcf948733" +dependencies = [ + "atomic-waker", + "bytes", + "fnv", + "futures-core", + "futures-sink", + "http", + "indexmap", + "slab", + "tokio", + "tokio-util", + "tracing", +] + +[[package]] +name = "hashbrown" +version = "0.15.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9229cfe53dfd69f0609a49f65461bd93001ea1ef889cd5529dd176593f5338a1" +dependencies = [ + "foldhash", +] + +[[package]] +name = "hashbrown" +version = "0.17.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ed5909b6e89a2db4456e54cd5f673791d7eca6732202bbf2a9cc504fe2f9b84a" + +[[package]] +name = "heck" +version = "0.5.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "2304e00983f87ffb38b55b444b5e3b60a884b5d30c0fca7d82fe33449bbe55ea" + +[[package]] +name = "hostname" +version = "0.4.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "617aaa3557aef3810a6369d0a99fac8a080891b68bd9f9812a1eeda0c0730cbd" +dependencies = [ + "cfg-if", + "libc", + "windows-link", +] + +[[package]] +name = "http" +version = "1.4.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e3ba2a386d7f85a81f119ad7498ebe444d2e22c2af0b86b069416ace48b3311a" +dependencies = [ + "bytes", + "itoa", +] + +[[package]] +name = "http-body" +version = "1.0.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "1efedce1fb8e6913f23e0c92de8e62cd5b772a67e7b3946df930a62566c93184" +dependencies = [ + "bytes", + "http", +] + +[[package]] +name = "http-body-util" +version = "0.1.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "b021d93e26becf5dc7e1b75b1bed1fd93124b374ceb73f43d4d4eafec896a64a" +dependencies = [ + "bytes", + "futures-core", + "http", + "http-body", + "pin-project-lite", +] + +[[package]] +name = "httparse" +version = "1.10.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "6dbf3de79e51f3d586ab4cb9d5c3e2c14aa28ed23d180cf89b4df0454a69cc87" + +[[package]] +name = "hyper" +version = "1.9.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "6299f016b246a94207e63da54dbe807655bf9e00044f73ded42c3ac5305fbcca" +dependencies = [ + "atomic-waker", + "bytes", + "futures-channel", + "futures-core", + "h2", + "http", + "http-body", + "httparse", + "itoa", + "pin-project-lite", + "smallvec", + "tokio", + "want", +] + +[[package]] +name = "hyper-rustls" +version = "0.27.9" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "33ca68d021ef39cf6463ab54c1d0f5daf03377b70561305bb89a8f83aab66e0f" +dependencies = [ + "http", + "hyper", + "hyper-util", + "rustls", + "tokio", + "tokio-rustls", + "tower-service", + "webpki-roots", +] + +[[package]] +name = "hyper-timeout" +version = "0.5.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "2b90d566bffbce6a75bd8b09a05aa8c2cb1fabb6cb348f8840c9e4c90a0d83b0" +dependencies = [ + "hyper", + "hyper-util", + "pin-project-lite", + "tokio", + "tower-service", +] + +[[package]] +name = "hyper-util" +version = "0.1.20" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "96547c2556ec9d12fb1578c4eaf448b04993e7fb79cbaad930a656880a6bdfa0" +dependencies = [ + "base64", + "bytes", + "futures-channel", + "futures-util", + "http", + "http-body", + "hyper", + "ipnet", + "libc", + "percent-encoding", + "pin-project-lite", + "socket2", + "tokio", + "tower-service", + "tracing", +] + +[[package]] +name = "iana-time-zone" +version = "0.1.65" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e31bc9ad994ba00e440a8aa5c9ef0ec67d5cb5e5cb0cc7f8b744a35b389cc470" +dependencies = [ + "android_system_properties", + "core-foundation-sys", + "iana-time-zone-haiku", + "js-sys", + "log", + "wasm-bindgen", + "windows-core", +] + +[[package]] +name = "iana-time-zone-haiku" +version = "0.1.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "f31827a206f56af32e590ba56d5d2d085f558508192593743f16b2306495269f" +dependencies = [ + "cc", +] + +[[package]] +name = "icu_collections" +version = "2.2.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "2984d1cd16c883d7935b9e07e44071dca8d917fd52ecc02c04d5fa0b5a3f191c" +dependencies = [ + "displaydoc", + "potential_utf", + "utf8_iter", + "yoke", + "zerofrom", + "zerovec", +] + +[[package]] +name = "icu_locale_core" +version = "2.2.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "92219b62b3e2b4d88ac5119f8904c10f8f61bf7e95b640d25ba3075e6cac2c29" +dependencies = [ + "displaydoc", + "litemap", + "tinystr", + "writeable", + "zerovec", +] + +[[package]] +name = "icu_normalizer" +version = "2.2.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "c56e5ee99d6e3d33bd91c5d85458b6005a22140021cc324cea84dd0e72cff3b4" +dependencies = [ + "icu_collections", + "icu_normalizer_data", + "icu_properties", + "icu_provider", + "smallvec", + "zerovec", +] + +[[package]] +name = "icu_normalizer_data" +version = "2.2.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "da3be0ae77ea334f4da67c12f149704f19f81d1adf7c51cf482943e84a2bad38" + +[[package]] +name = "icu_properties" +version = "2.2.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "bee3b67d0ea5c2cca5003417989af8996f8604e34fb9ddf96208a033901e70de" +dependencies = [ + "icu_collections", + "icu_locale_core", + "icu_properties_data", + "icu_provider", + "zerotrie", + "zerovec", +] + +[[package]] +name = "icu_properties_data" +version = "2.2.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "8e2bbb201e0c04f7b4b3e14382af113e17ba4f63e2c9d2ee626b720cbce54a14" + +[[package]] +name = "icu_provider" +version = "2.2.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "139c4cf31c8b5f33d7e199446eff9c1e02decfc2f0eec2c8d71f65befa45b421" +dependencies = [ + "displaydoc", + "icu_locale_core", + "writeable", + "yoke", + "zerofrom", + "zerotrie", + "zerovec", +] + +[[package]] +name = "id-arena" +version = "2.3.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "3d3067d79b975e8844ca9eb072e16b31c3c1c36928edf9c6789548c524d0d954" + +[[package]] +name = "idna" +version = "1.1.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "3b0875f23caa03898994f6ddc501886a45c7d3d62d04d2d90788d47be1b1e4de" +dependencies = [ + "idna_adapter", + "smallvec", + "utf8_iter", +] + +[[package]] +name = "idna_adapter" +version = "1.2.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "cb68373c0d6620ef8105e855e7745e18b0d00d3bdb07fb532e434244cdb9a714" +dependencies = [ + "icu_normalizer", + "icu_properties", +] + +[[package]] +name = "iii-sdk" +version = "0.11.7-next.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "abad1f632473931d97733a4b3d73a04f2cfa1759fa4d4bcdab9305b634563bf7" +dependencies = [ + "async-trait", + "futures-util", + "hostname", + "opentelemetry", + "opentelemetry-http", + "opentelemetry-proto", + "opentelemetry_sdk", + "prost", + "reqwest", + "schemars", + "serde", + "serde_json", + "sysinfo", + "thiserror", + "tokio", + "tokio-tungstenite", + "tracing", + "uuid", +] + +[[package]] +name = "indexmap" +version = "2.14.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d466e9454f08e4a911e14806c24e16fba1b4c121d1ea474396f396069cf949d9" +dependencies = [ + "equivalent", + "hashbrown 0.17.1", + "serde", + "serde_core", +] + +[[package]] +name = "ipnet" +version = "2.12.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d98f6fed1fde3f8c21bc40a1abb88dd75e67924f9cffc3ef95607bad8017f8e2" + +[[package]] +name = "is_terminal_polyfill" +version = "1.70.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "a6cb138bb79a146c1bd460005623e142ef0181e3d0219cb493e02f7d08a35695" + +[[package]] +name = "itertools" +version = "0.14.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "2b192c782037fadd9cfa75548310488aabdbf3d2da73885b31bd0abd03351285" +dependencies = [ + "either", +] + +[[package]] +name = "itoa" +version = "1.0.18" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "8f42a60cbdf9a97f5d2305f08a87dc4e09308d1276d28c869c684d7777685682" + +[[package]] +name = "js-sys" +version = "0.3.98" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "67df7112613f8bfd9150013a0314e196f4800d3201ae742489d999db2f979f08" +dependencies = [ + "cfg-if", + "futures-util", + "once_cell", + "wasm-bindgen", +] + +[[package]] +name = "lazy_static" +version = "1.5.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "bbd2bcb4c963f2ddae06a2efc7e9f3591312473c50c6685e1f298068316e66fe" + +[[package]] +name = "leb128fmt" +version = "0.1.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "09edd9e8b54e49e587e4f6295a7d29c3ea94d469cb40ab8ca70b288248a81db2" + +[[package]] +name = "libc" +version = "0.2.186" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "68ab91017fe16c622486840e4c83c9a37afeff978bd239b5293d61ece587de66" + +[[package]] +name = "litemap" +version = "0.8.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "92daf443525c4cce67b150400bc2316076100ce0b3686209eb8cf3c31612e6f0" + +[[package]] +name = "log" +version = "0.4.29" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "5e5032e24019045c762d3c0f28f5b6b8bbf38563a65908389bf7978758920897" + +[[package]] +name = "lru-slab" +version = "0.1.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "112b39cec0b298b6c1999fee3e31427f74f676e4cb9879ed1a121b43661a4154" + +[[package]] +name = "matchers" +version = "0.2.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d1525a2a28c7f4fa0fc98bb91ae755d1e2d1505079e05539e35bc876b5d65ae9" +dependencies = [ + "regex-automata", +] + +[[package]] +name = "memchr" +version = "2.8.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "f8ca58f447f06ed17d5fc4043ce1b10dd205e060fb3ce5b979b8ed8e59ff3f79" + +[[package]] +name = "mio" +version = "1.2.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "50b7e5b27aa02a74bac8c3f23f448f8d87ff11f92d3aac1a6ed369ee08cc56c1" +dependencies = [ + "libc", + "wasi", + "windows-sys 0.61.2", +] + +[[package]] +name = "ntapi" +version = "0.4.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "c3b335231dfd352ffb0f8017f3b6027a4917f7df785ea2143d8af2adc66980ae" +dependencies = [ + "winapi", +] + +[[package]] +name = "nu-ansi-term" +version = "0.50.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "7957b9740744892f114936ab4a57b3f487491bbeafaf8083688b16841a4240e5" +dependencies = [ + "windows-sys 0.61.2", +] + +[[package]] +name = "num-traits" +version = "0.2.19" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "071dfc062690e90b734c0b2273ce72ad0ffa95f0c74596bc250dcfd960262841" +dependencies = [ + "autocfg", +] + +[[package]] +name = "objc2-core-foundation" +version = "0.3.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "2a180dd8642fa45cdb7dd721cd4c11b1cadd4929ce112ebd8b9f5803cc79d536" +dependencies = [ + "bitflags", +] + +[[package]] +name = "objc2-io-kit" +version = "0.3.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "33fafba39597d6dc1fb709123dfa8289d39406734be322956a69f0931c73bb15" +dependencies = [ + "libc", + "objc2-core-foundation", +] + +[[package]] +name = "once_cell" +version = "1.21.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9f7c3e4beb33f85d45ae3e3a1792185706c8e16d043238c593331cc7cd313b50" + +[[package]] +name = "once_cell_polyfill" +version = "1.70.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "384b8ab6d37215f3c5301a95a4accb5d64aa607f1fcb26a11b5303878451b4fe" + +[[package]] +name = "openssl-probe" +version = "0.2.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "7c87def4c32ab89d880effc9e097653c8da5d6ef28e6b539d313baaacfbafcbe" + +[[package]] +name = "opentelemetry" +version = "0.31.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "b84bcd6ae87133e903af7ef497404dda70c60d0ea14895fc8a5e6722754fc2a0" +dependencies = [ + "futures-core", + "futures-sink", + "js-sys", + "pin-project-lite", + "thiserror", + "tracing", +] + +[[package]] +name = "opentelemetry-http" +version = "0.31.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d7a6d09a73194e6b66df7c8f1b680f156d916a1a942abf2de06823dd02b7855d" +dependencies = [ + "async-trait", + "bytes", + "http", + "opentelemetry", + "reqwest", +] + +[[package]] +name = "opentelemetry-proto" +version = "0.31.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "a7175df06de5eaee9909d4805a3d07e28bb752c34cab57fa9cff549da596b30f" +dependencies = [ + "base64", + "const-hex", + "opentelemetry", + "opentelemetry_sdk", + "prost", + "serde", + "serde_json", + "tonic", + "tonic-prost", +] + +[[package]] +name = "opentelemetry_sdk" +version = "0.31.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e14ae4f5991976fd48df6d843de219ca6d31b01daaab2dad5af2badeded372bd" +dependencies = [ + "futures-channel", + "futures-executor", + "futures-util", + "opentelemetry", + "percent-encoding", + "rand", + "thiserror", + "tokio", + "tokio-stream", +] + +[[package]] +name = "percent-encoding" +version = "2.3.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9b4f627cb1b25917193a259e49bdad08f671f8d9708acfd5fe0a8c1455d87220" + +[[package]] +name = "pin-project" +version = "1.1.13" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "2466b2336ed02bcdca6b294417127b90ec92038d1d5c4fbeac971a922e0e0924" +dependencies = [ + "pin-project-internal", +] + +[[package]] +name = "pin-project-internal" +version = "1.1.13" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "c96395f0a926bc13b1c17622aaddda1ecb55d49c8f1bf9777e4d877800a43f8b" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "pin-project-lite" +version = "0.2.17" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "a89322df9ebe1c1578d689c92318e070967d1042b512afbe49518723f4e6d5cd" + +[[package]] +name = "potential_utf" +version = "0.1.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0103b1cef7ec0cf76490e969665504990193874ea05c85ff9bab8b911d0a0564" +dependencies = [ + "zerovec", +] + +[[package]] +name = "ppv-lite86" +version = "0.2.21" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "85eae3c4ed2f50dcfe72643da4befc30deadb458a9b590d720cde2f2b1e97da9" +dependencies = [ + "zerocopy", +] + +[[package]] +name = "prettyplease" +version = "0.2.37" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "479ca8adacdd7ce8f1fb39ce9ecccbfe93a3f1344b3d0d97f20bc0196208f62b" +dependencies = [ + "proc-macro2", + "syn", +] + +[[package]] +name = "proc-macro2" +version = "1.0.106" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "8fd00f0bb2e90d81d1044c2b32617f68fcb9fa3bb7640c23e9c748e53fb30934" +dependencies = [ + "unicode-ident", +] + +[[package]] +name = "proptest" +version = "1.11.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "4b45fcc2344c680f5025fe57779faef368840d0bd1f42f216291f0dc4ace4744" +dependencies = [ + "bitflags", + "num-traits", + "rand", + "rand_chacha", + "rand_xorshift", + "regex-syntax", + "unarray", +] + +[[package]] +name = "prost" +version = "0.14.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d2ea70524a2f82d518bce41317d0fae74151505651af45faf1ffbd6fd33f0568" +dependencies = [ + "bytes", + "prost-derive", +] + +[[package]] +name = "prost-derive" +version = "0.14.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "27c6023962132f4b30eb4c172c91ce92d933da334c59c23cddee82358ddafb0b" +dependencies = [ + "anyhow", + "itertools", + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "quinn" +version = "0.11.9" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "b9e20a958963c291dc322d98411f541009df2ced7b5a4f2bd52337638cfccf20" +dependencies = [ + "bytes", + "cfg_aliases", + "pin-project-lite", + "quinn-proto", + "quinn-udp", + "rustc-hash", + "rustls", + "socket2", + "thiserror", + "tokio", + "tracing", + "web-time", +] + +[[package]] +name = "quinn-proto" +version = "0.11.14" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "434b42fec591c96ef50e21e886936e66d3cc3f737104fdb9b737c40ffb94c098" +dependencies = [ + "bytes", + "getrandom 0.3.4", + "lru-slab", + "rand", + "ring", + "rustc-hash", + "rustls", + "rustls-pki-types", + "slab", + "thiserror", + "tinyvec", + "tracing", + "web-time", +] + +[[package]] +name = "quinn-udp" +version = "0.5.14" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "addec6a0dcad8a8d96a771f815f0eaf55f9d1805756410b39f5fa81332574cbd" +dependencies = [ + "cfg_aliases", + "libc", + "once_cell", + "socket2", + "tracing", + "windows-sys 0.60.2", +] + +[[package]] +name = "quote" +version = "1.0.45" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "41f2619966050689382d2b44f664f4bc593e129785a36d6ee376ddf37259b924" +dependencies = [ + "proc-macro2", +] + +[[package]] +name = "r-efi" +version = "5.3.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "69cdb34c158ceb288df11e18b4bd39de994f6657d83847bdffdbd7f346754b0f" + +[[package]] +name = "r-efi" +version = "6.0.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "f8dcc9c7d52a811697d2151c701e0d08956f92b0e24136cf4cf27b57a6a0d9bf" + +[[package]] +name = "rand" +version = "0.9.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "44c5af06bb1b7d3216d91932aed5265164bf384dc89cd6ba05cf59a35f5f76ea" +dependencies = [ + "rand_chacha", + "rand_core", +] + +[[package]] +name = "rand_chacha" +version = "0.9.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d3022b5f1df60f26e1ffddd6c66e8aa15de382ae63b3a0c1bfc0e4d3e3f325cb" +dependencies = [ + "ppv-lite86", + "rand_core", +] + +[[package]] +name = "rand_core" +version = "0.9.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "76afc826de14238e6e8c374ddcc1fa19e374fd8dd986b0d2af0d02377261d83c" +dependencies = [ + "getrandom 0.3.4", +] + +[[package]] +name = "rand_xorshift" +version = "0.4.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "513962919efc330f829edb2535844d1b912b0fbe2ca165d613e4e8788bb05a5a" +dependencies = [ + "rand_core", +] + +[[package]] +name = "regex-automata" +version = "0.4.14" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "6e1dd4122fc1595e8162618945476892eefca7b88c52820e74af6262213cae8f" +dependencies = [ + "aho-corasick", + "memchr", + "regex-syntax", +] + +[[package]] +name = "regex-syntax" +version = "0.8.10" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "dc897dd8d9e8bd1ed8cdad82b5966c3e0ecae09fb1907d58efaa013543185d0a" + +[[package]] +name = "reqwest" +version = "0.12.28" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "eddd3ca559203180a307f12d114c268abf583f59b03cb906fd0b3ff8646c1147" +dependencies = [ + "base64", + "bytes", + "futures-core", + "http", + "http-body", + "http-body-util", + "hyper", + "hyper-rustls", + "hyper-util", + "js-sys", + "log", + "percent-encoding", + "pin-project-lite", + "quinn", + "rustls", + "rustls-pki-types", + "serde", + "serde_json", + "serde_urlencoded", + "sync_wrapper", + "tokio", + "tokio-rustls", + "tower", + "tower-http", + "tower-service", + "url", + "wasm-bindgen", + "wasm-bindgen-futures", + "web-sys", + "webpki-roots", +] + +[[package]] +name = "ring" +version = "0.17.14" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "a4689e6c2294d81e88dc6261c768b63bc4fcdb852be6d1352498b114f61383b7" +dependencies = [ + "cc", + "cfg-if", + "getrandom 0.2.17", + "libc", + "untrusted", + "windows-sys 0.52.0", +] + +[[package]] +name = "rustc-hash" +version = "2.1.