sqs/contracts-flask
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Avoid false garbage classification for UPD mentions
Deploy contracts-flask / validate (push) Canceled after 0s
Details

Browse Source
This commit is contained in:
“Naeel”
2026-08-27 10:44:48 +03:00
parent 475efefb40
commit e5c7c88073
5 changed files with 75 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files
site/services/classify.py
+6 -2
View File
@@ -47,9 +47,13 @@ _GARBAGE_HEADER_MARKERS = [
'СЧЕТ-ФАКТУРА', 'СЧЕТ НА ОПЛАТУ', 'АКТ СВЕРКИ',
'АКТ ОКАЗАННЫХ УСЛУГ', 'АКТ ВЫПОЛНЕННЫХ РАБОТ',
'ПЛАТЁЖНОЕ ПОРУЧЕНИЕ', 'УНИВЕРСАЛЬНЫЙ ПЕРЕДАТОЧНЫЙ',
'УПД', 'ПЛАТЕЖНОЕ ПОРУЧЕНИЕ',
'ПЛАТЕЖНОЕ ПОРУЧЕНИЕ',
]
_GARBAGE_HEADER_RE = re.compile(
r'(?m)^\s*(?:УПД|УНИВЕРСАЛЬНЫЙ ПЕРЕДАТОЧНЫЙ ДОКУМЕНТ)\b'
)
def _is_garbage_by_filename(filename: str) -> bool:
"""Stage 1: regex по имени файла — быстро, 0 токенов."""
@@ -59,7 +63,7 @@ def _is_garbage_by_filename(filename: str) -> bool:
def _is_garbage_by_header(text: str) -> bool:
"""Stage 2: ключевые слова в первых 2KB текста — быстро, 0 токенов."""
header = text[:2000].upper()
return any(marker in header for marker in _GARBAGE_HEADER_MARKERS)
return any(marker in header for marker in _GARBAGE_HEADER_MARKERS) or bool(_GARBAGE_HEADER_RE.search(header))
def _call_llm_classify(header_text, llm_client=None):