Pdftotree

Added
- Support for Python 3.8.
([86](https://github.com/HazyResearch/pdftotree/pull/86), [HiromuHota][HiromuHota])

Changed
- Switch the output format from "HTML-like" to hOCR.
([62](https://github.com/HazyResearch/pdftotree/pull/62), [HiromuHota][HiromuHota])
- Loosen Keras' version restriction, which is now unnecessarily strict.
([68](https://github.com/HazyResearch/pdftotree/pull/68), [HiromuHota][HiromuHota])
- Greedily extract contents from PDF even if it looks scanned.
([71](https://github.com/HazyResearch/pdftotree/pull/71), [HiromuHota][HiromuHota])
- Upgrade Keras to 2.4.0 or later (and TensorFlow 2.2 or later).
([86](https://github.com/HazyResearch/pdftotree/pull/86), [HiromuHota][HiromuHota])

Removed
- Remove "favor_figures" option and extract everything.
([77](https://github.com/HazyResearch/pdftotree/pull/77), [HiromuHota][HiromuHota])
- Remove "dry_run" option.
([89](https://github.com/HazyResearch/pdftotree/pull/89), [HiromuHota][HiromuHota])

Fixed
- Fix a bug that an html file is not created at a given path.
([64](https://github.com/HazyResearch/pdftotree/pull/64), [HiromuHota][HiromuHota])
- Extract LTChar even if they are not children of LTTextLine.
([79](https://github.com/HazyResearch/pdftotree/pull/79), [HiromuHota][HiromuHota])

Fixed
- Temporarily add `chardet` to requirements until
[pdfminer/pdfminer.six213](https://github.com/pdfminer/pdfminer.six/issues/213) is fixed.
([47](https://github.com/HazyResearch/pdftotree/issues/47), [lukehsiao][lh])
- Fix ValueError when a Node instance is a single element.
([49](https://github.com/HazyResearch/pdftotree/pull/49), [mgoo][mgoo])

[lh]: https://github.com/lukehsiao
[mgoo]: https://github.com/mgoo
[HiromuHota]: https://github.com/HiromuHota