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "94300abf3f1ae2e2b8ffb7b58043de3d399c73fa6f4b73826402a5c457614dbe" + +[[package]] +name = "rustls" +version = "0.23.40" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ef86cd5876211988985292b91c96a8f2d298df24e75989a43a3c73f2d4d8168b" +dependencies = [ + "once_cell", + "ring", + "rustls-pki-types", + "rustls-webpki", + "subtle", + "zeroize", +] + +[[package]] +name = "rustls-native-certs" +version = "0.8.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "612460d5f7bea540c490b2b6395d8e34a953e52b491accd6c86c8164c5932a63" +dependencies = [ + "openssl-probe", + "rustls-pki-types", + "schannel", + "security-framework", +] + +[[package]] +name = "rustls-pki-types" +version = "1.14.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "30a7197ae7eb376e574fe940d068c30fe0462554a3ddbe4eca7838e049c937a9" +dependencies = [ + "web-time", + "zeroize", +] + +[[package]] +name = "rustls-webpki" +version = "0.103.13" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "61c429a8649f110dddef65e2a5ad240f747e85f7758a6bccc7e5777bd33f756e" +dependencies = [ + "ring", + "rustls-pki-types", + "untrusted", +] + +[[package]] +name = "rustversion" +version = "1.0.22" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "b39cdef0fa800fc44525c84ccb54a029961a8215f9619753635a9c0d2538d46d" + +[[package]] +name = "ryu" +version = "1.0.23" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9774ba4a74de5f7b1c1451ed6cd5285a32eddb5cccb8cc655a4e50009e06477f" + +[[package]] +name = "schannel" +version = "0.1.29" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "91c1b7e4904c873ef0710c1f407dde2e6287de2bebc1bbbf7d430bb7cbffd939" +dependencies = [ + "windows-sys 0.61.2", +] + +[[package]] +name = "schemars" +version = "0.8.22" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "3fbf2ae1b8bc8e02df939598064d22402220cd5bbcca1c76f7d6a310974d5615" +dependencies = [ + "dyn-clone", + "schemars_derive", + "serde", + "serde_json", +] + +[[package]] +name = "schemars_derive" +version = "0.8.22" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "32e265784ad618884abaea0600a9adf15393368d840e0222d101a072f3f7534d" +dependencies = [ + "proc-macro2", + "quote", + "serde_derive_internals", + "syn", +] + +[[package]] +name = "security-framework" +version = "3.7.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "b7f4bc775c73d9a02cde8bf7b2ec4c9d12743edf609006c7facc23998404cd1d" +dependencies = [ + "bitflags", + "core-foundation", + "core-foundation-sys", + "libc", + "security-framework-sys", +] + +[[package]] +name = "security-framework-sys" +version = "2.17.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "6ce2691df843ecc5d231c0b14ece2acc3efb62c0a398c7e1d875f3983ce020e3" +dependencies = [ + "core-foundation-sys", + "libc", +] + +[[package]] +name = "semver" +version = "1.0.28" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "8a7852d02fc848982e0c167ef163aaff9cd91dc640ba85e263cb1ce46fae51cd" + +[[package]] +name = "serde" +version = "1.0.228" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9a8e94ea7f378bd32cbbd37198a4a91436180c5bb472411e48b5ec2e2124ae9e" +dependencies = [ + "serde_core", + "serde_derive", +] + +[[package]] +name = "serde_core" +version = "1.0.228" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "41d385c7d4ca58e59fc732af25c3983b67ac852c1a25000afe1175de458b67ad" +dependencies = [ + "serde_derive", +] + +[[package]] +name = "serde_derive" +version = "1.0.228" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d540f220d3187173da220f885ab66608367b6574e925011a9353e4badda91d79" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "serde_derive_internals" +version = "0.29.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "18d26a20a969b9e3fdf2fc2d9f21eda6c40e2de84c9408bb5d3b05d499aae711" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "serde_json" +version = "1.0.149" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "83fc039473c5595ace860d8c4fafa220ff474b3fc6bfdb4293327f1a37e94d86" +dependencies = [ + "itoa", + "memchr", + "serde", + "serde_core", + "zmij", +] + +[[package]] +name = "serde_urlencoded" +version = "0.7.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d3491c14715ca2294c4d6a88f15e84739788c1d030eed8c110436aafdaa2f3fd" +dependencies = [ + "form_urlencoded", + "itoa", + "ryu", + "serde", +] + +[[package]] +name = "serde_yaml" +version = "0.9.34+deprecated" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "6a8b1a1a2ebf674015cc02edccce75287f1a0130d394307b36743c2f5d504b47" +dependencies = [ + "indexmap", + "itoa", + "ryu", + "serde", + "unsafe-libyaml", +] + +[[package]] +name = "sha1" +version = "0.10.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e3bf829a2d51ab4a5ddf1352d8470c140cadc8301b2ae1789db023f01cedd6ba" +dependencies = [ + "cfg-if", + "cpufeatures", + "digest", +] + +[[package]] +name = "sharded-slab" +version = "0.1.7" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "f40ca3c46823713e0d4209592e8d6e826aa57e928f09752619fc696c499637f6" +dependencies = [ + "lazy_static", +] + +[[package]] +name = "shlex" +version = "1.3.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0fda2ff0d084019ba4d7c6f371c95d8fd75ce3524c3cb8fb653a3023f6323e64" + +[[package]] +name = "signal-hook-registry" +version = "1.4.8" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "c4db69cba1110affc0e9f7bcd48bbf87b3f4fc7c61fc9155afd4c469eb3d6c1b" +dependencies = [ + "errno", + "libc", +] + +[[package]] +name = "slab" +version = "0.4.12" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0c790de23124f9ab44544d7ac05d60440adc586479ce501c1d6d7da3cd8c9cf5" + +[[package]] +name = "smallvec" +version = "1.15.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "67b1b7a3b5fe4f1376887184045fcf45c69e92af734b7aaddc05fb777b6fbd03" + +[[package]] +name = "socket2" +version = "0.6.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "3a766e1110788c36f4fa1c2b71b387a7815aa65f88ce0229841826633d93723e" +dependencies = [ + "libc", + "windows-sys 0.61.2", +] + +[[package]] +name = "stable_deref_trait" +version = "1.2.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "6ce2be8dc25455e1f91df71bfa12ad37d7af1092ae736f3a6cd0e37bc7810596" + +[[package]] +name = "strsim" +version = "0.11.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "7da8b5736845d9f2fcb837ea5d9e2628564b3b043a70948a3f0b778838c5fb4f" + +[[package]] +name = "subtle" +version = "2.6.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "13c2bddecc57b384dee18652358fb23172facb8a2c51ccc10d74c157bdea3292" + +[[package]] +name = "syn" +version = "2.0.117" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e665b8803e7b1d2a727f4023456bbbbe74da67099c585258af0ad9c5013b9b99" +dependencies = [ + "proc-macro2", + "quote", + "unicode-ident", +] + +[[package]] +name = "sync_wrapper" +version = "1.0.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0bf256ce5efdfa370213c1dabab5935a12e49f2c58d15e9eac2870d3b4f27263" +dependencies = [ + "futures-core", +] + +[[package]] +name = "synstructure" +version = "0.13.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "728a70f3dbaf5bab7f0c4b1ac8d7ae5ea60a4b5549c8a5914361c99147a709d2" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "sysinfo" +version = "0.38.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "92ab6a2f8bfe508deb3c6406578252e491d299cbbf3bc0529ecc3313aee4a52f" +dependencies = [ + "libc", + "memchr", + "ntapi", + "objc2-core-foundation", + "objc2-io-kit", + "windows", +] + +[[package]] +name = "thiserror" +version = "2.0.18" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "4288b5bcbc7920c07a1149a35cf9590a2aa808e0bc1eafaade0b80947865fbc4" +dependencies = [ + "thiserror-impl", +] + +[[package]] +name = "thiserror-impl" +version = "2.0.18" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ebc4ee7f67670e9b64d05fa4253e753e016c6c95ff35b89b7941d6b856dec1d5" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "thread_local" +version = "1.1.9" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "f60246a4944f24f6e018aa17cdeffb7818b76356965d03b07d6a9886e8962185" +dependencies = [ + "cfg-if", +] + +[[package]] +name = "tinystr" +version = "0.8.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "c8323304221c2a851516f22236c5722a72eaa19749016521d6dff0824447d96d" +dependencies = [ + "displaydoc", + "zerovec", +] + +[[package]] +name = "tinyvec" +version = "1.11.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "3e61e67053d25a4e82c844e8424039d9745781b3fc4f32b8d55ed50f5f667ef3" +dependencies = [ + "tinyvec_macros", +] + +[[package]] +name = "tinyvec_macros" +version = "0.1.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "1f3ccbac311fea05f86f61904b462b55fb3df8837a366dfc601a0161d0532f20" + +[[package]] +name = "tokio" +version = "1.52.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "8fc7f01b389ac15039e4dc9531aa973a135d7a4135281b12d7c1bc79fd57fffe" +dependencies = [ + "bytes", + "libc", + "mio", + "pin-project-lite", + "signal-hook-registry", + "socket2", + "tokio-macros", + "windows-sys 0.61.2", +] + +[[package]] +name = "tokio-macros" +version = "2.7.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "385a6cb71ab9ab790c5fe8d67f1645e6c450a7ce006a33de03daa956cf70a496" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "tokio-rustls" +version = "0.26.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "1729aa945f29d91ba541258c8df89027d5792d85a8841fb65e8bf0f4ede4ef61" +dependencies = [ + "rustls", + "tokio", +] + +[[package]] +name = "tokio-stream" +version = "0.1.18" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "32da49809aab5c3bc678af03902d4ccddea2a87d028d86392a4b1560c6906c70" +dependencies = [ + "futures-core", + "pin-project-lite", + "tokio", +] + +[[package]] +name = "tokio-tungstenite" +version = "0.28.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d25a406cddcc431a75d3d9afc6a7c0f7428d4891dd973e4d54c56b46127bf857" +dependencies = [ + "futures-util", + "log", + "rustls", + "rustls-native-certs", + "rustls-pki-types", + "tokio", + "tokio-rustls", + "tungstenite", +] + +[[package]] +name = "tokio-util" +version = "0.7.18" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9ae9cec805b01e8fc3fd2fe289f89149a9b66dd16786abd8b19cfa7b48cb0098" +dependencies = [ + "bytes", + "futures-core", + "futures-sink", + "pin-project-lite", + "tokio", +] + +[[package]] +name = "tonic" +version = "0.14.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ac2a5518c70fa84342385732db33fb3f44bc4cc748936eb5833d2df34d6445ef" +dependencies = [ + "async-trait", + "base64", + "bytes", + "http", + "http-body", + "http-body-util", + "hyper", + "hyper-timeout", + "hyper-util", + "percent-encoding", + "pin-project", + "sync_wrapper", + "tokio", + "tokio-stream", + "tower", + "tower-layer", + "tower-service", + "tracing", +] + +[[package]] +name = "tonic-prost" +version = "0.14.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "50849f68853be452acf590cde0b146665b8d507b3b8af17261df47e02c209ea0" +dependencies = [ + "bytes", + "prost", + "tonic", +] + +[[package]] +name = "tower" +version = "0.5.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ebe5ef63511595f1344e2d5cfa636d973292adc0eec1f0ad45fae9f0851ab1d4" +dependencies = [ + "futures-core", + "futures-util", + "indexmap", + "pin-project-lite", + "slab", + "sync_wrapper", + "tokio", + "tokio-util", + "tower-layer", + "tower-service", + "tracing", +] + +[[package]] +name = "tower-http" +version = "0.6.10" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "68d6fdd9f81c2819c9a8b0e0cd91660e7746a8e6ea2ba7c6b2b057985f6bcb51" +dependencies = [ + "bitflags", + "bytes", + "futures-util", + "http", + "http-body", + "pin-project-lite", + "tower", + "tower-layer", + "tower-service", + "url", +] + +[[package]] +name = "tower-layer" +version = "0.3.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "121c2a6cda46980bb0fcd1647ffaf6cd3fc79a013de288782836f6df9c48780e" + +[[package]] +name = "tower-service" +version = "0.3.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "8df9b6e13f2d32c91b9bd719c00d1958837bc7dec474d94952798cc8e69eeec3" + +[[package]] +name = "tracing" +version = "0.1.44" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "63e71662fa4b2a2c3a26f570f037eb95bb1f85397f3cd8076caed2f026a6d100" +dependencies = [ + "pin-project-lite", + "tracing-attributes", + "tracing-core", +] + +[[package]] +name = "tracing-attributes" +version = "0.1.31" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "7490cfa5ec963746568740651ac6781f701c9c5ea257c58e057f3ba8cf69e8da" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "tracing-core" +version = "0.1.36" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "db97caf9d906fbde555dd62fa95ddba9eecfd14cb388e4f491a66d74cd5fb79a" +dependencies = [ + "once_cell", + "valuable", +] + +[[package]] +name = "tracing-log" +version = "0.2.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ee855f1f400bd0e5c02d150ae5de3840039a3f54b025156404e34c23c03f47c3" +dependencies = [ + "log", + "once_cell", + "tracing-core", +] + +[[package]] +name = "tracing-subscriber" +version = "0.3.23" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "cb7f578e5945fb242538965c2d0b04418d38ec25c79d160cd279bf0731c8d319" +dependencies = [ + "matchers", + "nu-ansi-term", + "once_cell", + "regex-automata", + "sharded-slab", + "smallvec", + "thread_local", + "tracing", + "tracing-core", + "tracing-log", +] + +[[package]] +name = "try-lock" +version = "0.2.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e421abadd41a4225275504ea4d6566923418b7f05506fbc9c0fe86ba7396114b" + +[[package]] +name = "tungstenite" +version = "0.28.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "8628dcc84e5a09eb3d8423d6cb682965dea9133204e8fb3efee74c2a0c259442" +dependencies = [ + "bytes", + "data-encoding", + "http", + "httparse", + "log", + "rand", + "rustls", + "rustls-pki-types", + "sha1", + "thiserror", + "utf-8", +] + +[[package]] +name = "typenum" +version = "1.20.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "40ce102ab67701b8526c123c1bab5cbe42d7040ccfd0f64af1a385808d2f43de" + +[[package]] +name = "unarray" +version = "0.1.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "eaea85b334db583fe3274d12b4cd1880032beab409c0d774be044d4480ab9a94" + +[[package]] +name = "unicode-ident" +version = "1.0.24" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e6e4313cd5fcd3dad5cafa179702e2b244f760991f45397d14d4ebf38247da75" + +[[package]] +name = "unicode-xid" +version = "0.2.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ebc1c04c71510c7f702b52b7c350734c9ff1295c464a03335b00bb84fc54f853" + +[[package]] +name = "unsafe-libyaml" +version = "0.2.11" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "673aac59facbab8a9007c7f6108d11f63b603f7cabff99fabf650fea5c32b861" + +[[package]] +name = "untrusted" +version = "0.9.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "8ecb6da28b8a351d773b68d5825ac39017e680750f980f3a1a85cd8dd28a47c1" + +[[package]] +name = "url" +version = "2.5.8" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ff67a8a4397373c3ef660812acab3268222035010ab8680ec4215f38ba3d0eed" +dependencies = [ + "form_urlencoded", + "idna", + "percent-encoding", + "serde", +] + +[[package]] +name = "utf-8" +version = "0.7.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "09cc8ee72d2a9becf2f2febe0205bbed8fc6615b7cb429ad062dc7b7ddd036a9" + +[[package]] +name = "utf8_iter" +version = "1.0.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "b6c140620e7ffbb22c2dee59cafe6084a59b5ffc27a8859a5f0d494b5d52b6be" + +[[package]] +name = "utf8parse" +version = "0.2.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "06abde3611657adf66d383f00b093d7faecc7fa57071cce2578660c9f1010821" + +[[package]] +name = "uuid" +version = "1.23.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ddd74a9687298c6858e9b88ec8935ec45d22e8fd5e6394fa1bd4e99a87789c76" +dependencies = [ + "getrandom 0.4.2", + "js-sys", + "serde_core", + "wasm-bindgen", +] + +[[package]] +name = "valuable" +version = "0.1.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ba73ea9cf16a25df0c8caa16c51acb937d5712a8429db78a3ee29d5dcacd3a65" + +[[package]] +name = "version_check" +version = "0.9.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0b928f33d975fc6ad9f86c8f283853ad26bdd5b10b7f1542aa2fa15e2289105a" + +[[package]] +name = "want" +version = "0.3.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "bfa7760aed19e106de2c7c0b581b509f2f25d3dacaf737cb82ac61bc6d760b0e" +dependencies = [ + "try-lock", +] + +[[package]] +name = "wasi" +version = "0.11.1+wasi-snapshot-preview1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ccf3ec651a847eb01de73ccad15eb7d99f80485de043efb2f370cd654f4ea44b" + +[[package]] +name = "wasip2" +version = "1.0.3+wasi-0.2.9" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "20064672db26d7cdc89c7798c48a0fdfac8213434a1186e5ef29fd560ae223d6" +dependencies = [ + "wit-bindgen 0.57.1", +] + +[[package]] +name = "wasip3" +version = "0.4.0+wasi-0.3.0-rc-2026-01-06" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "5428f8bf88ea5ddc08faddef2ac4a67e390b88186c703ce6dbd955e1c145aca5" +dependencies = [ + "wit-bindgen 0.51.0", +] + +[[package]] +name = "wasm-bindgen" +version = "0.2.121" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "49ace1d07c165b0864824eee619580c4689389afa9dc9ed3a4c75040d82e6790" +dependencies = [ + "cfg-if", + "once_cell", + "rustversion", + "wasm-bindgen-macro", + "wasm-bindgen-shared", +] + +[[package]] +name = "wasm-bindgen-futures" +version = "0.4.71" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "96492d0d3ffba25305a7dc88720d250b1401d7edca02cc3bcd50633b424673b8" +dependencies = [ + "js-sys", + "wasm-bindgen", +] + +[[package]] +name = "wasm-bindgen-macro" +version = "0.2.121" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "8e68e6f4afd367a562002c05637acb8578ff2dea1943df76afb9e83d177c8578" +dependencies = [ + "quote", + "wasm-bindgen-macro-support", +] + +[[package]] +name = "wasm-bindgen-macro-support" +version = "0.2.121" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d95a9ec35c64b2a7cb35d3fead40c4238d0940c86d107136999567a4703259f2" +dependencies = [ + "bumpalo", + "proc-macro2", + "quote", + "syn", + "wasm-bindgen-shared", +] + +[[package]] +name = "wasm-bindgen-shared" +version = "0.2.121" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "c4e0100b01e9f0d03189a92b96772a1fb998639d981193d7dbab487302513441" +dependencies = [ + "unicode-ident", +] + +[[package]] +name = "wasm-encoder" +version = "0.244.