Mark/infrastructure
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

AI project: duplicate method body in awm_flow.py, unreachable except blocks, stub LLM implementations #5

New issue
Closed
opened 2026-05-09 00:02:54 +00:00 by Mark · 3 comments
Mark commented 2026-05-09 00:02:54 +00:00
Owner
Copy link

AI project (~/Projects/AI/) has 15+ sub-projects. Critical bugs:

  1. awm_flow.py lines 306-362: duplicate method body (30 lines copied verbatim)
  2. awm_flow.py lines 133-146: two identical except Exception blocks, second is dead code
  3. llm.py: OpenAILLMClient and AnthropicLLMClient both raise NotImplementedError — all LLM implementations are stubs
  4. tool_manager.py: broken import path (from src. — PYTHONPATH issue)
  5. interpreter_crew.py: llm=None passed to ConversationSummaryMemory — will crash at runtime
AI project (~/Projects/AI/) has 15+ sub-projects. Critical bugs: 1. awm_flow.py lines 306-362: duplicate method body (30 lines copied verbatim) 2. awm_flow.py lines 133-146: two identical except Exception blocks, second is dead code 3. llm.py: OpenAILLMClient and AnthropicLLMClient both raise NotImplementedError — all LLM implementations are stubs 4. tool_manager.py: broken import path (from src. — PYTHONPATH issue) 5. interpreter_crew.py: llm=None passed to ConversationSummaryMemory — will crash at runtime
Mark commented 2026-05-17 04:15:29 +00:00
Author
Owner
Copy link

Feature Triage Assessment

Scope: medium
Risk: medium
Spec Alignment: misaligned
Summary: Fix 4 Python code quality and functional bugs in the AI project's crew_awm sub-project — duplicate 30-line method body in awm_flow.py:306-362, dead except block at awm_flow.py:141-146, stub OpenAI/Anthropic LLM clients in llm.py (all raise NotImplementedError), broken from src. import in tool_manager.py, and llm=None crash in interpreter_crew.py's ConversationSummaryMemory.
Recommendation: Close as misaligned — this issue describes application-level bugs in ~/Projects/AI/crew_awm, not infrastructure/system-level concerns defined in the infrastructure repo's README. If the bugs warrant fixing: (1) the duplicated method and dead except in awm_flow.py are trivial to remove, (2) llm.py's stubs comment that callers should use CrewAI directly — consider deleting the abstraction as dead code rather than implementing real clients, (3) tool_manager.py's import assumes PYTHONPATH is set, (4) interpreter_crew.py needs a fallback LLM before ConversationSummaryMemory is constructed. File these in a project-specific tracker, not infrastructure.
Rationale: SCOPE=medium because although only 4 files are touched, implementing real OpenAI/Anthropic LLM clients involves significant new API logic; the other 3 fixes are trivial. RISK=medium because LLM client implementation involves network I/O and credentials. SPEC=misaligned because the infrastructure repo's README explicitly scopes it to 'System-level bugs (Hindsight duplicates, CUDA mismatches, provider rotation)' — not application code bugs in a separate project. The llm.py stubs even tell callers to use CrewAI directly, suggesting the abstraction itself may be unused and could be deleted rather than implemented.

Labeled blocked:needs-decision — unblock on Forgejo to dispatch to the fixer pipeline.

