fix(llm): remove duplicated _appears_truncated stub

[NightingaleCen]

Related Issue

Regression introduced by #775 (fix(llm): resolve DeepSeek 500 / Failed to analyze resume (#706), commit 4dda99f). No dedicated issue tracks this — the symptom is uv run app failing immediately at import time with an IndentationError, so anyone on main cannot start the backend at all.

Description

app/llm.py ends up with two definitions of _appears_truncated, both broken, so the entire backend crashes at import time:

File ".../apps/backend/app/llm.py", line 735
    FALLBACK_MAX_TOKENS = 4096
    ^^^^^^^^^^^^^^^^^^^
IndentationError: expected an indented block after function definition on line 734

Root cause (introduced by #775)

git blame upstream/main -- apps/backend/app/llm.py shows the refactor in #775 was only partially applied:

• L734 (commit 4dda99f, PR fix(llm): resolve DeepSeek 500 / Failed to analyze resume #775): a new schema-aware signature def _appears_truncated(data: dict, schema_type: str = "resume") -> bool: was inserted, immediately followed by FALLBACK_MAX_TOKENS = 4096 -> IndentationError at module load.
• L782 (commit d2567520, original): the original implementation def _appears_truncated(data: dict) -> bool: was left in place. Its docstring was updated by fix(llm): resolve DeepSeek 500 / Failed to analyze resume #775, but its signature was not — so schema_type is referenced inside the body without being a parameter, and the only call site _appears_truncated(result, schema_type) at L1098 (also added by fix(llm): resolve DeepSeek 500 / Failed to analyze resume #775) would have raised NameError/TypeError even after fixing the IndentationError.

In other words: the new signature was pasted on top, the function body was never moved, and the old definition was never deleted.

Type

• Bug Fix
• Feature Enhancement
• Documentation Update
• Code Refactoring
• Other (please specify):

Proposed Changes

• Remove the empty stub at app/llm.py:734 (keep FALLBACK_MAX_TOKENS = 4096, which get_safe_max_tokens depends on).
• Update the real implementation at app/llm.py:782 to use the schema-aware signature def _appears_truncated(data: dict, schema_type: str = "resume") -> bool: so it matches both its body and its callers.

Screenshots / Code Snippets (if applicable)

Reproducer on main:

$ uv run app
Traceback (most recent call last):
  File ".../apps/backend/.venv/bin/app", line 4, in <module>
    from app.main import main
  File ".../apps/backend/app/main.py", line 21, in <module>
    from app.routers import config_router, enrichment_router, health_router, jobs_router, resumes_router
  File ".../apps/backend/app/routers/__init__.py", line 3, in <module>
    from app.routers.config import router as config_router
  File ".../apps/backend/app/routers/config.py", line 10, in <module>
    from app.llm import check_llm_health, LLMConfig, resolve_api_key
  File ".../apps/backend/app/llm.py", line 735
    FALLBACK_MAX_TOKENS = 4096
    ^^^^^^^^^^^^^^^^^^^
IndentationError: expected an indented block after function definition on line 734

How to Test

1. uv run python -c "from app.llm import _appears_truncated; print('ok')" -> prints ok (no IndentationError).
2. uv run pytest tests/unit/test_llm.py::TestAppearsTruncated -q -> 13/13 passing
3. uv run pytest -q -> 179 passing. One pre-existing failure unrelated to this PR remains.
4. uv run app -> backend boots without IndentationError.

Checklist

• The code compiles successfully without any errors or warnings
• The changes have been tested and verified
• The documentation has been updated (if applicable)
• The changes follow the project's coding guidelines and best practices
• The commit messages are descriptive and follow the project's guidelines
• All tests (if applicable) pass successfully
• [] This pull request has been linked to the related issue (if applicable)

Additional Information

While verifying this fix locally I had to pin python-dotenv back to 1.0.1 because 1.2.2 (the current pyproject.toml pin from #761) conflicts with litellm's ==1.0.1 requirement, so uv sync itself fails. The dependency fix is in #780.

---

Summary by cubic

Fixes backend import crash by removing a duplicate _appears_truncated stub and aligning the function signature with callers. Restores uv run app startup by resolving the IndentationError.

• Bug Fixes
  • Removed the empty _appears_truncated definition; kept FALLBACK_MAX_TOKENS = 4096.
  • Updated _appears_truncated to def _appears_truncated(data: dict, schema_type: str = "resume") -> bool: to match usage.

^{Written for commit 820dc37. Summary will update on new commits.}

[cubic-dev-ai]

No issues found across 1 file

[Copilot]

Pull request overview

Fixes a backend-breaking regression in apps/backend/app/llm.py where an incomplete duplicate _appears_truncated definition caused an IndentationError at import time, preventing the backend from starting.

Changes:

• Remove the stray _appears_truncated(...) stub that left FALLBACK_MAX_TOKENS = 4096 incorrectly positioned and triggered IndentationError.
• Update the actual _appears_truncated implementation to accept schema_type: str = "resume" to match its internal logic and call sites.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.