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "990065f2fe63003fe337b932cfb5e3b80e0b4d0f5ff650e6985b1048f62c8319" +dependencies = [ + "leb128fmt", + "wasmparser", +] + +[[package]] +name = "wasm-metadata" +version = "0.244.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "bb0e353e6a2fbdc176932bbaab493762eb1255a7900fe0fea1a2f96c296cc909" +dependencies = [ + "anyhow", + "indexmap", + "wasm-encoder", + "wasmparser", +] + +[[package]] +name = "wasmparser" +version = "0.244.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "47b807c72e1bac69382b3a6fb3dbe8ea4c0ed87ff5629b8685ae6b9a611028fe" +dependencies = [ + "bitflags", + "hashbrown 0.15.5", + "indexmap", + "semver", +] + +[[package]] +name = "web-sys" +version = "0.3.98" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "4b572dff8bcf38bad0fa19729c89bb5748b2b9b1d8be70cf90df697e3a8f32aa" +dependencies = [ + "js-sys", + "wasm-bindgen", +] + +[[package]] +name = "web-time" +version = "1.1.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "5a6580f308b1fad9207618087a65c04e7a10bc77e02c8e84e9b00dd4b12fa0bb" +dependencies = [ + "js-sys", + "wasm-bindgen", +] + +[[package]] +name = "webpki-roots" +version = "1.0.7" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "52f5ee44c96cf55f1b349600768e3ece3a8f26010c05265ab73f945bb1a2eb9d" +dependencies = [ + "rustls-pki-types", +] + +[[package]] +name = "winapi" +version = "0.3.9" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "5c839a674fcd7a98952e593242ea400abe93992746761e38641405d28b00f419" +dependencies = [ + "winapi-i686-pc-windows-gnu", + "winapi-x86_64-pc-windows-gnu", +] + +[[package]] +name = "winapi-i686-pc-windows-gnu" +version = "0.4.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ac3b87c63620426dd9b991e5ce0329eff545bccbbb34f3be09ff6fb6ab51b7b6" + +[[package]] +name = "winapi-x86_64-pc-windows-gnu" +version = "0.4.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "712e227841d057c1ee1cd2fb22fa7e5a5461ae8e48fa2ca79ec42cfc1931183f" + +[[package]] +name = "windows" +version = "0.62.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "527fadee13e0c05939a6a05d5bd6eec6cd2e3dbd648b9f8e447c6518133d8580" +dependencies = [ + "windows-collections", + "windows-core", + "windows-future", + "windows-numerics", +] + +[[package]] +name = "windows-collections" +version = "0.3.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "23b2d95af1a8a14a3c7367e1ed4fc9c20e0a26e79551b1454d72583c97cc6610" +dependencies = [ + "windows-core", +] + +[[package]] +name = "windows-core" +version = "0.62.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "b8e83a14d34d0623b51dce9581199302a221863196a1dde71a7663a4c2be9deb" +dependencies = [ + "windows-implement", + "windows-interface", + "windows-link", + "windows-result", + "windows-strings", +] + +[[package]] +name = "windows-future" +version = "0.3.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e1d6f90251fe18a279739e78025bd6ddc52a7e22f921070ccdc67dde84c605cb" +dependencies = [ + "windows-core", + "windows-link", + "windows-threading", +] + +[[package]] +name = "windows-implement" +version = "0.60.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "053e2e040ab57b9dc951b72c264860db7eb3b0200ba345b4e4c3b14f67855ddf" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "windows-interface" +version = "0.59.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "3f316c4a2570ba26bbec722032c4099d8c8bc095efccdc15688708623367e358" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "windows-link" +version = "0.2.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "f0805222e57f7521d6a62e36fa9163bc891acd422f971defe97d64e70d0a4fe5" + +[[package]] +name = "windows-numerics" +version = "0.3.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "6e2e40844ac143cdb44aead537bbf727de9b044e107a0f1220392177d15b0f26" +dependencies = [ + "windows-core", + "windows-link", +] + +[[package]] +name = "windows-result" +version = "0.4.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "7781fa89eaf60850ac3d2da7af8e5242a5ea78d1a11c49bf2910bb5a73853eb5" +dependencies = [ + "windows-link", +] + +[[package]] +name = "windows-strings" +version = "0.5.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "7837d08f69c77cf6b07689544538e017c1bfcf57e34b4c0ff58e6c2cd3b37091" +dependencies = [ + "windows-link", +] + +[[package]] +name = "windows-sys" +version = "0.52.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "282be5f36a8ce781fad8c8ae18fa3f9beff57ec1b52cb3de0789201425d9a33d" +dependencies = [ + "windows-targets 0.52.6", +] + +[[package]] +name = "windows-sys" +version = "0.60.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "f2f500e4d28234f72040990ec9d39e3a6b950f9f22d3dba18416c35882612bcb" +dependencies = [ + "windows-targets 0.53.5", +] + +[[package]] +name = "windows-sys" +version = "0.61.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ae137229bcbd6cdf0f7b80a31df61766145077ddf49416a728b02cb3921ff3fc" +dependencies = [ + "windows-link", +] + +[[package]] +name = "windows-targets" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9b724f72796e036ab90c1021d4780d4d3d648aca59e491e6b98e725b84e99973" +dependencies = [ + "windows_aarch64_gnullvm 0.52.6", + "windows_aarch64_msvc 0.52.6", + "windows_i686_gnu 0.52.6", + "windows_i686_gnullvm 0.52.6", + "windows_i686_msvc 0.52.6", + "windows_x86_64_gnu 0.52.6", + "windows_x86_64_gnullvm 0.52.6", + "windows_x86_64_msvc 0.52.6", +] + +[[package]] +name = "windows-targets" +version = "0.53.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "4945f9f551b88e0d65f3db0bc25c33b8acea4d9e41163edf90dcd0b19f9069f3" +dependencies = [ + "windows-link", + "windows_aarch64_gnullvm 0.53.1", + "windows_aarch64_msvc 0.53.1", + "windows_i686_gnu 0.53.1", + "windows_i686_gnullvm 0.53.1", + "windows_i686_msvc 0.53.1", + "windows_x86_64_gnu 0.53.1", + "windows_x86_64_gnullvm 0.53.1", + "windows_x86_64_msvc 0.53.1", +] + +[[package]] +name = "windows-threading" +version = "0.2.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "3949bd5b99cafdf1c7ca86b43ca564028dfe27d66958f2470940f73d86d75b37" +dependencies = [ + "windows-link", +] + +[[package]] +name = "windows_aarch64_gnullvm" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "32a4622180e7a0ec044bb555404c800bc9fd9ec262ec147edd5989ccd0c02cd3" + +[[package]] +name = "windows_aarch64_gnullvm" +version = "0.53.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "a9d8416fa8b42f5c947f8482c43e7d89e73a173cead56d044f6a56104a6d1b53" + +[[package]] +name = "windows_aarch64_msvc" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "09ec2a7bb152e2252b53fa7803150007879548bc709c039df7627cabbd05d469" + +[[package]] +name = "windows_aarch64_msvc" +version = "0.53.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "b9d782e804c2f632e395708e99a94275910eb9100b2114651e04744e9b125006" + +[[package]] +name = "windows_i686_gnu" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "8e9b5ad5ab802e97eb8e295ac6720e509ee4c243f69d781394014ebfe8bbfa0b" + +[[package]] +name = "windows_i686_gnu" +version = "0.53.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "960e6da069d81e09becb0ca57a65220ddff016ff2d6af6a223cf372a506593a3" + +[[package]] +name = "windows_i686_gnullvm" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0eee52d38c090b3caa76c563b86c3a4bd71ef1a819287c19d586d7334ae8ed66" + +[[package]] +name = "windows_i686_gnullvm" +version = "0.53.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "fa7359d10048f68ab8b09fa71c3daccfb0e9b559aed648a8f95469c27057180c" + +[[package]] +name = "windows_i686_msvc" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "240948bc05c5e7c6dabba28bf89d89ffce3e303022809e73deaefe4f6ec56c66" + +[[package]] +name = "windows_i686_msvc" +version = "0.53.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "1e7ac75179f18232fe9c285163565a57ef8d3c89254a30685b57d83a38d326c2" + +[[package]] +name = "windows_x86_64_gnu" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "147a5c80aabfbf0c7d901cb5895d1de30ef2907eb21fbbab29ca94c5b08b1a78" + +[[package]] +name = "windows_x86_64_gnu" +version = "0.53.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9c3842cdd74a865a8066ab39c8a7a473c0778a3f29370b5fd6b4b9aa7df4a499" + +[[package]] +name = "windows_x86_64_gnullvm" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "24d5b23dc417412679681396f2b49f3de8c1473deb516bd34410872eff51ed0d" + +[[package]] +name = "windows_x86_64_gnullvm" +version = "0.53.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0ffa179e2d07eee8ad8f57493436566c7cc30ac536a3379fdf008f47f6bb7ae1" + +[[package]] +name = "windows_x86_64_msvc" +version = "0.52.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "589f6da84c646204747d1270a2a5661ea66ed1cced2631d546fdfb155959f9ec" + +[[package]] +name = "windows_x86_64_msvc" +version = "0.53.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d6bbff5f0aada427a1e5a6da5f1f98158182f26556f345ac9e04d36d0ebed650" + +[[package]] +name = "wit-bindgen" +version = "0.51.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d7249219f66ced02969388cf2bb044a09756a083d0fab1e566056b04d9fbcaa5" +dependencies = [ + "wit-bindgen-rust-macro", +] + +[[package]] +name = "wit-bindgen" +version = "0.57.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "1ebf944e87a7c253233ad6766e082e3cd714b5d03812acc24c318f549614536e" + +[[package]] +name = "wit-bindgen-core" +version = "0.51.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ea61de684c3ea68cb082b7a88508a8b27fcc8b797d738bfc99a82facf1d752dc" +dependencies = [ + "anyhow", + "heck", + "wit-parser", +] + +[[package]] +name = "wit-bindgen-rust" +version = "0.51.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "b7c566e0f4b284dd6561c786d9cb0142da491f46a9fbed79ea69cdad5db17f21" +dependencies = [ + "anyhow", + "heck", + "indexmap", + "prettyplease", + "syn", + "wasm-metadata", + "wit-bindgen-core", + "wit-component", +] + +[[package]] +name = "wit-bindgen-rust-macro" +version = "0.51.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0c0f9bfd77e6a48eccf51359e3ae77140a7f50b1e2ebfe62422d8afdaffab17a" +dependencies = [ + "anyhow", + "prettyplease", + "proc-macro2", + "quote", + "syn", + "wit-bindgen-core", + "wit-bindgen-rust", +] + +[[package]] +name = "wit-component" +version = "0.244.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9d66ea20e9553b30172b5e831994e35fbde2d165325bec84fc43dbf6f4eb9cb2" +dependencies = [ + "anyhow", + "bitflags", + "indexmap", + "log", + "serde", + "serde_derive", + "serde_json", + "wasm-encoder", + "wasm-metadata", + "wasmparser", + "wit-parser", +] + +[[package]] +name = "wit-parser" +version = "0.244.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ecc8ac4bc1dc3381b7f59c34f00b67e18f910c2c0f50015669dde7def656a736" +dependencies = [ + "anyhow", + "id-arena", + "indexmap", + "log", + "semver", + "serde", + "serde_derive", + "serde_json", + "unicode-xid", + "wasmparser", +] + +[[package]] +name = "writeable" +version = "0.6.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "1ffae5123b2d3fc086436f8834ae3ab053a283cfac8fe0a0b8eaae044768a4c4" + +[[package]] +name = "yoke" +version = "0.8.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "abe8c5fda708d9ca3df187cae8bfb9ceda00dd96231bed36e445a1a48e66f9ca" +dependencies = [ + "stable_deref_trait", + "yoke-derive", + "zerofrom", +] + +[[package]] +name = "yoke-derive" +version = "0.8.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "de844c262c8848816172cef550288e7dc6c7b7814b4ee56b3e1553f275f1858e" +dependencies = [ + "proc-macro2", + "quote", + "syn", + "synstructure", +] + +[[package]] +name = "zerocopy" +version = "0.8.48" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "eed437bf9d6692032087e337407a86f04cd8d6a16a37199ed57949d415bd68e9" +dependencies = [ + "zerocopy-derive", +] + +[[package]] +name = "zerocopy-derive" +version = "0.8.48" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "70e3cd084b1788766f53af483dd21f93881ff30d7320490ec3ef7526d203bad4" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "zerofrom" +version = "0.1.8" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0ec05a11813ea801ff6d75110ad09cd0824ddba17dfe17128ea0d5f68e6c5272" +dependencies = [ + "zerofrom-derive", +] + +[[package]] +name = "zerofrom-derive" +version = "0.1.7" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "11532158c46691caf0f2593ea8358fed6bbf68a0315e80aae9bd41fbade684a1" +dependencies = [ + "proc-macro2", + "quote", + "syn", + "synstructure", +] + +[[package]] +name = "zeroize" +version = "1.8.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "b97154e67e32c85465826e8bcc1c59429aaaf107c1e4a9e53c8d8ccd5eff88d0" + +[[package]] +name = "zerotrie" +version = "0.2.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0f9152d31db0792fa83f70fb2f83148effb5c1f5b8c7686c3459e361d9bc20bf" +dependencies = [ + "displaydoc", + "yoke", + "zerofrom", +] + +[[package]] +name = "zerovec" +version = "0.11.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "90f911cbc359ab6af17377d242225f4d75119aec87ea711a880987b18cd7b239" +dependencies = [ + "yoke", + "zerofrom", + "zerovec-derive", +] + +[[package]] +name = "zerovec-derive" +version = "0.11.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "625dc425cab0dca6dc3c3319506e6593dcb08a9f387ea3b284dbd52a92c40555" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "zmij" +version = "1.0.21" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "b8848ee67ecc8aedbaf3e4122217aff892639231befc6a1b58d29fff4c2cabaa" diff --git a/context-compaction/Cargo.toml b/context-compaction/Cargo.toml new file mode 100644 index 00000000..bbde2fb8 --- /dev/null +++ b/context-compaction/Cargo.toml @@ -0,0 +1,38 @@ +[workspace] + +[package] +name = "context-compaction" +version = "0.1.0" +description = "Out-of-band session-history compactor. Subscribes to agent::events::TurnEnd, summarises older turns via provider-router, and appends a session-tree Compaction entry so the next turn reads a compressed transcript." +edition = "2021" +license = "Apache-2.0" +repository = "https://github.com/iii-hq/workers" +authors = ["iii"] +publish = false +build = "build.rs" + +[lib] +name = "context_compaction" +path = "src/lib.rs" + +[[bin]] +name = "context-compaction" +path = "src/main.rs" + +[dependencies] +iii-sdk = "=0.11.7-next.3" +serde = { version = "1", features = ["derive"] } +serde_json = "1" +serde_yaml = "0.9" +tokio = { version = "1", features = ["macros", "rt-multi-thread", "signal", "sync", "time"] } +anyhow = "1" +chrono = { version = "0.4", default-features = false, features = ["clock", "serde"] } +tracing = "0.1" +tracing-subscriber = { version = "0.3", features = ["fmt", "env-filter"] } +clap = { version = "4", features = ["derive", "env"] } + +[dev-dependencies] +serde_json = "1" + +[lints.rust] +unsafe_code = "forbid" diff --git a/context-compaction/README.md b/context-compaction/README.md new file mode 100644 index 00000000..12fa1a2a --- /dev/null +++ b/context-compaction/README.md @@ -0,0 +1,80 @@ +# context-compaction + +Out-of-band session-history compactor. Subscribes to `agent::events` on the +iii bus, watches every `TurnEnd` event for the per-turn token count, and +when the running transcript crosses the configured threshold, summarises +the older turns via `router::stream_assistant` and writes a +`SessionEntry::Compaction` to the session-tree. + +The orchestrator's next turn picks up the compacted transcript via +`session-tree::load_messages` (which already filters `Compaction` entries +out of the active path). No LLM-facing tools. Invisible to the model — the +only observable artefact is one extra entry in the session-tree per +compaction. + +## Why + +Long sessions ship the full conversation on every request because the +Anthropic / OpenAI APIs are stateless. Shipping a compacted prefix instead +of the raw transcript trades one cheap summariser LLM call for ongoing +per-turn savings against the larger, slower main-loop model. The +heuristic for "long enough to compact" is one env-tunable threshold; +defaults are conservative. + +## Install + +```bash +make compaction # spawn alongside the harness engine +``` + +The worker is intentionally not in `harness/iii.worker.yaml` `dependencies:` +— the upstream registry doesn't index it yet, so it ships as a background +process the harness `Makefile` starts after `engine` is up. Mirrors the +`iii-observability` "optional, side-of-config" pattern. PID file lives at +`$PIDS_DIR/context-compaction.pid`; logs at `$LOGS_DIR/context-compaction.log`. + +## Configuration + +All knobs are env vars; defaults are baked in. + +| Variable | Default | Effect | +|---|---|---| +| `COMPACT_TRIGGER_TOKENS` | `60000` | Token threshold above which `TurnEnd` triggers compaction | +| `COMPACT_KEEP_RECENT_TURNS` | `3` | Number of trailing turns kept verbatim in the post-compaction transcript | +| `COMPACT_SUMMARIZER_PROVIDER` | `anthropic` | Provider for the summariser LLM call | +| `COMPACT_SUMMARIZER_MODEL` | `claude-haiku-4-5` | Model for the summariser LLM call (use a cheap fast model) | + +`usage_total` for the threshold check sums `input + output + cache_read` so +the trigger fires on *true* transcript size, not on what happens to be +cache-hot. `cache_write` is excluded — it costs more per turn but doesn't +grow the transcript. + +## Coordination + +Single-writer correctness across multiple worker instances or rapid-fire +events is enforced via a nonce-and-readback lease at +`session//compaction_lease`. The engine's `state::*` ops have no CAS +primitive, so each acquisition writes a unique nonce and confirms +ownership via readback (`state::set` is last-write-wins; exactly one +writer sees its own nonce survive). Lease TTL is 300s — comfortably +above the 120s summariser timeout so a slow LLM call can't expire its +own lease and let a peer start a duplicate compaction. + +When a compaction lands, the worker stamps +`session//last_compaction_at` with `chrono::Utc::now().timestamp_millis()`. +The orchestrator watches that key and rebuilds its hot +`session//messages` view from session-tree on the next +`handle_streaming` entry — no synchronous coupling, no shared writers. + +## Testing + +```bash +cargo test # unit tests +cargo test --test manifest # manifest CLI smoke test +``` + +Pure logic (lease nonces, timestamp parsing, threshold policy, summary +rendering) is covered by the unit suite. End-to-end orchestration paths +(`acquire_lease`, `summarize_and_append`, `handle_event`) require a live +iii engine and aren't unit-tested today; live exercise is documented in +the implementation plan. diff --git a/context-compaction/build.rs b/context-compaction/build.rs new file mode 100644 index 00000000..81caa36d --- /dev/null +++ b/context-compaction/build.rs @@ -0,0 +1,6 @@ +fn main() { + println!