Htmlrag

- Fix the bug in BM25HTMLPruner.
- Fix the bug for duplicate child tags.
- Optimize the computing efficiency for super long HTML documents.

Thank LinXin04 for reporting some bugs.

- Fix some misguidence in reame.
- Parameter `max_node_words` is removed from class `GenHTMLPruner` since `v0.1.0`.

- Fix display problems when showing package with `pip show htmlrag`.
- Fix some problems in README.

- Add support for Chinese document.
- Optimize block tree construction.

- Add support for local embed inference.
- **BM25HTMLPruner** for bm25 embedding