Feature footnote support

[tiberlas]

Hi,
I've created footnote support. To the API I added functions:

• Can read footnotes from a paragraph as a list. The footnote object contains footnote reference id and paragraphs.

doc = Document('test.docx')
paragraph = doc.paragraphs[0]
footnotes = paragraph.footnotes
for footnote in footnotes:
    print(f'footnote reference id: {footnote.id}')
    print(f'this footnote has paragraphs: {len(footnote.paragraphs)}')

• Can access all footnotes in a document as a list where the list index is footnote reference id for that footnote

doc = Document('test.docx')
for footnote in doc.footnotes:
    print(f'footnote reference id: {footnote.id}')
    print(f'this footnote has paragraphs: {len(footnote.paragraphs)}')

• Can add a footnote to the end of a paragraph. The footnote reference id is calculated based on the other footnotes in the document. On a created footnote object you can add paragraphs.

doc = Document('test.docx')
paragraph = doc.paragraphs[0]
new_footnote = paragraph.add_footnote()
new_footnote.add_paragraph('paragraph in footnote')

• Can read and change footnote properties in a section.

doc = Document('test.docx')
s = doc.section[0]
print(s.footnote_position)
print(s.footnote_number_format)
print(s.footnote_numbering_start_value)
print(s.footnote_numbering_restart_location)
s.footnote_position = 'beneathText'
s.footnote_number_format = 'hex'
s.footnote_numbering_start_value = 4
s.footnote_numbering_restart_location = 'eachPage'

[Cerebex]

@scanny Any updates on merging footnote support into the library?

[Cerebex]

@tiberlas I tested this update locally and it works great. I wish it was merged since being able to pull footnotes from the exact paragraph that they are referenced in great!

@scanny Happy to help with work to merge this into main if you would like support.

[afultonuq]

Would also be keen to see this one merged in @scanny!

[jschultz-dk]

I tested your footnote support, and it works, but it has an issue with large files. Writing 1,200 pages with Python-docx takes only a few seconds, but adding 2,200 footnotes makes it painfully slow. It takes about seven minutes. You can see it getting slower and slower the more footnotes are added.

I have not investigated the cause yet but a typical error is to generate lists in loops for indexing in LXML, it is much faster (x100 or more) in large files to use iteratators and methods such as getprevious () and getnext ()