( + "cargo:rustc-env=TARGET={}", + std::env::var("TARGET").unwrap() + ); +} diff --git a/context-compaction/iii.worker.yaml b/context-compaction/iii.worker.yaml new file mode 100644 index 00000000..06fd8c40 --- /dev/null +++ b/context-compaction/iii.worker.yaml @@ -0,0 +1,18 @@ +iii: v1 +name: context-compaction +language: rust +deploy: binary +manifest: Cargo.toml +bin: context-compaction +description: Out-of-band session-history compactor. Subscribes to agent::events on each turn end, summarises older turns via provider-router, and writes a session-tree Compaction entry so the next turn reads a compressed transcript. + +runtime: + kind: rust + +scripts: + install: cargo build + start: cargo run + +dependencies: + session: "^0.2.0" + provider-router: "^0.2.0" diff --git a/context-compaction/prompts/compaction.txt b/context-compaction/prompts/compaction.txt new file mode 100644 index 00000000..c031d942 --- /dev/null +++ b/context-compaction/prompts/compaction.txt @@ -0,0 +1,27 @@ +You are an anchored context summarization assistant for an iii harness coding session. + +Summarize only the conversation history you are given. The newest turns are kept verbatim outside your summary, so focus on the older context that still matters for continuing the work. + +If the prompt includes a block, treat it as the current anchored summary. Update it with the new history by preserving still-true details, removing stale details, and merging in new facts. Never drop a still-true detail just because the new turns don't mention it. + +Follow this exact output structure. Keep every section heading even if a section is empty (write `(none)`). Prefer terse bullets to paragraphs. Preserve exact file paths, bus function ids (e.g. `shell::fs::read`, `agent_call`, `state::set`), session ids, call_ids, and other identifiers verbatim. Target a total length under 10% of the original conversation token count. + +# Goal +One or two bullets describing what the user is actually trying to accomplish in this session. + +# Worked on so far +Chronological bullets of the meaningful actions taken. Each bullet should name the agent_call function used and the salient argument or result (path, command, summary of effect). Skip routine reads and noise. + +# Files touched +Two sub-lists: `read:` (paths the agent inspected) and `modified:` (paths the agent created or edited). + +# Open questions +Anything the user asked or implied that has not yet been answered. Keep verbatim where possible. + +# Plan state +Where the agent currently is in any multi-step plan: which step just completed, which is next. + +# Identifiers worth remembering +Session ids, sandbox ids, branch names, ticket ids, urls, call_ids, anything the agent will need to refer to later. One per bullet, prefixed with what kind of id it is. + +Do not answer the conversation itself. Do not mention that you are summarising, compacting, or merging context. Respond in the same language as the conversation. diff --git a/context-compaction/skill.md b/context-compaction/skill.md new file mode 100644 index 00000000..19c6b18b --- /dev/null +++ b/context-compaction/skill.md @@ -0,0 +1,29 @@ +# context-compaction + +Out-of-band session-history compactor. + +Subscribes to `agent::events::TurnEnd` on every assistant turn. When the +running token count for a session exceeds a configurable threshold, the +worker summarises the older portion of the transcript via +`router::stream_assistant` and appends a `Compaction` entry to the +session-tree. `session-tree::load_messages` filters Compaction entries out +of the active-path transcript, so the next assistant turn reads a +compressed history without any orchestrator changes — except a tiny +"reload-from-session-tree-when-a-fresh-Compaction-is-at-the-tail" check. + +The worker has no LLM-facing tools (`tools: []`). It is invisible to the +model; the only artefact it produces is a side-effect on the session-tree. + +## Configuration + +| Env var | Default | What it does | +|---|---|---| +| `COMPACT_TRIGGER_TOKENS` | `60000` | Trigger when running `usage.input + output + cache_read` since the last Compaction crosses this many tokens. | +| `COMPACT_KEEP_RECENT_TURNS` | `3` | Number of trailing assistant/user turns kept verbatim. Older turns become the summary input. | +| `COMPACT_SUMMARIZER_PROVIDER` | (orchestrator's provider) | Cheap-model override for the summarisation call. | +| `COMPACT_SUMMARIZER_MODEL` | (orchestrator's model) | Cheap-model override. Default to a small model like `claude-haiku-4-5` to keep summarisation cost trivial. | + +## How to disable + +Either drop `context-compaction` from `harness/iii.worker.yaml` or set +`COMPACT_TRIGGER_TOKENS` to a number larger than any realistic session. diff --git a/context-compaction/src/config.rs b/context-compaction/src/config.rs new file mode 100644 index 00000000..33263a5d --- /dev/null +++ b/context-compaction/src/config.rs @@ -0,0 +1,62 @@ +//! Worker configuration. Pure helpers reading env vars with sane defaults. + +const TRIGGER_TOKENS_ENV: &str = "COMPACT_TRIGGER_TOKENS"; +const KEEP_RECENT_TURNS_ENV: &str = "COMPACT_KEEP_RECENT_TURNS"; +const SUMMARIZER_PROVIDER_ENV: &str = "COMPACT_SUMMARIZER_PROVIDER"; +const SUMMARIZER_MODEL_ENV: &str = "COMPACT_SUMMARIZER_MODEL"; + +const DEFAULT_TRIGGER_TOKENS: u64 = 60_000; +const DEFAULT_KEEP_RECENT_TURNS: usize = 3; + +/// Token-count threshold above which compaction fires. Reads from +/// `COMPACT_TRIGGER_TOKENS`; falls back to a sensible default. +pub fn trigger_tokens() -> u64 { + std::env::var(TRIGGER_TOKENS_ENV) + .ok() + .and_then(|s| s.parse::().ok()) + .filter(|&v| v > 0) + .unwrap_or(DEFAULT_TRIGGER_TOKENS) +} + +/// Number of trailing turns to keep verbatim. Older turns become summary +/// input. +pub fn keep_recent_turns() -> usize { + std::env::var(KEEP_RECENT_TURNS_ENV) + .ok() + .and_then(|s| s.parse::().ok()) + .filter(|&v| v > 0) + .unwrap_or(DEFAULT_KEEP_RECENT_TURNS) +} + +/// Optional provider override for the summariser call. `None` means +/// "use whatever the orchestrator's run_request had". +pub fn summarizer_provider() -> Option { + std::env::var(SUMMARIZER_PROVIDER_ENV) + .ok() + .filter(|s| !s.is_empty()) +} + +/// Optional model override for the summariser call. +pub fn summarizer_model() -> Option { + std::env::var(SUMMARIZER_MODEL_ENV) + .ok() + .filter(|s| !s.is_empty()) +} + +#[cfg(test)] +mod tests { + use super::*; + + #[test] + fn defaults_are_within_safe_range() { + // Don't assert exact equality because env vars may be set in + // CI/dev — just bound the values. + let t = trigger_tokens(); + assert!(t >= 1_000, "trigger threshold must be at least 1k tokens"); + assert!(t <= 1_000_000, "trigger threshold must be sane"); + + let k = keep_recent_turns(); + assert!(k >= 1, "must keep at least 1 recent turn"); + assert!(k <= 100, "100 recent turns is already an over-budget cap"); + } +} diff --git a/context-compaction/src/lib.rs b/context-compaction/src/lib.rs new file mode 100644 index 00000000..b77ad373 --- /dev/null +++ b/context-compaction/src/lib.rs @@ -0,0 +1,403 @@ +//! `context-compaction` worker. +//! +//! Subscribes to `agent::events` on the iii bus. On every `TurnEnd`, checks +//! whether the running transcript has crossed the configured token +//! threshold; if so, kicks off an out-of-band summarisation via +//! `router::stream_assistant` and appends a `Compaction` entry to the +//! session-tree. The orchestrator's next turn picks up the compacted +//! transcript via `session-tree::load_messages`. +//! +//! No LLM-facing tools. Invisible to the model. The only observable +//! artefact is an extra `Compaction` entry in the session-tree. + +pub mod config; +pub mod manifest; +pub mod summarize; +pub mod threshold; + +use std::sync::atomic::{AtomicU64, Ordering}; +use std::sync::Arc; + +use iii_sdk::{ + FunctionRef, IIIError, RegisterFunctionMessage, RegisterTriggerInput, TriggerRequest, Value, + III, +}; +use serde_json::json; + +const AGENT_EVENTS_STREAM: &str = "agent::events"; +const SUBSCRIBER_FN: &str = "context-compaction::on_agent_event"; +const STATE_SCOPE: &str = "agent"; +/// Must exceed `summarize::SUMMARIZER_TIMEOUT_MS / 1000` (currently 120s) so +/// a slow LLM call can't have its own lease expire underneath it and let a +/// peer start a duplicate compaction. 5 min also bounds how long a crashed +/// worker can block compaction: the next TurnEnd after the TTL elapses +/// re-acquires. +const LEASE_TTL_SECS: i64 = 300; + +/// Register the agent::events subscriber and return a handle the binary +/// can hold for the worker's lifetime. +pub fn register_with_iii(iii: &Arc) -> anyhow::Result { + let subscriber_fn = register_subscriber_fn(iii); + iii.register_trigger(RegisterTriggerInput { + trigger_type: "stream".into(), + function_id: SUBSCRIBER_FN.into(), + config: json!({ "stream_name": AGENT_EVENTS_STREAM }), + metadata: None, + }) + .map_err(|e| anyhow::anyhow!(e.to_string()))?; + Ok(Refs { subscriber_fn }) +} + +/// Handles returned by [`register_with_iii`]. Hold for the lifetime of the +/// worker; drop / unregister on shutdown. +pub struct Refs { + pub subscriber_fn: FunctionRef, +} + +fn register_subscriber_fn(iii: &Arc) -> FunctionRef { + let iii_for_handler = Arc::clone(iii); + iii.register_function(( + RegisterFunctionMessage::with_id(SUBSCRIBER_FN.to_string()).with_description( + "Internal: subscribes to agent::events; triggers session compaction on TurnEnd when running tokens exceed the configured threshold." + .to_string(), + ), + move |payload: Value| { + let iii = Arc::clone(&iii_for_handler); + async move { + handle_event(&iii, payload).await; + // Stream subscribers don't return meaningful data; the + // engine ignores anything that isn't an Err. + Ok::(Value::Null) + } + }, + )) +} + +/// Handle a single `agent::events` frame. Returns silently on every +/// non-actionable shape (non-TurnEnd events, missing fields, races) so a +/// malformed frame never aborts the subscriber. +async fn handle_event(iii: &III, payload: Value) { + let Some((session_id, event)) = extract_event_payload(&payload) else { + return; + }; + let Some(usage) = turn_end_usage(&event) else { + return; + }; + let total_tokens = usage_total(&usage); + if !threshold::should_compact(total_tokens) { + return; + } + let Some(nonce) = acquire_lease(iii, &session_id).await else { + tracing::debug!(%session_id, "compaction lease held; skipping"); + return; + }; + if let Err(e) = run_compaction(iii, &session_id).await { + tracing::warn!(%session_id, error = %e, "compaction failed"); + } + release_lease(iii, &session_id, &nonce).await; +} + +/// Decode the stream frame into `(session_id, event_data)`. Mirrors the +/// extraction logic in `harness/src/fanout.rs::extract_event_payload` so +/// both camelCase (`groupId`) and snake_case (`group_id`) envelopes work. +fn extract_event_payload(payload: &Value) -> Option<(String, Value)> { + let session_id = payload + .get("groupId") + .or_else(|| payload.get("group_id")) + .and_then(Value::as_str)? + .to_string(); + let data = payload + .get("event") + .and_then(|e| e.get("data")) + .cloned() + .or_else(|| payload.get("data").cloned()) + .unwrap_or(Value::Null); + Some((session_id, data)) +} + +/// Return `Some(usage)` if the event is a `TurnEnd` carrying a `message` +/// with `usage`. None otherwise — we only decide on TurnEnd to mirror +/// the orchestrator's commit point. +fn turn_end_usage(event: &Value) -> Option { + let kind = event.get("type").and_then(Value::as_str)?; + if kind != "TurnEnd" && kind != "turn_end" { + return None; + } + event + .get("message") + .and_then(|m| m.get("usage")) + .cloned() + .filter(|u| !u.is_null()) +} + +/// Anthropic and OpenAI both report only the *uncached* portion of the +/// prompt as `input` once caching is on. Adding `cache_read` recovers the +/// total transcript size we'd pay for without caching — which is the +/// right number to threshold on, because compaction should fire based on +/// *true* context-window pressure, not on what happens to be cache-hot. +fn usage_total(usage: &Value) -> u64 { + let pick = |k: &str| usage.get(k).and_then(Value::as_u64).unwrap_or(0); + pick("input") + pick("output") + pick("cache_read") +} + +async fn run_compaction(iii: &III, session_id: &str) -> anyhow::Result<()> { + summarize::summarize_and_append(iii, session_id).await +} + +fn lease_key(session_id: &str) -> String { + format!("session/{session_id}/compaction_lease") +} + +/// Monotonic per-process counter mixed into [`mint_lease_nonce`] so two +/// nonces minted in the same nanosecond within one process still diverge. +static NONCE_COUNTER: AtomicU64 = AtomicU64::new(0); + +/// Mint a per-attempt nonce for the lease's "nonce-and-readback" protocol. +/// pid covers cross-process collisions; nanos covers cross-attempt within +/// one process; the counter covers two concurrent attempts in the same +/// nanosecond. +fn mint_lease_nonce() -> String { + let pid = std::process::id(); + let nanos = chrono::Utc::now().timestamp_nanos_opt().unwrap_or(0); + let seq = NONCE_COUNTER.fetch_add(1, Ordering::Relaxed); + format!("{pid}-{nanos}-{seq}") +} + +/// Acquire a single-writer lease for compaction on this session via +/// nonce-and-readback. Returns the winning nonce when granted, `None` when +/// another writer holds an unexpired lease OR raced us and won. +/// +/// Why a nonce: the engine's `state::*` ops have no CAS primitive. A naive +/// check-then-write is a race — two concurrent writers both see an empty +/// lease, both write their timestamp, and both proceed. By stamping a +/// unique nonce and reading it back, exactly one writer sees its own +/// nonce surviving in the store: state::set is last-write-wins, so +/// whichever writer landed last "owns" the readback. Loser sees a +/// different nonce and bails. +async fn acquire_lease(iii: &III, session_id: &str) -> Option { + let key = lease_key(session_id); + let now_ms = chrono::Utc::now().timestamp_millis(); + let now_secs = now_ms / 1000; + + // Bail early if a non-expired lease is already held. Saves the + // readback round-trip in the common no-contention case. + if let Some(existing) = state_get(iii, &key).await { + let ts_secs = read_lease_timestamp_secs(&existing); + if ts_secs > 0 && now_secs - ts_secs < LEASE_TTL_SECS { + return None; + } + } + + let nonce = mint_lease_nonce(); + state_set(iii, &key, json!({ "nonce": nonce, "ts": now_ms })).await; + + let stored = state_get(iii, &key).await; + let stored_nonce = stored + .as_ref() + .and_then(|v| v.get("nonce")) + .and_then(Value::as_str); + if stored_nonce == Some(nonce.as_str()) { + Some(nonce) + } else { + None + } +} + +/// Read the lease's millisecond timestamp from either the current +/// `{nonce, ts}` shape or the legacy bare-i64 shape, and return it in +/// seconds. Returns 0 when the value is missing or unrecognised, which +/// makes any caller treat the lease as expired. +fn read_lease_timestamp_secs(v: &Value) -> i64 { + if let Some(ms) = v.get("ts").and_then(Value::as_i64) { + return ms / 1000; + } + if let Some(secs) = v.as_i64() { + return secs; + } + 0 +} + +/// Release the lease only if we still hold it. Without this check, a +/// long-running compaction that exceeded the TTL could clear the lease +/// belonging to a successor worker. +async fn release_lease(iii: &III, session_id: &str, our_nonce: &str) { + let key = lease_key(session_id); + let stored = state_get(iii, &key).await; + let stored_nonce = stored + .as_ref() + .and_then(|v| v.get("nonce")) + .and_then(Value::as_str); + if stored_nonce == Some(our_nonce) { + state_set(iii, &key, Value::Null).await; + } +} + +/// Stamp the orchestrator-watched key so `handle_streaming` knows it must +/// rebuild its hot `session//messages` view from session-tree before +/// the next assistant turn. Called from +/// [`summarize::summarize_and_append`] after the Compaction entry has +/// been committed. +pub async fn stamp_last_compaction(iii: &III, session_id: &str) { + let now_ms = chrono::Utc::now().timestamp_millis(); + state_set( + iii, + &format!("session/{session_id}/last_compaction_at"), + json!(now_ms), + ) + .await; +} + +async fn state_get(iii: &III, key: &str) -> Option { + iii.trigger(TriggerRequest { + function_id: "state::get".into(), + payload: json!({ "scope": STATE_SCOPE, "key": key }), + action: None, + timeout_ms: None, + }) + .await + .ok() + .filter(|v| !v.is_null()) +} + +async fn state_set(iii: &III, key: &str, value: Value) { + if let Err(e) = iii + .trigger(TriggerRequest { + function_id: "state::set".into(), + payload: json!({ "scope": STATE_SCOPE, "key": key, "value": value }), + action: None, + timeout_ms: None, + }) + .await + { + tracing::warn!(error = %e, %key, "context-compaction: state::set failed"); + } +} + +#[cfg(test)] +mod tests { + use super::*; + + #[test] + fn extract_event_payload_handles_camelcase() { + let env = json!