## Feature Triage Assessment **Scope:** medium **Risk:** medium **Spec Alignment:** misaligned **Summary:** Fix 4 Python code quality and functional bugs in the AI project's crew_awm sub-project — duplicate 30-line method body in awm_flow.py:306-362, dead except block at awm_flow.py:141-146, stub OpenAI/Anthropic LLM clients in llm.py (all raise NotImplementedError), broken from src. import in tool_manager.py, and llm=None crash in interpreter_crew.py's ConversationSummaryMemory. **Recommendation:** Close as misaligned — this issue describes application-level bugs in ~/Projects/AI/crew_awm, not infrastructure/system-level concerns defined in the infrastructure repo's README. If the bugs warrant fixing: (1) the duplicated method and dead except in awm_flow.py are trivial to remove, (2) llm.py's stubs comment that callers should use CrewAI directly — consider deleting the abstraction as dead code rather than implementing real clients, (3) tool_manager.py's import assumes PYTHONPATH is set, (4) interpreter_crew.py needs a fallback LLM before ConversationSummaryMemory is constructed. File these in a project-specific tracker, not infrastructure. **Rationale:** SCOPE=medium because although only 4 files are touched, implementing real OpenAI/Anthropic LLM clients involves significant new API logic; the other 3 fixes are trivial. RISK=medium because LLM client implementation involves network I/O and credentials. SPEC=misaligned because the infrastructure repo's README explicitly scopes it to 'System-level bugs (Hindsight duplicates, CUDA mismatches, provider rotation)' — not application code bugs in a separate project. The llm.py stubs even tell callers to use CrewAI directly, suggesting the abstraction itself may be unused and could be deleted rather than implemented. Labeled `blocked:needs-decision` — unblock on Forgejo to dispatch to the fixer pipeline.
Mark commented 2026-05-28 10:12:53 +00:00
Author
Owner
Copy link

⏸ Needs Decision

What needs to be decided?

  • Is this issue still relevant? If yes, remove the blocked:needs-decision label to route back to triage.

Blocking factors:

  • Didn't meet auto-approve criteria — scope=small, risk=low, spec=partial

Scope: small Risk: low Spec: partial Type: bug
Summary: Fix 5 bugs in the AI project (~/Projects/AI/): duplicate method body in awm_flow.py, dead except blocks, unimplemented LLM client stubs in llm.py, broken import path in tool_manager.py, and llm=None crash in interpreter_crew.py.
Suggestion: Fix each bug with targeted patches: (1) remove duplicated lines 306-362 in awm_flow.py, (2) remove the second identical except block at lines 133-146, (3) implement OpenAI/Anthropic LLM client methods with actual API calls, (4) fix import path in tool_manager.py from src. to correct project structure, (5) pass a valid llm instance to ConversationSummaryMemory. Could be done as a single commit given small scope.
Rationale: Scope=small because all 5 bugs touch <5 files with trivial-to-modest fixes (mostly deletions and one-liners). Risk=low because all changes are additive/restorative — fixing broken code, not touching working paths. Spec=partial because the bugs are in ~/Projects/AI/ code being tracked through the infrastructure repo, which is a scope stretch from the repo's stated purpose of infrastructure issue tracking. Type=bug throughout.

To unblock: Remove the blocked:needs-decision label. The triage will apply the appropriate outcome based on your choice above.

## ⏸ Needs Decision **What needs to be decided?** - **Is this issue still relevant?** If yes, remove the `blocked:needs-decision` label to route back to triage. **Blocking factors:** - **Didn't meet auto-approve criteria** — scope=small, risk=low, spec=partial --- **Scope:** small **Risk:** low **Spec:** partial **Type:** bug **Summary:** Fix 5 bugs in the AI project (~/Projects/AI/): duplicate method body in awm_flow.py, dead except blocks, unimplemented LLM client stubs in llm.py, broken import path in tool_manager.py, and llm=None crash in interpreter_crew.py. **Suggestion:** Fix each bug with targeted patches: (1) remove duplicated lines 306-362 in awm_flow.py, (2) remove the second identical except block at lines 133-146, (3) implement OpenAI/Anthropic LLM client methods with actual API calls, (4) fix import path in tool_manager.py from src. to correct project structure, (5) pass a valid llm instance to ConversationSummaryMemory. Could be done as a single commit given small scope. **Rationale:** Scope=small because all 5 bugs touch <5 files with trivial-to-modest fixes (mostly deletions and one-liners). Risk=low because all changes are additive/restorative — fixing broken code, not touching working paths. Spec=partial because the bugs are in ~/Projects/AI/ code being tracked through the infrastructure repo, which is a scope stretch from the repo's stated purpose of infrastructure issue tracking. Type=bug throughout. **To unblock:** Remove the `blocked:needs-decision` label. The triage will apply the appropriate outcome based on your choice above.
Mark commented 2026-06-16 05:23:37 +00:00
Author
Owner
Copy link

