Repository navigation
fix(sensor): keep Warp assistant replies to user queries - #199
Open
rudycelekli wants to merge 1 commit into
Open
rudycelekli wants to merge 1 commit into
rudycelekli wants to merge 1 commit into
Conversation
On the legacy Warp schema each exchange's model output lives in ai_blocks.output. The parser used that output for ActionResult exchanges but ignored it for Query exchanges, so the assistant's reply to every user prompt was dropped from chat_history. Emit it as an assistant message after the query. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
|
This branch has not been deployed
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What type of PR is this? (check all applicable)
Related issue: N/A (found by reading the code; reproduced on
mainat 6b33dff)What changed?
In the Query branch, also emit
_extract_llm_text(llm_output)as an assistant message (sequence_id<exchange_id>_response). Current schema (no ai_blocks, llm_output NULL) is unchanged.Why?
Sensor/adr_sensor/parsers/warp_parser.py:218-223. On the legacy schema (ai_blocks present) each exchange's model output is in ai_blocks.output. The parser uses it for ActionResult exchanges but in the Query branch only emits the user message, so the assistant's reply to every user prompt is dropped.
How did you test it?
user 'what is in this dir?'; branch also printsassistant 'Let me check with ls.'. New testtest_legacy_schema_retains_query_response; existingtest_parses_conversation_contentupdated (its fixture already stored "reply for ex-1" in the Query exchange's llm_output and the old assertion len==2 encoded the drop; now 3 with the reply checked). Main: 2 failed; branch: all Warp tests pass.Potential risks
The change only affects the input described above; other paths go through the same code as before, and the component's existing tests pass.
AI assistance: Claude Code (model Claude Opus 5.5) found this bug, wrote the fix and test, and drafted this description; submitted by @rudycelekli.
🤖 Generated with Claude Code