({ + "groupId": "sess-1", + "event": { "data": { "type": "TurnEnd", "message": {} } }, + }); + let (sid, data) = extract_event_payload(&env).expect("extracts"); + assert_eq!(sid, "sess-1"); + assert_eq!(data["type"], "TurnEnd"); + } + + #[test] + fn extract_event_payload_handles_snake_case() { + let env = json!({ + "group_id": "sess-2", + "data": { "type": "TurnEnd" }, + }); + let (sid, data) = extract_event_payload(&env).expect("extracts"); + assert_eq!(sid, "sess-2"); + assert_eq!(data["type"], "TurnEnd"); + } + + #[test] + fn turn_end_usage_extracts_when_present() { + let event = json!({ + "type": "TurnEnd", + "message": { "usage": { "input": 100, "output": 50, "cache_read": 800 } }, + }); + let u = turn_end_usage(&event).expect("has usage"); + assert_eq!(u["input"], 100); + } + + #[test] + fn turn_end_usage_skips_non_turn_end() { + for kind in ["TurnStart", "MessageStart", "FunctionExecutionStart", "AgentStart"] { + let event = json!({ "type": kind, "message": { "usage": { "input": 9999 } }}); + assert!(turn_end_usage(&event).is_none(), "{kind} must be ignored"); + } + } + + #[test] + fn turn_end_usage_skips_missing_usage() { + let event = json!({ "type": "TurnEnd", "message": {} }); + assert!(turn_end_usage(&event).is_none()); + } + + #[test] + fn usage_total_sums_all_three_buckets() { + let u = json!({ "input": 100, "output": 50, "cache_read": 800, "cache_write": 200 }); + // cache_write is intentionally excluded — it counts toward cost + // but not toward transcript size. + assert_eq!(usage_total(&u), 950); + } + + #[test] + fn usage_total_handles_missing_fields() { + let u = json!({ "input": 100 }); + assert_eq!(usage_total(&u), 100); + } + + #[test] + fn lease_key_namespaces_by_session() { + assert!(lease_key("sess-9").contains("sess-9")); + assert!(lease_key("sess-9").contains("compaction_lease")); + } + + #[test] + fn mint_lease_nonce_returns_unique_values_across_calls() { + let n1 = mint_lease_nonce(); + let n2 = mint_lease_nonce(); + let n3 = mint_lease_nonce(); + assert_ne!(n1, n2); + assert_ne!(n2, n3); + assert_ne!(n1, n3); + } + + #[test] + fn mint_lease_nonce_diverges_for_same_nanosecond_calls() { + // The pid+nanos+counter recipe must survive the tightest possible + // back-to-back call. The counter alone guarantees this even when + // nanos repeats. + let mut seen = std::collections::HashSet::new(); + for _ in 0..1000 { + assert!(seen.insert(mint_lease_nonce()), "nonces must not collide"); + } + } + + #[test] + fn read_lease_timestamp_secs_accepts_new_object_shape() { + let v = json!({ "nonce": "abc", "ts": 1_700_000_000_000_i64 }); + assert_eq!(read_lease_timestamp_secs(&v), 1_700_000_000); + } + + #[test] + fn read_lease_timestamp_secs_accepts_legacy_bare_int_shape() { + // Pre-nonce releases wrote a bare i64 (seconds). Restart compat: a + // worker upgraded mid-flight may read a legacy lease and must still + // interpret its TTL. + let v = json!(1_700_000_000_i64); + assert_eq!(read_lease_timestamp_secs(&v), 1_700_000_000); + } + + #[test] + fn read_lease_timestamp_secs_returns_zero_for_garbage() { + // Anything we can't parse must read as "expired" so a writer + // attempts re-acquisition instead of being permanently blocked. + assert_eq!(read_lease_timestamp_secs(&json!({"unrelated": true})), 0); + assert_eq!(read_lease_timestamp_secs(&json!("not an int")), 0); + assert_eq!(read_lease_timestamp_secs(&Value::Null), 0); + } + + #[test] + fn lease_ttl_exceeds_summarizer_timeout() { + // The acquire_lease comment says LEASE_TTL_SECS must exceed the + // summariser LLM call timeout so a long call can't lose its lease + // mid-flight. This test pins that invariant. + let summarizer_timeout_secs = + (crate::summarize::SUMMARIZER_TIMEOUT_MS_FOR_TEST / 1000) as i64; + assert!( + LEASE_TTL_SECS > summarizer_timeout_secs, + "lease TTL ({LEASE_TTL_SECS}s) must exceed summarizer timeout ({summarizer_timeout_secs}s)" + ); + } +} diff --git a/context-compaction/src/main.rs b/context-compaction/src/main.rs new file mode 100644 index 00000000..61c026f4 --- /dev/null +++ b/context-compaction/src/main.rs @@ -0,0 +1,67 @@ +use std::sync::Arc; + +use anyhow::Result; +use clap::Parser; +use context_compaction::{manifest, register_with_iii}; +use iii_sdk::{register_worker, InitOptions, OtelConfig, WorkerMetadata}; + +#[derive(Parser, Debug)] +#[command( + name = "context-compaction", + about = "Out-of-band session-history compactor. Subscribes to agent::events::TurnEnd and writes a session-tree Compaction entry when the running token count crosses the configured threshold." +)] +struct Cli { + #[arg(long, env = "III_URL", default_value = "ws://127.0.0.1:49134")] + url: String, + + #[arg(long)] + manifest: bool, +} + +#[tokio::main] +async fn main() -> Result<()> { + tracing_subscriber::fmt() + .with_env_filter( + tracing_subscriber::EnvFilter::try_from_default_env() + .unwrap_or_else(|_| tracing_subscriber::EnvFilter::new("info")), + ) + .init(); + + let cli = Cli::parse(); + + if cli.manifest { + let m = manifest::build(); + println!("{}", serde_json::to_string_pretty(&m)?); + return Ok(()); + } + + let iii = register_worker( + &cli.url, + InitOptions { + otel: Some(OtelConfig::default()), + metadata: Some(WorkerMetadata { + runtime: "rust".to_string(), + version: env!("CARGO_PKG_VERSION").to_string(), + name: "context-compaction".to_string(), + os: std::env::consts::OS.to_string(), + pid: Some(std::process::id()), + telemetry: None, + ..WorkerMetadata::default() + }), + ..InitOptions::default() + }, + ); + let iii = Arc::new(iii); + + let _refs = register_with_iii(&iii)?; + tracing::info!( + threshold = context_compaction::config::trigger_tokens(), + keep_recent = context_compaction::config::keep_recent_turns(), + "context-compaction ready, subscribed to agent::events" + ); + + tokio::signal::ctrl_c().await?; + tracing::info!("context-compaction shutting down"); + iii.shutdown_async().await; + Ok(()) +} diff --git a/context-compaction/src/manifest.rs b/context-compaction/src/manifest.rs new file mode 100644 index 00000000..30eaf500 --- /dev/null +++ b/context-compaction/src/manifest.rs @@ -0,0 +1,49 @@ +//! Worker manifest. Emitted when `context-compaction --manifest` is invoked +//! so the harness manifest generator and registry can read worker metadata +//! without booting an iii engine. + +use serde::Serialize; + +pub const SKILL_ID: &str = "context-compaction"; +pub const SKILL_BODY: &str = include_str!("../skill.md"); + +#[derive(Debug, Serialize)] +pub struct Manifest { + pub name: &'static str, + pub version: &'static str, + pub description: &'static str, + pub functions: Vec<&'static str>, + pub subscriptions: Vec<&'static str>, + pub skill_id: &'static str, + pub skill: &'static str, +} + +pub fn build() -> Manifest { + Manifest { + name: "context-compaction", + version: env!("CARGO_PKG_VERSION"), + description: "Out-of-band session-history compactor (subscribes to agent::events::TurnEnd).", + functions: vec![], + subscriptions: vec!["agent::events"], + skill_id: SKILL_ID, + skill: SKILL_BODY, + } +} + +#[cfg(test)] +mod tests { + use super::*; + + #[test] + fn manifest_carries_no_llm_facing_functions() { + let m = build(); + assert!(m.functions.is_empty(), "compactor is not LLM-facing"); + assert_eq!(m.subscriptions, vec!["agent::events"]); + } + + #[test] + fn manifest_includes_skill_body() { + let m = build(); + assert!(m.skill.contains("compaction")); + } +} diff --git a/context-compaction/src/summarize.rs b/context-compaction/src/summarize.rs new file mode 100644 index 00000000..ac1e65e8 --- /dev/null +++ b/context-compaction/src/summarize.rs @@ -0,0 +1,298 @@ +//! Compaction execution: load active path, split keep-recent / older-prefix, +//! summarise via provider-router, append a Compaction entry to session-tree. + +use anyhow::{anyhow, Context}; +use iii_sdk::{TriggerRequest, Value, III}; +use serde_json::json; + +use crate::config; + +const COMPACTION_PROMPT: &str = include_str!("../prompts/compaction.txt"); +const SUMMARIZER_TIMEOUT_MS: u64 = 120_000; + +/// Test-only re-export so `lib::tests::lease_ttl_exceeds_summarizer_timeout` +/// can pin the LEASE_TTL_SECS > SUMMARIZER_TIMEOUT_MS/1000 invariant +/// without making the constant `pub` in non-test builds. +#[cfg(test)] +pub(crate) const SUMMARIZER_TIMEOUT_MS_FOR_TEST: u64 = SUMMARIZER_TIMEOUT_MS; + +/// Top-level entry point. Reads the active session-tree path, decides +/// what to summarise, calls the summariser LLM, then writes the +/// `Compaction` entry. Errors are surfaced to the caller for tracing +/// but never panic. +pub async fn summarize_and_append(iii: &III, session_id: &str) -> anyhow::Result<()> { + let messages = load_active_messages(iii, session_id).await?; + let keep = config::keep_recent_turns(); + let (older, _recent) = split_messages(&messages, keep); + if older.is_empty() { + // Nothing to summarise — the session is shorter than + // keep_recent_turns. Don't waste a summariser call. + return Ok(()); + } + let tokens_before = estimate_token_count(&older); + let summary = summarize_with_router(iii, session_id, &older).await?; + append_compaction(iii, session_id, summary, tokens_before).await?; + Ok(()) +} + +/// Walk the session-tree's active path and return the in-transcript +/// `AgentMessage`s (the session worker already filters Compaction entries +/// out, so the model would never have seen them either). +async fn load_active_messages(iii: &III, session_id: &str) -> anyhow::Result> { + let resp = iii + .trigger(TriggerRequest { + function_id: "session-tree::messages".into(), + payload: json!({ "session_id": session_id }), + action: None, + timeout_ms: Some(10_000), + }) + .await + .map_err(|e| anyhow!("session-tree::messages failed: {e}"))?; + let messages = resp + .get("messages") + .cloned() + .unwrap_or(resp) + .as_array() + .cloned() + .ok_or_else(|| anyhow!("session-tree::messages returned non-array"))?; + Ok(messages) +} + +/// Split the message list so the trailing `keep_recent` turns stay +/// verbatim and the older prefix becomes the summariser's input. A +/// "turn" here is one assistant message plus any leading user / tool +/// results; we approximate by counting messages from the end, which is +/// close enough for the threshold-driven trigger. +fn split_messages(messages: &[Value], keep_recent: usize) -> (Vec, Vec) { + if messages.len() <= keep_recent { + return (Vec::new(), messages.to_vec()); + } + let split_at = messages.len() - keep_recent; + ( + messages[..split_at].to_vec(), + messages[split_at..].to_vec(), + ) +} + +/// Rough token-count proxy. The summariser doesn't need byte-accurate +/// numbers for the `tokens_before` field — it's metadata for later +/// analysis, not a billing input — so a 4-chars-per-token heuristic is +/// fine. +fn estimate_token_count(messages: &[Value]) -> u64 { + let total_chars: usize = messages + .iter() + .map(|m| serde_json::to_string(m).map(|s| s.len()).unwrap_or(0)) + .sum(); + (total_chars / 4) as u64 +} + +async fn summarize_with_router( + iii: &III, + session_id: &str, + older: &[Value], +) -> anyhow::Result { + let payload = build_summarizer_payload(session_id, older); + let resp = iii + .trigger(TriggerRequest { + function_id: "router::stream_assistant".into(), + payload, + action: None, + timeout_ms: Some(SUMMARIZER_TIMEOUT_MS), + }) + .await + .map_err(|e| anyhow!("router::stream_assistant failed: {e}"))?; + + extract_summary_text(&resp).context("summariser response had no text content") +} + +/// Build the payload sent to `router::stream_assistant` for the +/// summarisation call. The provider/model are configurable so operators +/// can route to a cheap model without touching this code. +pub(crate) fn build_summarizer_payload(session_id: &str, older: &[Value]) -> Value { + let provider = config::summarizer_provider().unwrap_or_else(|| "anthropic".to_string()); + let model = config::summarizer_model().unwrap_or_else(|| "claude-haiku-4-5".to_string()); + json!({ + "session_id": format!("{session_id}::compaction"), + "provider": provider, + "model": model, + "system_prompt": COMPACTION_PROMPT, + "messages": [ + { + "role": "user", + "content": [ + { "type": "text", "text": render_user_prompt(older) } + ], + "timestamp": chrono::Utc::now().timestamp_millis() + } + ], + "tools": [] + }) +} + +fn render_user_prompt(older: &[Value]) -> String { + let mut out = String::new(); + out.push_str("Summarise the conversation below following the system-prompt structure exactly. Keep identifiers verbatim.\n\n"); + out.push_str("\n"); + for msg in older { + if let Some(role) = msg.get("role").and_then(Value::as_str) { + out.push_str(&format!("\n[{role}]\n")); + } + if let Some(content) = msg.get("content").and_then(Value::as_array) { + for block in content { + if let Some(text) = block.get("text").and_then(Value::as_str) { + out.push_str(text); + out.push('\n'); + } else if let Some(name) = block.get("name").and_then(Value::as_str) { + let args = block.get("input").cloned().unwrap_or_else(|| json!({})); + out.push_str(&format!( + "[tool_call] {name} {}\n", + serde_json::to_string(&args).unwrap_or_default() + )); + } + } + } else if let Some(text) = msg.get("content").and_then(Value::as_str) { + out.push_str(text); + out.push('\n'); + } + } + out.push_str("\n"); + out +} + +fn extract_summary_text(resp: &Value) -> Option { + // router::stream_assistant returns an AssistantMessage shape with + // `content: [ContentBlock]`. Concatenate text blocks. + let content = resp.get("content").and_then(Value::as_array)?; + let mut out = String::new(); + for block in content { + if let Some(t) = block.get("text").and_then(Value::as_str) { + if !out.is_empty() { + out.push('\n'); + } + out.push_str(t); + } + } + if out.is_empty() { + None + } else { + Some(out) + } +} + +async fn append_compaction( + iii: &III, + session_id: &str, + summary: String, + tokens_before: u64, +) -> anyhow::Result<()> { + iii.trigger(TriggerRequest { + function_id: "session-tree::compact".into(), + payload: json!({ + "session_id": session_id, + "summary": summary, + "tokens_before": tokens_before, + "file_ops": { "read_files": [], "modified_files": [] } + }), + action: None, + timeout_ms: Some(10_000), + }) + .await + .map_err(|e| anyhow!("session-tree::compact failed: {e}"))?; + + // Stamp the orchestrator-watched watermark so the next turn's + // `handle_streaming` reload check fires. + crate::stamp_last_compaction(iii, session_id).await; + Ok(()) +} + +#[cfg(test)] +mod tests { + use super::*; + + #[test] + fn split_messages_keeps_all_when_below_threshold() { + let msgs = vec![json!({"role": "user"}), json!({"role": "assistant"})]; + let (older, recent) = split_messages(&msgs, 3); + assert!(older.is_empty()); + assert_eq!(recent.len(), 2); + } + + #[test] + fn split_messages_splits_at_keep_boundary() { + let msgs: Vec = (0..10).map(|i| json!({"i": i})).collect(); + let (older, recent) = split_messages(&msgs, 3); + assert_eq!(older.len(), 7); + assert_eq!(recent.len(), 3); + assert_eq!(recent[0]["i"], 7); + assert_eq!(recent[2]["i"], 9); + } + + #[test] + fn render_user_prompt_includes_role_and_text() { + let msgs = vec![ + json!({"role": "user", "content": [{"type": "text", "text": "hello"}]}), + json!({"role": "assistant", "content": [{"type": "text", "text": "hi"}]}), + ]; + let out = render_user_prompt(&msgs); + assert!(out.contains("[user]")); + assert!(out.contains("[assistant]")); + assert!(out.contains("hello")); + assert!(out.contains("hi")); + assert!(out.contains("")); + assert!(out.contains("")); + } + + #[test] + fn render_user_prompt_handles_tool_calls() { + let msgs = vec![json!({ + "role": "assistant", + "content": [ + {"type": "tool_use", "name": "shell__run", "input": {"command": "ls"}} + ] + })]; + let out = render_user_prompt(&msgs); + assert!(out.contains("[tool_call] shell__run")); + assert!(out.contains("command")); + } + + #[test] + fn extract_summary_text_concatenates_text_blocks() { + let resp = json!({ + "content": [ + {"type": "text", "text": "# Goal\nbuild a thing"}, + {"type": "text", "text": "# Plan state\ndone"} + ] + }); + let s = extract_summary_text(&resp).expect("has text"); + assert!(s.contains("Goal")); + assert!(s.contains("Plan state")); + } + + #[test] + fn extract_summary_text_returns_none_when_empty() { + let resp = json!({ "content": [] }); + assert!(extract_summary_text(&resp).is_none()); + } + + #[test] + fn estimate_token_count_scales_with_message_size() { + let small = vec![json!({"text": "x"})]; + let big = vec![json!({"text": "x".repeat(4000)})]; + assert!(estimate_token_count(&big) > estimate_token_count(&small)); + } + + #[test] + fn build_summarizer_payload_uses_configured_or_default_model() { + let payload = build_summarizer_payload("sess-1", &[json!({"role": "user", "content": []})]); + assert!(payload["session_id"] + .as_str() + .unwrap() + .contains("compaction")); + assert!(payload["model"].is_string()); + assert!(payload["provider"].is_string()); + assert!(payload["system_prompt"] + .as_str() + .unwrap() + .contains("anchored context summarization")); + } +} diff --git a/context-compaction/src/threshold.rs b/context-compaction/src/threshold.rs new file mode 100644 index 00000000..3698a58e --- /dev/null +++ b/context-compaction/src/threshold.rs @@ -0,0 +1,26 @@ +//! Trigger policy for compaction. Single function so unit tests and the +//! lib subscriber agree on the decision boundary. + +use crate::config::trigger_tokens; + +/// Returns `true` when the most recent turn's transcript size has crossed +/// the configured token threshold and compaction should fire. +pub fn should_compact(total_tokens: u64) -> bool { + total_tokens >= trigger_tokens() +} + +#[cfg(test)] +mod tests { + use super::*; + + #[test] + fn zero_tokens_never_compacts() { + assert!(!should_compact(0)); + } + + #[test] + fn huge_token_count_always_compacts() { + // 10× the highest sane default — must always cross. + assert!