1234567891011121314151617181920212223242526272829303132333435363738 |
- import os
- import sys
- sys.path.append(os.path.dirname(__file__) + "/../")
- import logging
- import traceback
- import pandas
- from format_convert import get_memory_info
- @get_memory_info.memory_decorator
- def xlsx2text(path, unique_type_dir):
- logging.info("into xlsx2text")
- try:
- try:
- # sheet_name=None, 即拿取所有sheet,存为dict
- df_dict = pandas.read_excel(path, header=None, keep_default_na=False, sheet_name=None)
- except Exception as e:
- logging.info("xlsx format error!")
- return [-3]
- df_list = [sheet for sheet in df_dict.values()]
- sheet_text = ""
- for df in df_list:
- text = '<table border="1">' + "\n"
- for index, row in df.iterrows():
- text = text + "<tr>"
- for r in row:
- text = text + "<td>" + str(r) + "</td>" + "\n"
- # print(text)
- text = text + "</tr>" + "\n"
- text = text + "</table>" + "\n"
- sheet_text += text
- return [sheet_text]
- except Exception as e:
- logging.info("xlsx2text error!")
- print("xlsx2text", traceback.print_exc())
- return [-1]
|