Repository navigation
feat(memory): consolidate working and coding experiences across sessions - #451
Draft
Leeminjing wants to merge 2 commits into
Draft
Leeminjing wants to merge 2 commits into
Leeminjing wants to merge 2 commits into
Conversation
Leeminjing
marked this pull request as ready for review
November 14, 2026 04:17
Leeminjing
marked this pull request as draft
November 15, 2026 08:02
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Change
Adds explicit cross-session experience consolidation to the Working/Coding SDK. Callers select source sessions, publish reusable derived experiences, query SQL-only publication receipts, and withdraw results. The change includes source-fingerprint invalidation, whole-batch output validation, SQLite/PostgreSQL migrations and publication contracts, sync/async client APIs, 11-language prompts, and usage/configuration documentation.
This is one independent SDK PR based on upstream
7b85404be1b1. It does not depend on #447 or include that PR's Qdrant native-retrieval interfaces. No VDB adapter, dependency manifest, lockfile, workflow, or MemHub implementation changes are included.User impact and acceptance
no_resultdecisions.outcome_unknownfor SQL verification.The acceptance criteria existed before implementation. A separately verifiable failing-test-before-implementation history has not been established; none is backfilled.
Validation
Current tested code:
ea99d85. Head94c7e4dadds documentation only.A complete unmodified archive of upstream
7b85404be1b1, using the same Python/dependencies/test services, produced 1919 passed / 73 failed / 33 skipped / 18 errors / 45 deselected, exit 1. Comparing complete JUnit testcase IDs found no current-only failure/error IDs; 90 IDs were shared. One baseline-only reembed failure was not counted as a fix. This comparison does not establish that every failure has the same cause or that the full suite is green.Environment: Windows, Python 3.12.10, SQLite 3.49.1, PostgreSQL 17.7/psycopg 3.3.6, Chroma 1.5.9, Qdrant client 1.19.1/Server 1.20.0, FAISS 1.15.1. PostgreSQL and Qdrant used newly created disposable Docker containers.
Commands, versions, results, baseline comparison, native-interface evidence, and limitations.
Native retrieval declaration
Derived experiences reuse existing native dense-vector interfaces: Chroma
Collection.upsert/query, Qdrantupsert/query_points, and FAISSIndexFlatIP/add/search. Fixed-version official interface references are linked in the validation record.New SQL receipt/read-set tables and reverse dependency indexes manage publication, validity, withdrawal, idempotency, and business visibility. They do not produce semantic or term-search candidates. No sidecar retrieval index or new keyword/BM25/full-text/hybrid channel is introduced.
Qdrant Local coverage is limited to the existing experience collection, not full dual-store initialization. Chroma HTTP, multiple Python versions, scale performance, and fresh real-model quality sampling were not run. Historical author-provided quality results are retained separately and are not presented as current-head validation.
Privacy, safety, and dependencies
Rollback
Stop invoking the new consolidation APIs. Back up upgraded business databases before reverting code; do not assume older releases can open the new PostgreSQL schema. Use a validated migration or restore procedure rather than dropping receipt/source data. Database downgrade/restore has not been rehearsed in this draft.
Draft status
Keep this PR as draft while current-head CI, the red full/focused regression results, validation limitations, and code review remain to be resolved. No existing #447 CI success is attributed to this PR.