Code clean

This commit is contained in:
Timothyxxx
2024-03-14 12:54:10 +08:00
parent 26d52a7231
commit 0aae756538
12 changed files with 120 additions and 384 deletions

View File

@@ -2,6 +2,7 @@ import operator
from typing import Any
from typing import Dict
import fitz # PyMuPDF
from pypdf import PdfReader
@@ -11,3 +12,20 @@ def check_pdf_pages(pdf_file: str, rules: Dict[str, Any]) -> float:
reader = PdfReader(pdf_file)
nb_pages: int = len(reader.pages)
return float(getattr(operator, rules["relation"])(nb_pages, rules["ref_value"]))
def extract_answers_from_pdf(pdf_file):
doc = fitz.open(pdf_file)
answers = []
for page in doc:
text = page.get_text()
lines = text.split('\n')
for line in lines:
if line.strip():
parts = line.split('=')
if len(parts) > 1:
answer = parts[-1].strip()
answers.append(answer)
return answers