Add HOCRDocProprocessor and HocrVisualParser

[HiromuHota]

Description of the problems or issues

Is your pull request related to a problem? Please describe.

This is the second patch that follows #518 .

Does your pull request fix any issue.

N/A.

Description of the proposed changes

Add HOCRDocProprocessor and HocrVisualParser

Test plan

I added a few real hOCR example files.

Checklist

• I have updated the documentation accordingly.
• I have added tests to cover my changes.
• All new and existing tests passed.
• I have updated the CHANGELOG.rst accordingly.

[codecov-commenter]

Codecov Report

Merging #519 into master will increase coverage by 0.22%.
The diff coverage is 90.96%.

@@            Coverage Diff             @@
##           master     #519      +/-   ##
==========================================
+ Coverage   85.81%   86.03%   +0.22%     
==========================================
  Files          90       92       +2     
  Lines        4582     4769     +187     
  Branches      852      896      +44     
==========================================
+ Hits         3932     4103     +171     
- Misses        467      475       +8     
- Partials      183      191       +8     

Flag	Coverage Δ
#unittests	86.03% <90.96%> (+0.22%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
setup.py	0.00% <ø> (ø)
.../fonduer/parser/visual_parser/pdf_visual_parser.py	84.76% <ø> (ø)
...duer/parser/preprocessors/hocr_doc_preprocessor.py	86.66% <86.66%> (ø)
...fonduer/parser/visual_parser/hocr_visual_parser.py	95.23% <95.23%> (ø)
src/fonduer/parser/preprocessors/__init__.py	100.00% <100.00%> (ø)
src/fonduer/parser/visual_parser/__init__.py	100.00% <100.00%> (ø)
src/fonduer/parser/parser.py	93.39% <0.00%> (+0.23%)	⬆️

[HiromuHota]

Unfortunately, this PR would not fix the #12. Moreover, #12 won't be fixed by its nature.
I'd recommend the hOCR format if visual information is to be parsed because bboxes and words are perfectly linked in the hOCR.
If hOCR cannot be used for any reason, use PdfVisualParser as a last resort.

[senwu]

Can we get all conflicts resolved first? Thanks!

[HiromuHota]

I rebased on the master branch and resolved the conflicts. Thanks.

[lukehsiao]

Just a few nits.

[HiromuHota]

@lukehsiao thank you so much for reviewing such a big PR and your comments.
Please review my edits and replies. Feel free to ask any further clarification.

[lukehsiao]

Thanks!

I think @senwu wanted to take a look to, so I'll wait for him. But LGTM.

[senwu]

LGTM

[senwu]

Let's plan to have a tutorial about this PR. This is a really awesome improvement!!! Thoughts? @HiromuHota @lukehsiao

[HiromuHota]

I'll definitely update the existing tutorials.
Rather than adding a new tutorial, I'd like to

• pick one of the existing ones and replace HTMLDocPreprocessor and PdfVisualParser with HOCRDocProprocessor and HocrVisualParser if it is just drop-in replacement,
• Or enrich the "Parsing Documents into the Data Model" section of intro/Intro_Data_Model.ipynb.

[senwu]

That works as well. We need to show two things: 1) some basics about what's in it and how to use it, and 2) end to end run with high quality.

[HiromuHota]

An update: I've started replacing html with hocr (html -> pdfy -> pdf -> pdftotree -> hocr) in the wikipedia tutorial.
With a minor change at parser, all cells run with no error but #527 up until the LF part, which needs a major change.