2.18 空的docx无法被识别为非招标文件
This commit is contained in:
parent
bd8addd788
commit
518f5d7ac7
@ -1,9 +1,9 @@
|
||||
import time
|
||||
from PyPDF2 import PdfReader # 确保已安装 PyPDF2: pip install PyPDF2
|
||||
from docx import Document
|
||||
|
||||
from flask_app.general.llm.通义千问long import upload_file, qianwen_long
|
||||
|
||||
|
||||
def judge_zbfile_exec(file_path):
|
||||
"""
|
||||
判断文件是否属于招标文件,并返回结果。
|
||||
@ -18,7 +18,23 @@ def judge_zbfile_exec(file_path):
|
||||
num_pages = len(reader.pages)
|
||||
if num_pages <= 5:
|
||||
return False
|
||||
# 模拟使用大模型进行判断
|
||||
elif file_path.lower().endswith('.docx'):
|
||||
doc = Document(file_path)
|
||||
accumulated_text = ""
|
||||
chunk_size = 10 # 每次读取10个段落
|
||||
paragraphs = doc.paragraphs
|
||||
|
||||
for i in range(0, len(paragraphs), chunk_size):
|
||||
chunk = paragraphs[i:i + chunk_size]
|
||||
for para in chunk:
|
||||
accumulated_text += para.text
|
||||
# 判断累计字符数是否已经达到1000字,
|
||||
if len(accumulated_text) >= 1000:
|
||||
break
|
||||
# 若累计内容不足1000字,则直接返回False
|
||||
if len(accumulated_text) < 1000:
|
||||
return False
|
||||
# 使用大模型进行判断
|
||||
user_query = """该文件是否属于招标文件?如果是的话,请返回'是',如果不是的话,返回'否'。请不要返回其他解释或内容。
|
||||
以下是常见的招标文件类型:
|
||||
公开招标文件、邀请招标文件、竞争性谈判文件、竞争性磋商文件、询价文件、问询文件、货物类招标文件、工程类招标文件、施工类招标文件、服务类招标文件、比选文件。
|
||||
@ -42,7 +58,7 @@ def judge_zbfile_exec(file_path):
|
||||
|
||||
if __name__ == '__main__':
|
||||
start_time = time.time()
|
||||
pdf_path = r"C:\Users\Administrator\Desktop\新建文件夹 (3)\temp\乱码文件测试.doc"
|
||||
pdf_path = r"C:\Users\Administrator\Downloads\file1739842556194.docx"
|
||||
res = judge_zbfile_exec(pdf_path)
|
||||
if res:
|
||||
print("yes")
|
||||
|
Loading…
x
Reference in New Issue
Block a user