- Fix CoreNLPClient call in `ConstituencyParserCoreNLP`. It seems like the context manager does not work will with multiple initialization and call. Now instead, we instantiate and start the client in `__init__` and use it in the for loop and close it on `__dell__`.
0.0.25
- Fix CoreNLPClient call in `ConstituencyParserCoreNLP`. We were creating the server in a for loop which is faster that the Client init thus it was creating multiple client instead of one.
0.0.24
- Change the approach to pass endpoint and start_server as keyword arguments in the parse method for `ConstituencyParserCoreNLP`.
0.0.23
- `ConstituencyParserCoreNLP` has now the start_server option to start if not already started. Based on [this](https://github.com/stanfordnlp/stanza/issues/810).