merge doc in a wrong way

[BaiYouShiWo]

In the file “pdf_recovery_doc.py”, function "merge_docx_v1" and “merge_docx_v2” use sort() simply like

docx_files_list = sorted(docx_files_list)

I don't think this can achieve the expected results, the correct way may be

      import re
      docx_files_list = sorted(
      docx_files_list,
      key=lambda x: int(re.search(r'_(\d+)_', x).group(1))
      )

If possible, I can submit a pull request later

[jiangnanboy]

ok