fix(async-context-compression): parse summary responses safely#73
Merged
Fu-Jie merged 1 commit intoMay 22, 2026
Merged
Conversation
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
Comment on lines
+4821
to
+4828
| if isinstance(message, dict): | ||
| text = collect_text(message.get("content")) | ||
| if text.strip(): | ||
| return text.strip() | ||
|
|
||
| text = collect_text(first_choice.get("text")) | ||
| if text.strip(): | ||
| return text.strip() |
There was a problem hiding this comment.
在处理 choices 列表时，目前的逻辑对 message 和 first_choice 的取值较为固定（仅尝试了 content 或 text 键）。为了更充分地利用 collect_text 函数提供的多键名遍历和递归提取能力，建议直接对 message 和 first_choice 对象调用 collect_text。这样可以自动兼容那些将摘要文本放在 output_text 或其他常用字段中的非标准 Provider 响应，使解析逻辑更加稳健。
Suggested change
| if isinstance(message, dict): | |
| text = collect_text(message.get("content")) | |
| if text.strip(): | |
| return text.strip() | |
| text = collect_text(first_choice.get("text")) | |
| if text.strip(): | |
| return text.strip() | |
| if isinstance(message, dict): | |
| text = collect_text(message) | |
| if text.strip(): | |
| return text.strip() | |
| text = collect_text(first_choice) | |
| if text.strip(): | |
| return text.strip() |
Owner
|
I still think the risk is relatively high here. This code path has quite a few moving parts, and there are multiple places where the behavior can diverge depending on the provider response shape, so I would prefer to be cautious before merging. Could you please provide real runtime evidence from actual OpenWebUI testing? Specifically, it would help to see:
If the runtime logs look clean in both scenarios, I can reconsider merging. |
Contributor
Author
|
Hi. You can check |
Contributor
Author
|
Owner
|
@NexZhu I pushed a follow-up commit on top of this PR: Reason for the change:
I also synced the related v1.6.3 release-prep documentation files so the release surface stays consistent. Could you please test this updated branch in your environment? It would be especially helpful to confirm:
The targeted regression tests pass on my side, but since your environment is the one that originally exposed this path, I would still like a real runtime check from you before we finalize it. |
- accept responses-style summary payloads without a choices array - ignore reasoning-only output and surface compact empty-summary diagnostics - sync README, docs, indexes, and release notes for v1.6.4
Fu-Jie
force-pushed
the
feat/async-context-compression-response-parsing
branch
from
481ef4b to
e064c02
Compare
Owner
|
I rebased this PR on top of the merged v1.6.3 changes from #76 and force-updated the branch as v1.6.4, because the plugin-version check now requires a new version for the next sequential PR to main.\n\nWhat changed in this update:\n- Kept the response-shape parsing fix so _call_summary_llm() can accept output-only / Responses-style payloads instead of failing on a stale choices-only gate.\n- Preserved the current SUMMARY_FAIL_MODE behavior from main, so this PR only adds the remaining response-parsing compatibility layer on top of the shipped 1.6.3 behavior.\n- Refreshed the tests and release/documentation surfaces for v1.6.4.\n\nCould you please re-test on the updated branch? If possible, please include:\n1. a normal successful run\n2. a provider response that does not use choices (Responses-style output payload)\n3. browser-console logs or screenshots if anything still looks off |
Contributor
Author
|
@Fu-Jie I've re-tested your version in my environment, compression worked as expected. |
Fu-Jie
added a commit
that referenced
this pull request
May 22, 2026
…tions Merge PR #77 from mhajder/openwebui-extensions - Add complete pl-PL translation set to TRANSLATIONS dict - Update GEMINI.md to list pl-PL in supported languages - Add contributors NexZhu (code, PR #73) and mhajder (translation, PR #77) to README.md, README_CN.md, and .all-contributorsrc (badge: 8 → 10) Co-authored-by: mhajder <mhajder@users.noreply.github.com>
Owner
|
Thank you for the contribution, @NexZhu! 🎉 This PR was merged via a rebase into You've been added to the Contributors section of the project. Thanks for making the plugin more robust! 🚀 |
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
2 participants
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.
Summary
Fixes Async Context Compression summary extraction when the summary provider returns text in non-standard response shapes.
What Changed
choices[].message.contentoutput_textoutputmessage itemsreasoning_content,thinking, and reasoning output items, so private reasoning isnot saved as chat memory.
1.6.3.Verification
PYTHONPYCACHEPREFIX=/private/tmp/pycache-openwebui /usr/bin/python3 -m unittest plugins.filters.async-context- compression.test_async_context_compression.TestAsyncContextCompression.test_extract_summary_text_supports_alternate_response_shapes plugins.filters.async-context- compression.test_async_context_compression.TestAsyncContextCompression.test_call_summary_llm_rejects_empty_message_contentPYTHONPYCACHEPREFIX=/private/tmp/pycache-openwebui /usr/bin/python3 -m py_compile plugins/filters/async-context-compression/ async_context_compression.py plugins/filters/async-context-compression/test_async_context_compression.py