codeflash-admin/codeflash-internal
1
0
Fork 0
mirror of https://github.com/codeflash-ai/codeflash-internal.git synced 2026-05-04 18:25:18 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

fix: handle syntactically invalid LLM output in testgen repair (#2472)

Browse source 
## Summary
- Catch `ParserSyntaxError` when parsing LLM-repaired code instead of
letting it bubble to the generic 500 handler
- Reduces Sentry noise from expected LLM failures
- The CLI already handles non-200 responses gracefully (returns `None`,
continues)
This commit is contained in:
Kevin Turcios 2026-03-06 07:32:30 +00:00 committed by GitHub
parent 4edd183d82
commit 14c0b3acca
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 5 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

6
django/aiservice/core/shared/testgen_review/repair.py
View file

@ -90,7 +90,11 @@ async def testgen_repair(
from core.languages.python.testgen.validate import instrument_tests, validate_request_data # noqa: PLC0415
from core.shared.testgen_models import TestGenSchema # noqa: PLC0415
repaired_cst = parse_module_to_cst(repaired_code)
try:
repaired_cst = parse_module_to_cst(repaired_code)
except Exception:
logging.warning("LLM returned syntactically invalid repaired code, falling back to original")
return 500, TestRepairErrorSchema(error="LLM returned syntactically invalid code")
original_cst = parse_module_to_cst(data.test_source)
# Extract repaired function nodes by name