An error is encountered when attempting to load a Confluence page that has attachments into Helix GPT. 

The ErrorDetails in the record show the following trace information:

  Traceback (most recent call last):
  File "/opt/bmc/data-connection/app/jobs/service.py", line 417, in handle_job_step
    handler(job, job_step, chain, connection)
  File "/opt/bmc/data-connection/app/connections/confluence/loader.py", line 25, in load_confluence_page
    load_page_attachments(job, job_step, chain, page, confluence_service, connection.id)
  File "/opt/bmc/data-connection/app/connections/confluence/loader.py", line 54, in load_page_attachments
    load_attachment(job, job_step, chain, attachment, confluence_service, temp_dir_name, connection_id)
  File "/opt/bmc/data-connection/app/connections/confluence/loader.py", line 68, in load_attachment
    langchain_document = loader.load()
                         ^^^^^^^^^^^^^
  File "/opt/bmc/.local/lib/python3.11/site-packages/langchain_community/document_loaders/word_document.py", line 57, in load
    page_content=docx2txt.process(self.file_path),
                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/bmc/.local/lib/python3.11/site-packages/docx2txt/docx2txt.py", line 76, in process
    zipf = zipfile.ZipFile(docx)
           ^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/zipfile.py", line 1312, in __init__
    self._RealGetContents()
  File "/usr/local/lib/python3.11/zipfile.py", line 1379, in _RealGetContents
    raise BadZipFile("File is not a zip file")
zipfile.BadZipFile: File is not a zip file

Make sure you can open the attachment from the Confluence page directly. If not, then it will not load into Helix GPT successfully. 

The attachment file has to be fixed for the entire page to process properly since the load step is all inclusive. There is not a separate load step for the attachment and then the page. For this connector it is one in the same. In order for the page to load, any allowed attachments files types must be valid. Fix the attachments on those pages for everything to process properly.