word-documents

Library

Use the installed python-docx through python for .docx. Do not install packages, and do not use bash or a shell to read one.

For legacy .doc, antiword is installed; run LC_ALL=C antiword -m UTF-8.txt -w 0 "/source/path/to/file.doc" with bash to get the text. For an attachment, use its listed /run/attachments/... path instead.

Find The Files

Search /source recursively for files ending in .docx or .doc, case-insensitive.

Recipe

from docx import Document

document = Document(path)
for paragraph in document.paragraphs:
    print(paragraph.text)
for table in document.tables:
    for row in table.rows:
        print([cell.text for cell in row.cells])

To create or edit a .docx, load it with Document(path), add or change paragraphs, table rows, or styles, then document.save(path). Never create or edit a .doc; produce a .docx deliverable instead.

Verify

Reopen every .docx you write with Document(path) and assert its paragraph or table row count matches what you intended.

Gotchas

• A table's text lives in table.rows, not in document.paragraphs.
• Tables become plain text; .doc layout and embedded content are lost when read through antiword.
• antiword output is source text only; you cannot edit a .doc file with it.
• On an encrypted, corrupt, or mislabeled file, stop and report it; there is no fallback reader.
• Cite results by path and section, or by paragraph number for a .doc extract.