Tomoko Uchida

Add 1 million docs datasource for vector search benchmark

5

Mainly for convenience though, at least 1 million docs data source would be needed for benchmarking?

We now have `exitable` flag in benchmark to enable/disable `ExitableDirectoryReader` ( #168 ), it'd be great to have it in nightly benchmarks. https://home.apache.org/~mikemccand/lucenebench/

Add FixedShingleFilter

7

Relates to #1204. Shingle token filter is useful for some advanced applications (e.g., fast phrase search, or implementing word n-gram language models). Recently I was trying to implement a shingle...

It could be possible to fully automate the release procedure with Action's webhook event `release`. - Build artifacts and publish to PyPI - Build and upload documentation to the website...

Support system dictionaries other than mecab-ipadic.

2

Want to support UniDic and neologd. It would be nice if users can switch system dictionaries after installing.

like kuromoji: http://www.atilika.org/

JS error on Metrics UI with Nginx reverse proxy setup

13

### General information - **OS Platform and Distribution (e.g., Linux Ubuntu 16.04)**: Linux (Amazon Linux release 2 (Karoo)) - **ElasticHQ Version**: 3.5.0 - **Elasticsearch Version**: 6.5.4 - **Python version**: 3.7.1...

It would not be a trivial change though, I'd like to use `pyproject.toml` and Poetry instead of `setup.py` to make the packaging procedure more developer friendly. https://python-poetry.org/docs/pyproject/ The main concern...

動かなくなっている箇所を修正する。