Overseer recovery — session timed out (10min idle)

Findings from quickfix-infrastructure-5:

  • The session never started editing the target files. Last log lines show it was globbing the wrong paths (**/awm_flow.py, **/llm.py, etc.) against the infrastructure repo, getting 0 matches, and idling.
  • Issue #5's target files (awm_flow.py, llm.py, tool_manager.py, interpreter_crew.py) live in ~/Projects/AI/crew_awm, not in infrastructure. The quickfix pipeline checked out Mark/infrastructure and expected the model to discover the cross-repo target, which didn't happen.
  • The 2 unpushed commits on infrastructure/main (0fd316e, 1630018 — pre-commit hooks) are unrelated; they belong to issue #77, not #5.

Recovery actions taken:

  • Killed the stuck session (was already gone — tmux reported can't find session).
  • Replaced quickfix label with triage-approved + blocked:needs-human (note: blocked:needs-attention does not exist in this repo; blocked:needs-human is the closest semantic match).
  • Left ~/Projects/AI/crew_awm untouched (only pre-staged chroma_tools/embeddings deletions from an unrelated cleanup, not part of this issue).

Recommended next step: Issue #5 needs to be re-dispatched with crew_awm as the target repo, or its target-path needs to be clarified in the issue body. It cannot be fixed by a session running in infrastructure.

**Overseer recovery — session timed out (10min idle)** Findings from `quickfix-infrastructure-5`: - The session never started editing the target files. Last log lines show it was globbing the wrong paths (`**/awm_flow.py`, `**/llm.py`, etc.) against the `infrastructure` repo, getting 0 matches, and idling. - Issue #5's target files (`awm_flow.py`, `llm.py`, `tool_manager.py`, `interpreter_crew.py`) live in `~/Projects/AI/crew_awm`, not in `infrastructure`. The quickfix pipeline checked out `Mark/infrastructure` and expected the model to discover the cross-repo target, which didn't happen. - The 2 unpushed commits on `infrastructure/main` (`0fd316e`, `1630018` — pre-commit hooks) are unrelated; they belong to issue #77, not #5. Recovery actions taken: - Killed the stuck session (was already gone — `tmux` reported `can't find session`). - Replaced `quickfix` label with `triage-approved` + `blocked:needs-human` (note: `blocked:needs-attention` does not exist in this repo; `blocked:needs-human` is the closest semantic match). - Left `~/Projects/AI/crew_awm` untouched (only pre-staged chroma_tools/embeddings deletions from an unrelated cleanup, not part of this issue). **Recommended next step:** Issue #5 needs to be re-dispatched with `crew_awm` as the target repo, or its target-path needs to be clarified in the issue body. It cannot be fixed by a session running in `infrastructure`.
Sign in to join this conversation.
No Branch/Tag specified
Branches Tags
main
No results found.
Labels
Clear labels   
blocked:gate-research

   
blocked:needs-decision

   
blocked:needs-human

   
blocked:no-changes

   
bug

Something is broken

  
documentation

   
enhancement

New feature or request

  
feature-request

   
in-progress

Issue is currently being worked on by a quickfix runner

  
quickfix

   
research

Needs investigation before action

  
triage-approved
No labels
blocked:gate-research
blocked:needs-decision
blocked:needs-human
blocked:no-changes
bug
documentation
enhancement
feature-request
in-progress
quickfix
research
triage-approved
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
Mark/infrastructure#5
No description provided.