(should_compact(10_000_000)); + } +} diff --git a/context-compaction/tests/manifest.rs b/context-compaction/tests/manifest.rs new file mode 100644 index 00000000..ac80ecc2 --- /dev/null +++ b/context-compaction/tests/manifest.rs @@ -0,0 +1,79 @@ +//! Smoke test for the `--manifest` CLI flag. Mirrors `policy-denylist`'s +//! approach: spawn the built binary, parse stdout as JSON, assert the +//! shape matches what the harness manifest generator and registry consume. + +use std::process::Command; + +use serde_json::Value; + +fn binary_path() -> String { + // Prefer the Cargo-injected path when available (standard cargo test + // builds). Fall back to constructing it from CARGO_MANIFEST_DIR so the + // test still works in environments where CARGO_BIN_EXE_* is not + // injected (e.g. some sandboxed CI runners). + if let Some(path) = option_env!("CARGO_BIN_EXE_context-compaction") { + return path.to_string(); + } + let manifest_dir = env!("CARGO_MANIFEST_DIR"); + format!("{manifest_dir}/target/debug/context-compaction") +} + +#[test] +fn manifest_subcommand_emits_valid_json() { + let bin = binary_path(); + let output = Command::new(&bin) + .arg("--manifest") + .output() + .unwrap_or_else(|e| panic!("spawn {bin}: {e}")); + + assert!( + output.status.success(), + "binary exited with {:?}; stderr: {}", + output.status, + String::from_utf8_lossy(&output.stderr), + ); + + let stdout = String::from_utf8(output.stdout).expect("manifest stdout is utf-8"); + let manifest: Value = serde_json::from_str(&stdout).expect("manifest stdout is valid JSON"); + + assert_eq!(manifest["name"], "context-compaction"); + assert_eq!(manifest["version"], env!("CARGO_PKG_VERSION")); + assert!( + manifest["description"] + .as_str() + .is_some_and(|s| !s.is_empty()), + "description must be non-empty" + ); + + // The compactor is not LLM-facing: it must not register any callable + // function. Regressing this would expose `result::fetch`-style surface + // the model could discover via `engine::functions::list`. + let functions = manifest["functions"] + .as_array() + .expect("functions must be an array"); + assert!( + functions.is_empty(), + "compactor must not register LLM-facing functions, got {functions:?}" + ); + + // The whole worker's purpose is the agent::events subscription; if this + // ever drops out of the manifest the harness wiring stops registering + // the trigger and compaction silently never fires. + let subscriptions = manifest["subscriptions"] + .as_array() + .expect("subscriptions must be an array"); + assert!( + subscriptions + .iter() + .any(|s| s.as_str() == Some("agent::events")), + "subscriptions must include agent::events, got {subscriptions:?}" + ); + + assert_eq!(manifest["skill_id"], "context-compaction"); + assert!( + manifest["skill"] + .as_str() + .is_some_and(|s| !s.is_empty()), + "skill body must be embedded" + ); +} diff --git a/harness/Makefile b/harness/Makefile index fd91bccb..44cfb9be 100644 --- a/harness/Makefile +++ b/harness/Makefile @@ -45,17 +45,17 @@ CONFIG_FILE := $(HARNESS_DIR)/config.yaml # Local worker crates that need to be cargo-built and symlinked into # ~/.iii/workers/ so the engine spawns them on host instead of trying # `cargo run` inside libkrun (where cargo isn't installed). -LOCAL_WORKERS := approval-gate auth-credentials hook-fanout iii-directory \ - llm-budget models-catalog policy-denylist provider-anthropic \ - provider-openai provider-router session shell turn-orchestrator \ - harness +LOCAL_WORKERS := approval-gate auth-credentials context-compaction hook-fanout \ + iii-directory llm-budget models-catalog policy-denylist \ + provider-anthropic provider-openai provider-router session \ + shell turn-orchestrator harness -.PHONY: help all build config observability engine verify web stop restart logs clean ensure-dirs +.PHONY: help all build config observability engine compaction verify web stop restart logs clean ensure-dirs help: @awk '/^[^#]/ && !/^$$/ {exit} /^#/ {sub(/^# ?/, ""); print}' $(firstword $(MAKEFILE_LIST)) -all: build config observability engine verify +all: build config observability engine compaction verify ensure-dirs: @mkdir -p $(PIDS_DIR) $(LOGS_DIR) $(HARNESS_DIR)/data/skills $(WORKERS_BIN) @@ -102,6 +102,35 @@ observability: config @echo "==> iii worker add iii-observability (idempotent)" @cd "$(HARNESS_DIR)" && iii worker add iii-observability --no-wait +# ─── compaction ────────────────────────────────────────────────────────────── + +# Spawn the `context-compaction` worker as a background process. Kept OUT of +# `iii.worker.yaml` `dependencies:` because the upstream registry doesn't yet +# know about it; that means `iii worker add` can't resolve it and it stays +# absent from `EXPECTED_WORKERS` / `harness::status`. Mirrors the same +# "optional, side-of-config" pattern used by `iii-observability`. The worker +# is its own iii-sdk client — it connects to the engine over WS on its own +# once started. +# +# Requires the engine to be up first. The PID file lives under $(PIDS_DIR) +# alongside `engine.pid` / `web.pid` so `make stop` cleans it up uniformly. +COMPACTION_BIN := $(WORKERS_BIN)/context-compaction +compaction: ensure-dirs + @[[ -x "$(COMPACTION_BIN)" ]] || { echo "==> $(COMPACTION_BIN) missing; run 'make build' first"; exit 1; } + @if [[ -f "$(PIDS_DIR)/context-compaction.pid" ]] && kill -0 "$$(cat $(PIDS_DIR)/context-compaction.pid)" 2>/dev/null; then \ + echo "==> context-compaction already running (pid $$(cat $(PIDS_DIR)/context-compaction.pid))"; \ + exit 0; \ + fi + @echo "==> starting context-compaction worker (engine=$(DEMO_ENGINE_WS))" + @{ nohup "$(COMPACTION_BIN)" --url "$(DEMO_ENGINE_WS)" > "$(LOGS_DIR)/context-compaction.log" 2>&1 & echo $$! > "$(PIDS_DIR)/context-compaction.pid"; } + @sleep 1 + @if kill -0 "$$(cat $(PIDS_DIR)/context-compaction.pid)" 2>/dev/null; then \ + echo " pid=$$(cat $(PIDS_DIR)/context-compaction.pid) logs: make logs W=context-compaction"; \ + else \ + echo " [error] context-compaction exited immediately; tail $(LOGS_DIR)/context-compaction.log"; \ + exit 1; \ + fi + # ─── engine ────────────────────────────────────────────────────────────────── engine: ensure-dirs config @@ -154,7 +183,7 @@ web: ensure-dirs W ?= engine logs: @case "$(W)" in \ - engine|web) exec tail -f "$(LOGS_DIR)/$(W).log" ;; \ + engine|web|context-compaction) exec tail -f "$(LOGS_DIR)/$(W).log" ;; \ *) echo "==> tailing engine.log (filter for '$(W)')"; exec tail -f "$(LOGS_DIR)/engine.log" ;; \ esac diff --git a/harness/crates/harness-types/src/function.rs b/harness/crates/harness-types/src/function.rs index 5620a04e..197e426d 100644 --- a/harness/crates/harness-types/src/function.rs +++ b/harness/crates/harness-types/src/function.rs @@ -63,6 +63,23 @@ pub struct FunctionResult { pub details: serde_json::Value, #[serde(default)] pub terminate: bool, + /// Set when `content` has been truncated for context-budget reasons. + /// The orchestrator stashes the full payload under + /// `session//result/` and the model can retrieve + /// it via `agent_call(function="result::fetch", payload={call_id})`. + /// Absent when the result was kept verbatim. + #[serde(default, skip_serializing_if = "Option::is_none")] + pub truncated: Option, +} + +/// Metadata attached to a [`FunctionResult`] whose `content` has been +/// truncated. Lets the model decide whether to fetch the full payload. +#[derive(Debug, Clone, PartialEq, Serialize, Deserialize)] +pub struct TruncationInfo { + /// Original serialized size of the un-truncated `{content, details}`. + pub original_bytes: u64, + /// Side-key suffix; full payload at `session//result/`. + pub call_id: String, } /// Outcome of prepare. Either ready to execute, or short-circuited. diff --git a/harness/crates/harness-types/src/lib.rs b/harness/crates/harness-types/src/lib.rs index 904fccbb..7a354ec5 100644 --- a/harness/crates/harness-types/src/lib.rs +++ b/harness/crates/harness-types/src/lib.rs @@ -18,7 +18,7 @@ pub use agent_message::{ pub use content::{ContentBlock, ImageContent, TextContent}; pub use function::{ AgentFunction, CacheRetention, ExecutionMode, FinalizedFunctionCall, FunctionCall, - FunctionResult, PreparedFunctionCall, Transport, + FunctionResult, PreparedFunctionCall, Transport, TruncationInfo, }; pub use stream_event::{AssistantMessageEvent, ErrorKind, StopReason, Usage}; pub use thinking::{TextPhase, TextSignature, ThinkingBudgets, ThinkingLevel}; diff --git a/provider-anthropic/crates/harness-types/src/function.rs b/provider-anthropic/crates/harness-types/src/function.rs index 5620a04e..197e426d 100644 --- a/provider-anthropic/crates/harness-types/src/function.rs +++ b/provider-anthropic/crates/harness-types/src/function.rs @@ -63,6 +63,23 @@ pub struct FunctionResult { pub details: serde_json::Value, #[serde(default)] pub terminate: bool, + /// Set when `content` has been truncated for context-budget reasons. + /// The orchestrator stashes the full payload under + /// `session//result/` and the model can retrieve + /// it via `agent_call(function="result::fetch", payload={call_id})`. + /// Absent when the result was kept verbatim. + #[serde(default, skip_serializing_if = "Option::is_none")] + pub truncated: Option, +} + +/// Metadata attached to a [`FunctionResult`] whose `content` has been +/// truncated. Lets the model decide whether to fetch the full payload. +#[derive(Debug, Clone, PartialEq, Serialize, Deserialize)] +pub struct TruncationInfo { + /// Original serialized size of the un-truncated `{content, details}`. + pub original_bytes: u64, + /// Side-key suffix; full payload at `session//result/`. + pub call_id: String, } /// Outcome of prepare. Either ready to execute, or short-circuited. diff --git a/provider-anthropic/crates/harness-types/src/lib.rs b/provider-anthropic/crates/harness-types/src/lib.rs index 904fccbb..7a354ec5 100644 --- a/provider-anthropic/crates/harness-types/src/lib.rs +++ b/provider-anthropic/crates/harness-types/src/lib.rs @@ -18,7 +18,7 @@ pub use agent_message::{ pub use content::{ContentBlock, ImageContent, TextContent}; pub use function::{ AgentFunction, CacheRetention, ExecutionMode, FinalizedFunctionCall, FunctionCall, - FunctionResult, PreparedFunctionCall, Transport, + FunctionResult, PreparedFunctionCall, Transport, TruncationInfo, }; pub use stream_event::{AssistantMessageEvent, ErrorKind, StopReason, Usage}; pub use thinking::{TextPhase, TextSignature, ThinkingBudgets, ThinkingLevel}; diff --git a/provider-anthropic/crates/provider-base/src/openai_compat.rs b/provider-anthropic/crates/provider-base/src/openai_compat.rs index 3bc514f5..3dec4f12 100644 --- a/provider-anthropic/crates/provider-base/src/openai_compat.rs +++ b/provider-anthropic/crates/provider-base/src/openai_compat.rs @@ -446,6 +446,21 @@ fn merge_usage(usage: &serde_json::Value, into: &mut Usage) { { into.output += v; } + // OpenAI auto-caches prefixes >1024 tokens with no opt-in. The cached + // count surfaces nested under either `prompt_tokens_details.cached_tokens` + // (Chat Completions API) or `input_tokens_details.cached_tokens` + // (Responses API). Either path discounts the cached portion at 50% of + // input price; surfacing the count lets downstream cost calculation + // (e.g. llm-budget) reflect that. + for parent in ["prompt_tokens_details", "input_tokens_details"] { + if let Some(v) = usage + .get(parent) + .and_then(|d| d.get("cached_tokens")) + .and_then(serde_json::Value::as_u64) + { + into.cache_read += v; + } + } } fn map_finish_reason(reason: &str) -> StopReason { @@ -504,3 +519,52 @@ fn build_final(state: &PartialState, model: &str, provider: &str) -> AssistantMe fn _kept(_: harness_types::FunctionCall, _: fn(&str, Option) -> harness_types::ErrorKind) { let _ = classify_provider_error; } + +#[cfg(test)] +mod tests { + use super::*; + use harness_types::Usage; + + #[test] + fn merge_usage_extracts_chat_completions_cached_tokens() { + let mut u = Usage::default(); + merge_usage( + &serde_json::json!({ + "prompt_tokens": 1500, + "completion_tokens": 200, + "prompt_tokens_details": { "cached_tokens": 1200 } + }), + &mut u, + ); + assert_eq!(u.input, 1500); + assert_eq!(u.output, 200); + assert_eq!( + u.cache_read, 1200, + "Chat Completions cached_tokens must flow into Usage.cache_read" + ); + } + + #[test] + fn merge_usage_extracts_responses_api_cached_tokens() { + let mut u = Usage::default(); + merge_usage( + &serde_json::json!({ + "input_tokens": 2000, + "output_tokens": 100, + "input_tokens_details": { "cached_tokens": 1700 } + }), + &mut u, + ); + assert_eq!(u.cache_read, 1700); + } + + #[test] + fn merge_usage_no_cache_field_leaves_cache_read_zero() { + let mut u = Usage::default(); + merge_usage( + &serde_json::json!({ "prompt_tokens": 50, "completion_tokens": 25 }), + &mut u, + ); + assert_eq!(u.cache_read, 0); + } +} diff --git a/provider-anthropic/src/lib.rs b/provider-anthropic/src/lib.rs index e362e431..56354a7a 100644 --- a/provider-anthropic/src/lib.rs +++ b/provider-anthropic/src/lib.rs @@ -220,6 +220,141 @@ pub fn functions_to_wire(tools: &[harness_types::AgentFunction]) -> Vec bool { + static CACHED: std::sync::OnceLock = std::sync::OnceLock::new(); + *CACHED.get_or_init(|| match std::env::var(CACHE_FLAG_ENV) { + Ok(v) => !matches!(v.as_str(), "0" | "false" | "FALSE" | "False"), + Err(_) => true, + }) +} + +fn ephemeral_marker() -> serde_json::Value { + serde_json::json!({ "type": "ephemeral" }) +} + +/// Build the `system` wire field. Emits the typed-block array form with a +/// `cache_control: ephemeral` marker when caching is enabled and the prompt +/// is long enough to be cache-eligible. Otherwise emits the plain-string +/// form (which Anthropic also accepts) so short system prompts don't trigger +/// HTTP 400 on too-small cacheable spans. +pub fn build_system_field(system_prompt: &str) -> serde_json::Value { + if cache_enabled() && system_prompt.len() >= CACHE_MIN_CHARS { + serde_json::json!([{ + "type": "text", + "text": system_prompt, + "cache_control": ephemeral_marker(), + }]) + } else { + serde_json::Value::String(system_prompt.to_string()) + } +} + +/// Attach a `cache_control: ephemeral` marker to the last entry of the +/// `tools` array. Anthropic caches the entire prefix up to the marker, so +/// one marker on the last tool caches the whole tools array as a unit. +/// No-op when caching is disabled, the array is empty, or the serialized +/// size of the array falls below the cache-eligibility floor. +pub fn apply_tools_cache_control(tools: &mut [serde_json::Value]) { + if !cache_enabled() || tools.is_empty() { + return; + } + let serialized_size: usize = tools + .iter() + .map(|v| serde_json::to_string(v).map(|s| s.len()).unwrap_or(0)) + .sum(); + if serialized_size < CACHE_MIN_CHARS { + return; + } + if let Some(last) = tools.last_mut() { + if let Some(obj) = last.as_object_mut() { + obj.insert("cache_control".into(), ephemeral_marker()); + } + } +} + +/// Stamp a `cache_control: ephemeral` marker on the last content block of +/// the most recent "stable" assistant turn — i.e. one whose `tool_use` +/// blocks all have matching downstream `tool_result` blocks. Marking an +/// unstable turn (in-flight tool calls) would cache a transient state and +/// invalidate on the next turn, defeating the point. +pub fn apply_messages_cache_anchor(wire: &mut [serde_json::Value]) { + if !cache_enabled() || wire.is_empty() { + return; + } + let last_stable = (0..wire.len()) + .rev() + .find(|&idx| is_stable_assistant(wire, idx)); + let Some(idx) = last_stable else { return }; + let Some(content) = wire[idx].get_mut("content").and_then(|c| c.as_array_mut()) else { + return; + }; + if let Some(last_block) = content.last_mut() { + if let Some(obj) = last_block.as_object_mut() { + obj.insert("cache_control".into(), ephemeral_marker()); + } + } +} + +fn is_stable_assistant(wire: &[serde_json::Value], idx: usize) -> bool { + let msg = &wire[idx]; + if msg.get("role").and_then(|r| r.as_str()) != Some("assistant") { + return false; + } + let tool_use_ids: Vec<&str> = msg + .get("content") + .and_then(|c| c.as_array()) + .map(|arr| { + arr.iter() + .filter_map(|b| { + if b.get("type").and_then(|t| t.as_str()) == Some("tool_use") { + b.get("id").and_then(|i| i.as_str()) + } else { + None + } + }) + .collect() + }) + .unwrap_or_default(); + if tool_use_ids.is_empty() { + return true; + } + tool_use_ids + .iter() + .all(|id| has_downstream_tool_result(&wire[idx + 1..], id)) +} + +fn has_downstream_tool_result(later: &[serde_json::Value], id: &str) -> bool { + later.iter().any(|m| { + m.get("role").and_then(|r| r.as_str()) == Some("user") + && m.get("content") + .and_then(|c| c.as_array()) + .map(|blocks| { + blocks.iter().any(|b| { + b.get("type").and_then(|t| t.as_str()) == Some("tool_result") + && b.get("tool_use_id").and_then(|i| i.as_str()) == Some(id) + }) + }) + .unwrap_or(false) + }) +} + /// Stream a response from Anthropic. Returns an event stream that closes with /// `done` on success or `error` on failure. Never throws. pub async fn stream( @@ -296,12 +431,16 @@ async fn stream_inner( "anthropic.request.build", Some(iii_sdk::SpanKind::Internal), || async { + let mut wire_messages = to_wire_messages(&messages); + apply_messages_cache_anchor(&mut wire_messages); + let mut wire_tools = functions_to_wire(&tools); + apply_tools_cache_control(&mut wire_tools); let body = serde_json::json!({ "model": cfg.model, "max_tokens": cfg.max_tokens, - "system": system_prompt, - "messages": to_wire_messages(&messages), - "tools": functions_to_wire(&tools), + "system": build_system_field(&system_prompt), + "messages": wire_messages, + "tools": wire_tools, "stream": true, }); let client = reqwest::Client::builder() @@ -774,6 +913,151 @@ mod tests { assert_eq!(u.output, 26); } + #[test] + fn merge_usage_captures_cache_fields() { + let mut u = Usage::default(); + merge_usage( + &serde_json::json!({ + "input_tokens": 100, + "output_tokens": 50, + "cache_read_input_tokens": 80, + "cache_creation_input_tokens": 20, + }), + &mut u, + ); + assert_eq!(u.cache_read, 80); + assert_eq!(u.cache_write, 20); + } + + #[test] + fn build_system_field_short_returns_string() { + let out = build_system_field("hi"); + assert!(out.is_string()); + assert_eq!(out.as_str(), Some("hi")); + } + + #[test] + fn build_system_field_long_returns_typed_block_with_cache_marker() { + let long = "x".repeat(CACHE_MIN_CHARS); + let out = build_system_field(&long); + let arr = out.as_array().expect("typed-block array"); + assert_eq!(arr.len(), 1); + assert_eq!(arr[0]["type"], "text"); + assert_eq!(arr[0]["text"].as_str().unwrap().len(), CACHE_MIN_CHARS); + assert_eq!(arr[0]["cache_control"]["type"], "ephemeral"); + } + + #[test] + fn apply_tools_cache_control_skips_empty() { + let mut tools: Vec = vec![]; + apply_tools_cache_control(&mut tools); + assert!(tools.is_empty()); + } + + #[test] + fn apply_tools_cache_control_skips_small_arrays() { + // A single tiny tool entry — far below the 4 KB floor. + let mut tools = vec![serde_json::json!({ + "name": "agent_call", + "description": "noop", + "input_schema": {"type": "object"}, + })]; + apply_tools_cache_control(&mut tools); + assert!( + tools[0].get("cache_control").is_none(), + "tiny tools array must not be marked (would 400 on Anthropic)" + ); + } + + #[test] + fn apply_tools_cache_control_marks_last_when_eligible() { + // Pad description so the serialized array exceeds CACHE_MIN_CHARS. + let bulky = "x".repeat(CACHE_MIN_CHARS); + let mut tools = vec![ + serde_json::json!({"name": "a", "description": "small", "input_schema": {}}), + serde_json::json!({"name": "b", "description": bulky, "input_schema": {}}), + ]; + apply_tools_cache_control(&mut tools); + assert!( + tools[0].get("cache_control").is_none(), + "only the last entry is marked" + ); + assert_eq!(tools[1]["cache_control"]["type"], "ephemeral"); + } + + #[test] + fn apply_messages_cache_anchor_marks_last_stable_assistant() { + let mut wire = vec![ + serde_json::json!({"role": "user", "content": [{"type": "text", "text": "hi"}]}), + serde_json::json!({ + "role": "assistant", + "content": [{"type": "text", "text": "first reply"}] + }), + serde_json::json!({"role": "user", "content": [{"type": "text", "text": "more"}]}), + serde_json::json!({ + "role": "assistant", + "content": [{"type": "text", "text": "second reply"}] + }), + ]; + apply_messages_cache_anchor(&mut wire); + assert!( + wire[1]["content"][0].get("cache_control").is_none(), + "earlier assistant should not be marked" + ); + assert_eq!(wire[3]["content"][0]["cache_control"]["type"], "ephemeral"); + } + + #[test] + fn apply_messages_cache_anchor_skips_assistant_with_unresolved_tool_use() { + let mut wire = vec![ + serde_json::json!({"role": "user", "content": [{"type": "text", "text": "hi"}]}), + serde_json::json!({ + "role": "assistant", + "content": [ + {"type": "text", "text": "running"}, + {"type": "tool_use", "id": "tc1", "name": "shell__run", "input": {}} + ] + }), + // No matching tool_result follows — assistant is "in-flight". + ]; + apply_messages_cache_anchor(&mut wire); + for block in wire[1]["content"].as_array().unwrap() { + assert!( + block.get("cache_control").is_none(), + "in-flight assistant must not be marked" + ); + } + } + + #[test] + fn apply_messages_cache_anchor_marks_assistant_when_tool_result_follows() { + let mut wire = vec![ + serde_json::json!({ + "role": "assistant", + "content": [ + {"type": "text", "text": "running"}, + {"type": "tool_use", "id": "tc1", "name": "shell__run", "input": {}} + ] + }), + serde_json::json!({ + "role": "user", + "content": [{"type": "tool_result", "tool_use_id": "tc1", "content": "ok"}] + }), + ]; + apply_messages_cache_anchor(&mut wire); + // The tool_use is the last content block of the assistant; the + // marker lands there. + let tool_use = &wire[0]["content"][1]; + assert_eq!(tool_use["cache_control"]["type"], "ephemeral"); + } + + #[test] + fn apply_messages_cache_anchor_noop_on_empty() { + let mut wire: Vec = vec![]; + apply_messages_cache_anchor(&mut wire); + assert!(wire.is_empty()); + } + #[test] fn with_credential_api_key() { let cred = auth_credentials::Credential::ApiKey { diff --git a/provider-openai/crates/harness-types/src/function.rs b/provider-openai/crates/harness-types/src/function.rs index 5620a04e..197e426d 100644 --- a/provider-openai/crates/harness-types/src/function.rs +++ b/provider-openai/crates/harness-types/src/function.rs @@ -63,6 +63,23 @@ pub struct FunctionResult { pub details: serde_json::Value, #[serde(default)] pub terminate: bool, + /// Set when `content` has been truncated for context-budget reasons. + /// The orchestrator stashes the full payload under + /// `session//result/` and the model can retrieve + /// it via `agent_call(function="result::fetch", payload={call_id})`. + /// Absent when the result was kept verbatim. + #[serde(default, skip_serializing_if = "Option::is_none")] + pub truncated: Option, +} + +/// Metadata attached to a [`FunctionResult`] whose `content` has been +/// truncated. Lets the model decide whether to fetch the full payload. +#[derive(Debug, Clone, PartialEq, Serialize, Deserialize)] +pub struct TruncationInfo { + /// Original serialized size of the un-truncated `{content, details}`. + pub original_bytes: u64, + /// Side-key suffix; full payload at `session//result/`. + pub call_id: String, } /// Outcome of prepare. Either ready to execute, or short-circuited. diff --git a/provider-openai/crates/harness-types/src/lib.rs b/provider-openai/crates/harness-types/src/lib.rs index 904fccbb..7a354ec5 100644 --- a/provider-openai/crates/harness-types/src/lib.rs +++ b/provider-openai/crates/harness-types/src/lib.rs @@ -18,7 +18,7 @@ pub use agent_message::{ pub use content::{ContentBlock, ImageContent, TextContent}; pub use function::{ AgentFunction, CacheRetention, ExecutionMode, FinalizedFunctionCall, FunctionCall, - FunctionResult, PreparedFunctionCall, Transport, + FunctionResult, PreparedFunctionCall, Transport, TruncationInfo, }; pub use stream_event::{AssistantMessageEvent, ErrorKind, StopReason, Usage}; pub use thinking::{TextPhase, TextSignature, ThinkingBudgets, ThinkingLevel}; diff --git a/provider-openai/crates/provider-base/src/openai_compat.rs b/provider-openai/crates/provider-base/src/openai_compat.rs index 6648ff5e..fe09149b 100644 --- a/provider-openai/crates/provider-base/src/openai_compat.rs +++ b/provider-openai/crates/provider-base/src/openai_compat.rs @@ -446,6 +446,21 @@ fn merge_usage(usage: &serde_json::Value, into: &mut Usage) { { into.output += v; } + // OpenAI auto-caches prefixes >1024 tokens with no opt-in. The cached + // count surfaces nested under either `prompt_tokens_details.cached_tokens` + // (Chat Completions API) or `input_tokens_details.cached_tokens` + // (Responses API). Either path discounts the cached portion at 50% of + // input price; surfacing the count lets downstream cost calculation + // (e.g. llm-budget) reflect that. + for parent in ["prompt_tokens_details", "input_tokens_details"] { + if let Some(v) = usage + .get(parent) + .and_then(|d| d.get("cached_tokens")) + .and_then(serde_json::Value::as_u64) + { + into.cache_read += v; + } + } } fn map_finish_reason(reason: &str) -> StopReason { @@ -504,3 +519,52 @@ fn build_final(state: &PartialState, model: &str, provider: &str) -> AssistantMe fn _kept(_: harness_types::FunctionCall, _: fn(&str, Option) -> harness_types::ErrorKind) { let _ = classify_provider_error; } + +#[cfg(test)] +mod tests { + use super::*; + use harness_types::Usage; + + #[test] + fn merge_usage_extracts_chat_completions_cached_tokens() { + let mut u = Usage::default(); + merge_usage( + &serde_json::json!({ + "prompt_tokens": 1500, + "completion_tokens": 200, + "prompt_tokens_details": { "cached_tokens": 1200 } + }), + &mut u, + ); + assert_eq!(u.input, 1500); + assert_eq!(u.output, 200); + assert_eq!( + u.cache_read, 1200, + "Chat Completions cached_tokens must flow into Usage.cache_read" + ); + } + + #[test] + fn merge_usage_extracts_responses_api_cached_tokens() { + let mut u = Usage::default(); + merge_usage( + &serde_json::json!({ + "input_tokens": 2000, + "output_tokens": 100, + "input_tokens_details": { "cached_tokens": 1700 } + }), + &mut u, + ); + assert_eq!(u.cache_read, 1700); + } + + #[test] + fn merge_usage_no_cache_field_leaves_cache_read_zero() { + let mut u = Usage::default(); + merge_usage( + &serde_json::json!({ "prompt_tokens": 50, "completion_tokens": 25 }), + &mut u, + ); + assert_eq!(u.cache_read, 0); + } +} diff --git a/provider-router/crates/harness-types/src/function.rs b/provider-router/crates/harness-types/src/function.rs index 5620a04e..197e426d 100644 --- a/provider-router/crates/harness-types/src/function.rs +++ b/provider-router/crates/harness-types/src/function.rs @@ -63,6 +63,23 @@ pub struct FunctionResult { pub details: serde_json::Value, #[serde(default)] pub terminate: bool, + /// Set when `content` has been truncated for context-budget reasons. + /// The orchestrator stashes the full payload under + /// `session//result/` and the model can retrieve + /// it via `agent_call(function="result::fetch", payload={call_id})`. + /// Absent when the result was kept verbatim. + #[serde(default, skip_serializing_if = "Option::is_none")] + pub truncated: Option, +} + +/// Metadata attached to a [`FunctionResult`] whose `content` has been +/// truncated. Lets the model decide whether to fetch the full payload. +#[derive(Debug, Clone, PartialEq, Serialize, Deserialize)] +pub struct TruncationInfo { + /// Original serialized size of the un-truncated `{content, details}`. + pub original_bytes: u64, + /// Side-key suffix; full payload at `session//result/`. + pub call_id: String, } /// Outcome of prepare. Either ready to execute, or short-circuited. diff --git a/provider-router/crates/harness-types/src/lib.rs b/provider-router/crates/harness-types/src/lib.rs index 904fccbb..7a354ec5 100644 --- a/provider-router/crates/harness-types/src/lib.rs +++ b/provider-router/crates/harness-types/src/lib.rs @@ -18,7 +18,7 @@ pub use agent_message::{ pub use content::{ContentBlock, ImageContent, TextContent}; pub use function::{ AgentFunction, CacheRetention, ExecutionMode, FinalizedFunctionCall, FunctionCall, - FunctionResult, PreparedFunctionCall, Transport, + FunctionResult, PreparedFunctionCall, Transport, TruncationInfo, }; pub use stream_event::{AssistantMessageEvent, ErrorKind, StopReason, Usage}; pub use thinking::{TextPhase, TextSignature, ThinkingBudgets, ThinkingLevel}; diff --git a/session/crates/harness-types/src/function.rs b/session/crates/harness-types/src/function.rs index 5620a04e..197e426d 100644 --- a/session/crates/harness-types/src/function.rs +++ b/session/crates/harness-types/src/function.rs @@ -63,6 +63,23 @@ pub struct FunctionResult { pub details: serde_json::Value, #[serde(default)] pub terminate: bool, + /// Set when `content` has been truncated for context-budget reasons. + /// The orchestrator stashes the full payload under + /// `session//result/` and the model can retrieve + /// it via `agent_call(function="result::fetch", payload={call_id})`. + /// Absent when the result was kept verbatim. + #[serde(default, skip_serializing_if = "Option::is_none")] + pub truncated: Option, +} + +/// Metadata attached to a [`FunctionResult`] whose `content` has been +/// truncated. Lets the model decide whether to fetch the full payload. +#[derive(Debug, Clone, PartialEq, Serialize, Deserialize)] +pub struct TruncationInfo { + /// Original serialized size of the un-truncated `{content, details}`. + pub original_bytes: u64, + /// Side-key suffix; full payload at `session//result/`. + pub call_id: String, } /// Outcome of prepare. Either ready to execute, or short-circuited. diff --git a/session/crates/harness-types/src/lib.rs b/session/crates/harness-types/src/lib.rs index 904fccbb..7a354ec5 100644 --- a/session/crates/harness-types/src/lib.rs +++ b/session/crates/harness-types/src/lib.rs @@ -18,7 +18,7 @@ pub use agent_message::{ pub use content::{ContentBlock, ImageContent, TextContent}; pub use function::{ AgentFunction, CacheRetention, ExecutionMode, FinalizedFunctionCall, FunctionCall, - FunctionResult, PreparedFunctionCall, Transport, + FunctionResult, PreparedFunctionCall, Transport, TruncationInfo, }; pub use stream_event::{AssistantMessageEvent, ErrorKind, StopReason, Usage}; pub use thinking::{TextPhase, TextSignature, ThinkingBudgets, ThinkingLevel}; diff --git a/turn-orchestrator/crates/harness-types/src/function.rs b/turn-orchestrator/crates/harness-types/src/function.rs index 5620a04e..197e426d 100644 --- a/turn-orchestrator/crates/harness-types/src/function.rs +++ b/turn-orchestrator/crates/harness-types/src/function.rs @@ -63,6 +63,23 @@ pub struct FunctionResult { pub details: serde_json::Value, #[serde(default)] pub terminate: bool, + /// Set when `content` has been truncated for context-budget reasons. + /// The orchestrator stashes the full payload under + /// `session//result/` and the model can retrieve + /// it via `agent_call(function="result::fetch", payload={call_id})`. + /// Absent when the result was kept verbatim. + #[serde(default, skip_serializing_if = "Option::is_none")] + pub truncated: Option, +} + +/// Metadata attached to a [`FunctionResult`] whose `content` has been +/// truncated. Lets the model decide whether to fetch the full payload. +#[derive(Debug, Clone, PartialEq, Serialize, Deserialize)] +pub struct TruncationInfo { + /// Original serialized size of the un-truncated `{content, details}`. + pub original_bytes: u64, + /// Side-key suffix; full payload at `session//result/`. + pub call_id: String, } /// Outcome of prepare. Either ready to execute, or short-circuited. diff --git a/turn-orchestrator/crates/harness-types/src/lib.rs b/turn-orchestrator/crates/harness-types/src/lib.rs index 904fccbb..7a354ec5 100644 --- a/turn-orchestrator/crates/harness-types/src/lib.rs +++ b/turn-orchestrator/crates/harness-types/src/lib.rs @@ -18,7 +18,7 @@ pub use agent_message::{ pub use content::{ContentBlock, ImageContent, TextContent}; pub use function::{ AgentFunction, CacheRetention, ExecutionMode, FinalizedFunctionCall, FunctionCall, - FunctionResult, PreparedFunctionCall, Transport, + FunctionResult, PreparedFunctionCall, Transport, TruncationInfo, }; pub use stream_event::{AssistantMessageEvent, ErrorKind, StopReason, Usage}; pub use thinking::{TextPhase, TextSignature, ThinkingBudgets, ThinkingLevel}; diff --git a/turn-orchestrator/src/agent_call.rs b/turn-orchestrator/src/agent_call.rs index 5798a015..396be887 100644 --- a/turn-orchestrator/src/agent_call.rs +++ b/turn-orchestrator/src/agent_call.rs @@ -22,6 +22,14 @@ pub const FUNCTION_ID: &str = "agent::call"; /// override per call site when a tighter ceiling is wanted. pub(crate) const DEFAULT_DISPATCH_TIMEOUT_MS: Option = None; +/// Reserved virtual function id. When `dispatch()` sees this, it skips the +/// bus hop and reads the stashed full payload from +/// `session//result/` directly. Not registered on the +/// bus (no entry in `engine::functions::list`) — discoverable only via the +/// `[result truncated …]` marker emitted by the truncator and the recovery +/// clause in [`agent_call_tool`]'s description. +const RESULT_FETCH_FUNCTION: &str = "result::fetch"; + /// The single tool schema sent to the provider. /// /// Snapshot-tested below — any change here is a wire-format change for every @@ -29,7 +37,7 @@ pub(crate) const DEFAULT_DISPATCH_TIMEOUT_MS: Option = None; pub fn agent_call_tool() -> Value { json!({ "name": TOOL_NAME, - "description": "Call any iii function on the bus. The argument `function` is the function id (use `::` separators, e.g. `shell::fs::ls`). The argument `payload` is the function-specific JSON arguments. Skills loaded into your context tell you which functions exist and what arguments they take. The result is whatever that function returns.", + "description": "Call any iii function on the bus. The argument `function` is the function id (use `::` separators, e.g. `shell::fs::ls`). The argument `payload` is the function-specific JSON arguments. Skills loaded into your context tell you which functions exist and what arguments they take. The result is whatever that function returns. If a previous tool result was replaced with a `[result truncated …]` marker, retrieve the full payload by calling `agent_call` with `function=\"result::fetch\"` and `payload={\"call_id\": \"\"}`.", "parameters": { "type": "object", "properties": { @@ -60,6 +68,7 @@ fn error_result(envelope: Value) -> FunctionResult { })], details: envelope, terminate: false, + truncated: None, } } @@ -115,6 +124,7 @@ pub(crate) fn decode_or_passthrough(value: Value) -> FunctionResult { content: vec![ContentBlock::Text(TextContent { text })], details: value, terminate: false, + truncated: None, } } @@ -129,10 +139,9 @@ pub(crate) fn decode_or_passthrough(value: Value) -> FunctionResult { /// dispatch via `iii.trigger`, and map errors back to envelopes the model /// can read. /// -/// `_session_id` kept in the signature for caller symmetry; not consumed. pub async fn dispatch( iii: &III, - _session_id: &str, + session_id: &str, function: &Value, payload: Value, ) -> FunctionResult { @@ -141,6 +150,10 @@ pub async fn dispatch( Err(result) => return result, }; + if function_id == RESULT_FETCH_FUNCTION { + return handle_result_fetch(iii, session_id, &payload).await; + } + let response = iii .trigger(TriggerRequest { function_id: function_id.clone(), @@ -170,6 +183,38 @@ pub async fn dispatch( } } +/// Handle the virtual `result::fetch` function — reads the stashed full +/// payload for a previously-truncated tool call out of iii-state and +/// returns it verbatim (as if it had been returned the first time). +async fn handle_result_fetch(iii: &III, session_id: &str, payload: &Value) -> FunctionResult { + if session_id.is_empty() { + return error_result(json!({ + "error": "missing_session", + "function": RESULT_FETCH_FUNCTION, + "message": "result::fetch requires session context (dispatcher session_id was empty)" + })); + } + let call_id = match payload.get("call_id").and_then(Value::as_str) { + Some(s) if !s.is_empty() => s, + _ => { + return error_result(json!({ + "error": "missing_call_id", + "function": RESULT_FETCH_FUNCTION, + "message": "result::fetch requires a non-empty `call_id` string in payload" + })) + } + }; + match crate::persistence::load_full_result(iii, session_id, call_id).await { + Some(value) => decode_or_passthrough(value), + None => error_result(json!({ + "error": "result_not_found", + "function": RESULT_FETCH_FUNCTION, + "call_id": call_id, + "message": "no stashed result for this call_id (not truncated, or state pruned)" + })), + } +} + /// Register `agent::call` as a regular iii function so it appears in /// `engine::functions::list` and can be invoked through `bridge::trigger` /// for testing. The browser does not call this directly — `bridge::trigger` @@ -223,6 +268,29 @@ mod tests { "description should explain what `function` is" ); } + + #[test] + fn agent_call_description_documents_result_fetch_recovery() { + // The model only learns about `result::fetch` from the tool + // description plus the `[result truncated …]` marker emitted by the + // truncator. If we drop this from the description, agents will see + // truncated results in their context with no way to recover the + // full payload. + let tool = agent_call_tool(); + let desc = tool["description"].as_str().unwrap(); + assert!( + desc.contains("result::fetch"), + "tool description must mention the result::fetch recovery path" + ); + assert!( + desc.contains("call_id"), + "tool description must explain the call_id argument" + ); + assert!( + desc.contains("truncated"), + "tool description must reference the truncation marker" + ); + } } #[cfg(test)] diff --git a/turn-orchestrator/src/lib.rs b/turn-orchestrator/src/lib.rs index ed2c9356..3436033b 100644 --- a/turn-orchestrator/src/lib.rs +++ b/turn-orchestrator/src/lib.rs @@ -17,6 +17,7 @@ pub mod transitions; pub use config::TurnOrchestratorConfig; pub use register::register_with_iii; pub use state::{ - cwd_index_key, cwd_key, function_schemas_key, messages_key, run_request_key, sandbox_id_key, + cwd_index_key, cwd_key, function_schemas_key, last_compaction_at_key, + last_compaction_consumed_at_key, messages_key, run_request_key, sandbox_id_key, tool_schemas_key, turn_state_key, TurnState, TurnStateRecord, }; diff --git a/turn-orchestrator/src/persistence.rs b/turn-orchestrator/src/persistence.rs index 0e9dc34e..1a53e9f4 100644 --- a/turn-orchestrator/src/persistence.rs +++ b/turn-orchestrator/src/persistence.rs @@ -34,6 +34,82 @@ pub async fn load_messages(iii: &III, session_id: &str) -> Vec { serde_json::from_value(value).unwrap_or_default() } +/// Rebuild the hot `session//messages` key from the session-tree when +/// the `context-compaction` worker has stamped a fresh `last_compaction_at` +/// since our last reload. The tree's `load_messages` already filters +/// `Compaction` entries out of the active transcript, so this is the +/// canonical way to consume a freshly-landed compaction. +/// +/// No-op when no compaction is pending. Cheap (two `state::get`s on the +/// fast path), so safe to call at the top of every turn. +pub async fn maybe_reload_after_compaction(iii: &III, session_id: &str) { + let last_key = crate::state::last_compaction_at_key(session_id); + let watermark_key = crate::state::last_compaction_consumed_at_key(session_id); + let last = state_get(iii, &last_key) + .await + .and_then(|v| v.as_i64()) + .unwrap_or(0); + if last == 0 { + return; + } + let consumed = state_get(iii, &watermark_key) + .await + .and_then(|v| v.as_i64()) + .unwrap_or(0); + if last <= consumed { + return; + } + let resp = match iii + .trigger(TriggerRequest { + function_id: "session-tree::messages".into(), + payload: json!({ "session_id": session_id }), + action: None, + timeout_ms: Some(10_000), + }) + .await + { + Ok(v) => v, + Err(e) => { + tracing::warn!(error = %e, %session_id, "reload-after-compaction: session-tree::messages failed"); + return; + } + }; + // `session-tree::messages` returns `{messages: [{entry_id, message}, ...]}`. + let Some(rows) = resp.get("messages").and_then(|v| v.as_array()) else { + return; + }; + let mut rebuilt: Vec = Vec::with_capacity(rows.len()); + for row in rows { + let msg = row.get("message").cloned().unwrap_or(JsonValue::Null); + if msg.is_null() { + continue; + } + match serde_json::from_value::(msg) { + Ok(m) => rebuilt.push(m), + Err(e) => { + tracing::warn!(error = %e, %session_id, "reload-after-compaction: failed to decode AgentMessage; aborting reload"); + return; + } + } + } + if rebuilt.is_empty() { + return; + } + if let Ok(value) = serde_json::to_value(&rebuilt) { + state_set(iii, &messages_key(session_id), value).await; + } + // Reset the mirror watermark so future `save_messages` calls don't try + // to re-append messages the tree already has. + let mirror_key = crate::state::last_session_tree_len_key(session_id); + state_set(iii, &mirror_key, json!(rebuilt.len() as u64)).await; + state_set(iii, &watermark_key, json!(last)).await; + tracing::info!( + %session_id, + new_len = rebuilt.len(), + "context-compaction landed; reloaded messages from session-tree" + ); +} + pub async fn save_messages(iii: &III, session_id: &str, messages: &[AgentMessage]) { let key = messages_key(session_id); if let Ok(value) = serde_json::to_value(messages) { @@ -254,6 +330,23 @@ const EXECUTED_KEY: &str = "function_executed"; const LEGACY_PREPARED_KEY: &str = "tool_prepared"; const LEGACY_EXECUTED_KEY: &str = "tool_executed"; +/// Stash the full, untruncated payload for one function-call result. The +/// truncated `FunctionResult` that the orchestrator persists into the +/// message stream carries a `TruncationInfo` pointing back here; the model +/// recovers the full payload via `agent_call(function="result::fetch")`. +pub async fn save_full_result(iii: &III, session_id: &str, call_id: &str, payload: &Value) { + let key = staging_key(session_id, &format!("result/{call_id}")); + state_set(iii, &key, payload.clone()).await; +} + +/// Fetch a previously-stashed full result. Returns `None` if no payload +/// exists at the expected key (e.g. the result wasn't truncated, the +/// call_id is wrong, or state retention pruned it). +pub async fn load_full_result(iii: &III, session_id: &str, call_id: &str) -> Option { + let key = staging_key(session_id, &format!("result/{call_id}")); + state_get(iii, &key).await +} + fn staging_key(session_id: &str, suffix: &str) -> String { format!("session/{session_id}/{suffix}") } @@ -387,6 +480,7 @@ mod tests { content: vec![ContentBlock::Text(TextContent { text: text.into() })], details: json!({ "text": text }), terminate: false, + truncated: None, } } diff --git a/turn-orchestrator/src/state.rs b/turn-orchestrator/src/state.rs index 4b830a77..bce0cf42 100644 --- a/turn-orchestrator/src/state.rs +++ b/turn-orchestrator/src/state.rs @@ -131,6 +131,22 @@ pub fn last_session_tree_len_key(session_id: &str) -> String { format!("session/{session_id}/session_tree_mirror_len") } +/// Key under scope=agent that the `context-compaction` worker stamps with +/// `chrono::Utc::now().timestamp_millis()` after each successful +/// `session-tree::compact` append. The orchestrator watches it to decide +/// whether to rebuild the hot `messages_key` from session-tree on the +/// next turn. +pub fn last_compaction_at_key(session_id: &str) -> String { + format!("session/{session_id}/last_compaction_at") +} + +/// Watermark advanced by [`persistence::maybe_reload_after_compaction`] +/// whenever it consumes a fresh `last_compaction_at`. Reload only fires +/// when `last_compaction_at > last_compaction_consumed_at`. +pub fn last_compaction_consumed_at_key(session_id: &str) -> String { + format!("session/{session_id}/last_compaction_consumed_at") +} + #[allow(dead_code)] fn _ensure_message_types_in_scope(_: AgentMessage) {} diff --git a/turn-orchestrator/src/states/assistant.rs b/turn-orchestrator/src/states/assistant.rs index 71cc74f0..41f02059 100644 --- a/turn-orchestrator/src/states/assistant.rs +++ b/turn-orchestrator/src/states/assistant.rs @@ -75,6 +75,7 @@ pub async fn handle_awaiting(iii: &III, record: &mut TurnStateRecord) -> anyhow: } pub async fn handle_streaming(iii: &III, record: &mut TurnStateRecord) -> anyhow::Result<()> { + persistence::maybe_reload_after_compaction(iii, &record.session_id).await; let request = persistence::load_run_request(iii, &record.session_id).await; let messages = persistence::load_messages(iii, &record.session_id).await; let schemas = persistence::load_function_schemas(iii, &record.session_id).await; diff --git a/turn-orchestrator/src/states/functions.rs b/turn-orchestrator/src/states/functions.rs index 2a4a67d3..7af573fb 100644 --- a/turn-orchestrator/src/states/functions.rs +++ b/turn-orchestrator/src/states/functions.rs @@ -2,7 +2,7 @@ use harness_types::{ AgentEvent, AgentMessage, AssistantMessage, ContentBlock, FunctionCall, FunctionResult, - FunctionResultMessage, TextContent, + FunctionResultMessage, TextContent, TruncationInfo, }; use iii_sdk::{TriggerRequest, Value, III}; use serde_json::json; @@ -16,6 +16,123 @@ const TOPIC_BEFORE: &str = "agent::before_function_call"; const TOPIC_AFTER: &str = "agent::after_function_call"; const HOOK_TIMEOUT_MS: u64 = 10_000; +// ─── Tool-result truncation ────────────────────────────────────────────── +// +// Large tool outputs (multi-MB shell::run, big shell::fs::read) dominate +// per-turn token cost. We cap each FunctionResult at a serialized-bytes +// budget; oversized payloads get stashed under `session//result/` +// and the in-stream `content` is replaced with a head+tail-elided preview +// plus a marker telling the model how to call `result::fetch` to recover +// the full payload (intercepted in `agent_call::dispatch`). + +const DEFAULT_TRUNCATE_BYTES: usize = 8192; +const TRUNCATE_ENV: &str = "HARNESS_RESULT_TRUNCATE_BYTES"; +const TRUNCATE_HEAD_BYTES: usize = 2048; +const TRUNCATE_TAIL_BYTES: usize = 2048; + +fn truncate_threshold() -> usize { + static CACHED: std::sync::OnceLock = std::sync::OnceLock::new(); + *CACHED.get_or_init(|| { + std::env::var(TRUNCATE_ENV) + .ok() + .and_then(|s| s.parse::().ok()) + .filter(|&v| v > 0) + .unwrap_or(DEFAULT_TRUNCATE_BYTES) + }) +} + +/// If `result`'s serialized size exceeds the truncation threshold, stash the +/// full payload under `session//result/` and return a +/// compact replacement carrying a `TruncationInfo` pointer. Otherwise return +/// `result` unchanged. +async fn maybe_truncate_result( + iii: &III, + session_id: &str, + call_id: &str, + result: FunctionResult, +) -> FunctionResult { + let threshold = truncate_threshold(); + let serialized_size = match serde_json::to_string(&result) { + Ok(s) => s.len(), + Err(_) => return result, + }; + if serialized_size <= threshold { + return result; + } + // Persist the full payload first; if state::set fails, fall through and + // return the original result so we never lose data. + let full_json = match serde_json::to_value(&result) { + Ok(v) => v, + Err(_) => return result, + }; + persistence::save_full_result(iii, session_id, call_id, &full_json).await; + + let summary_text = render_truncated_text(&result, serialized_size, call_id); + FunctionResult { + content: vec![ContentBlock::Text(TextContent { text: summary_text })], + details: json!({ + "truncated": true, + "original_bytes": serialized_size, + "call_id": call_id, + }), + terminate: result.terminate, + truncated: Some(TruncationInfo { + original_bytes: serialized_size as u64, + call_id: call_id.to_string(), + }), + } +} + +/// Render the model-facing replacement text for a truncated result. Head + +/// tail elision keeps the most semantically useful parts (shell errors +/// typically live at the bottom, while file headers / call signatures live +/// at the top). +fn render_truncated_text(result: &FunctionResult, original_bytes: usize, call_id: &str) -> String { + let mut combined = String::new(); + for block in &result.content { + if let ContentBlock::Text(t) = block { + if !combined.is_empty() { + combined.push('\n'); + } + combined.push_str(&t.text); + } + } + let total = combined.len(); + let body = if total > TRUNCATE_HEAD_BYTES + TRUNCATE_TAIL_BYTES { + let head_end = char_boundary_floor(&combined, TRUNCATE_HEAD_BYTES); + let tail_start = char_boundary_ceil(&combined, total - TRUNCATE_TAIL_BYTES); + format!( + "{head}\n\n[... {elided} bytes elided ...]\n\n{tail}", + head = &combined[..head_end], + elided = total - head_end - (total - tail_start), + tail = &combined[tail_start..], + ) + } else { + combined + }; + format!( + "[result truncated — {original_bytes} bytes — call agent_call with \ + function=\"result::fetch\", payload={{\"call_id\": \"{call_id}\"}} \ + to retrieve the full output]\n\n{body}" + ) +} + +fn char_boundary_floor(s: &str, idx: usize) -> usize { + let mut i = idx.min(s.len()); + while i > 0 && !s.is_char_boundary(i) { + i -= 1; + } + i +} + +fn char_boundary_ceil(s: &str, idx: usize) -> usize { + let mut i = idx.min(s.len()); + while i < s.len() && !s.is_char_boundary(i) { + i += 1; + } + i +} + /// Map `tool_use {name: "agent_call", input: {function, payload}}` back to /// a normal [`FunctionCall`] carrying the inner function id. Non-`agent_call` /// calls pass through unchanged so legacy/test fixtures keep working. @@ -88,6 +205,7 @@ pub async fn handle_prepare(iii: &III, record: &mut TurnStateRecord) -> anyhow:: content: vec![ContentBlock::Text(TextContent { text: reason })], details: json!({ "blocked": true }), terminate: false, + truncated: None, }) } else { None @@ -162,6 +280,7 @@ pub async fn handle_execute(iii: &III, record: &mut TurnStateRecord) -> anyhow:: .get("error") .and_then(Value::as_str) .is_some(); + let result = maybe_truncate_result(iii, &record.session_id, &fc.id, result).await; persistence::upsert_executed_call(&mut results, (fc.clone(), result.clone(), is_error)); persistence::save_executed_calls(iii, &record.session_id, &results).await; @@ -431,6 +550,7 @@ mod tests { content: vec![], details: json!({}), terminate: false, + truncated: None, }, false, )]; @@ -453,6 +573,7 @@ mod tests { content: vec![ContentBlock::Text(TextContent { text: "ok".into() })], details: json!({}), terminate: false, + truncated: None, }; let evt = build_function_execution_event(&fc, &result, false); match evt { @@ -481,6 +602,7 @@ mod tests { })], details: json!({"blocked": true}), terminate: false, + truncated: None, }; let evt = build_function_execution_event(&fc, &blocked, true); match evt { @@ -573,4 +695,97 @@ mod tests { "handle_finalize must not .expect() last_assistant" ); } + + // ─── Truncation helpers ────────────────────────────────────────────── + + #[test] + fn render_truncated_text_short_input_kept_verbatim() { + let result = FunctionResult { + content: vec![ContentBlock::Text(TextContent { + text: "hello world".into(), + })], + details: json!({}), + terminate: false, + truncated: None, + }; + let out = render_truncated_text(&result, 12345, "call-abc"); + assert!(out.contains("result truncated")); + assert!(out.contains("12345 bytes")); + assert!(out.contains("call-abc")); + assert!(out.contains("hello world")); + assert!( + !out.contains("[... "), + "short content must not have ellipsis" + ); + } + + #[test] + fn render_truncated_text_large_input_head_tail_elides() { + let head = "H".repeat(3000); + let middle = "M".repeat(50_000); + let tail = "T".repeat(3000); + let combined = format!("{head}{middle}{tail}"); + let result = FunctionResult { + content: vec![ContentBlock::Text(TextContent { + text: combined.clone(), + })], + details: json!({}), + terminate: false, + truncated: None, + }; + let out = render_truncated_text(&result, combined.len(), "call-x"); + assert!(out.contains("[... "), "must contain elision marker"); + assert!(out.contains("bytes elided")); + // Head and tail preserved. + assert!(out.contains("HHHHHHHHHH")); + assert!(out.contains("TTTTTTTTTT")); + // Middle should be largely gone — fewer than 100 M's survive. + let m_count = out.chars().filter(|c| *c == 'M').count(); + assert!( + m_count < 100, + "middle should be elided, found {} 'M' chars", + m_count + ); + } + + #[test] + fn render_truncated_text_concatenates_multiple_text_blocks() { + let result = FunctionResult { + content: vec![ + ContentBlock::Text(TextContent { text: "alpha".into() }), + ContentBlock::Text(TextContent { text: "beta".into() }), + ], + details: json!({}), + terminate: false, + truncated: None, + }; + let out = render_truncated_text(&result, 999, "cid"); + assert!(out.contains("alpha")); + assert!(out.contains("beta")); + } + + #[test] + fn char_boundary_helpers_respect_utf8() { + let s = "héllo🦀world"; + for idx in 0..=s.len() { + let f = char_boundary_floor(s, idx); + assert!(f <= idx); + assert!(s.is_char_boundary(f)); + } + for idx in 0..=s.len() { + let c = char_boundary_ceil(s, idx); + assert!(c >= idx); + assert!(s.is_char_boundary(c)); + } + } + + #[test] + fn truncate_threshold_has_sane_default() { + // OnceLock means we can't reliably set/reset the env in-process; + // assert only that the default is the documented constant. + // (Per-env override is exercised manually / via integration tests.) + let t = truncate_threshold(); + assert!(t >= 1024, "threshold should never round down below 1 KB"); + assert!(t <= 1_000_000, "threshold should be sane"); + } } diff --git a/turn-orchestrator/tests/integration.rs b/turn-orchestrator/tests/integration.rs index ecf15455..83115a12 100644 --- a/turn-orchestrator/tests/integration.rs +++ b/turn-orchestrator/tests/integration.rs @@ -18,7 +18,24 @@ fn state_keys_distinct_per_facet() { turn_orchestrator::cwd_key(s), turn_orchestrator::sandbox_id_key(s), turn_orchestrator::function_schemas_key(s), + turn_orchestrator::last_compaction_at_key(s), + turn_orchestrator::last_compaction_consumed_at_key(s), ]; let unique: std::collections::HashSet<_> = keys.iter().collect(); assert_eq!(unique.len(), keys.len(), "every facet has a distinct key"); } + +#[test] +fn compaction_watermark_keys_namespace_by_session() { + let s = "sess-9"; + assert!(turn_orchestrator::last_compaction_at_key(s).contains(s)); + assert!(turn_orchestrator::last_compaction_consumed_at_key(s).contains(s)); + // The compactor writes `last_compaction_at`; the orchestrator advances + // `last_compaction_consumed_at`. If those names accidentally collided, + // the orchestrator could overwrite the compactor's stamp. + assert_ne!( + turn_orchestrator::last_compaction_at_key(s), + turn_orchestrator::last_compaction_consumed_at_key(s), + "compactor's stamp key must not collide with the orchestrator's watermark" + ); +}