Hi, I had a very brief look. I don’t know why there is a Python server implementation. We have a fully GRPC capable backend implementation located here: https://github.com/apache/opennlp-sandbox/tree/main/opennlp-grpc You can just run that via Docker (for example). This would also be the thing to upgrade / enhance for further NLP tasks.
On the other side, we would need to implement a Python client (based on the proto definition) and wrap that into something a Python person is more familiar with. I guess that the best way would be to fork opennlp-sandbox and provide your updates in your fork of the sandbox, so it can easily be integrated. Gruß Richard > Am 16.09.2025 um 13:56 schrieb Jobin Sabu <85jobins...@gmail.com>: > > Good evening OpenNLP devs, > I hope you’re doing well. I wanted to apologise for the delay in following > up on the gRPC-based Python integration we discussed earlier — university > exams and related academic commitments took up more time than I expected. > Things have settled down now, and I’m ready to continue the work. > > Quick status update (minimal prototype): > > I generated and tested the Python gRPC stubs from proto/opennlp.proto. > > I implemented a small Python gRPC server that exposes the tokenizer service > and a simple client example. > > I verified the flow locally using the official Apache OpenNLP CLI (OpenNLP > 2.5.5). > > Code + usage examples are on GitHub: > https://github.com/JOBIN-SABU/Apache_opennlp_grpc > > > Next steps I intend to take (and would value your input on): > > Open a short PR with this minimal prototype so it’s easy for reviewers to > run and test. > > Expand to additional services (POS tagging, sentence detection) based on > your feedback and preferences. > > Ensure the implementation and packaging meet ASF guidelines for > contribution. > > > If you or other maintainers have a moment, could you try the repo and share > any initial feedback? I can open a PR with a concise README and test > instructions so it’s straightforward to run locally. I’d greatly appreciate > any guidance on preferred design choices or packaging requirements before I > expand the scope. > > Thank you again for the earlier encouragement — I’m excited to contribute > and will follow your lead on next steps. > > Best regards, > Jobin Sabu > 85jobins...@gmail.com > https://github.com/JOBIN-SABU/Apache_opennlp_grpc > > > On Sun, 18 May, 2025, 12:38 am Richard Zowalla, <r...@apache.org> wrote: > >> Hi, >> >> Feel free to contribute to OpenNLP. >> >> You can ask any question on the list or we can open a discussion with code >> (i.e. in a PR). >> Every contribution in that area is valuable, imho. >> >> Gruß >> Richard >> >>> Am 10.05.2025 um 05:38 schrieb Jobin Sabu <85jobins...@gmail.com>: >>> >>> Respected Apache OpenNLP Team, >>> >>> I hope you're doing well. >>> >>> Although I wasn't selected for GSoC this year, I remain deeply interested >>> in contributing to OpenNLP — particularly the gRPC-based Python >> integration >>> project we previously discussed. I found the idea both technically >> exciting >>> and valuable for the community, and I’d still love to pursue it >>> independently. >>> >>> Currently, I’m in the middle of my 4th semester university exams, which >>> will conclude on *May 17*. I plan to begin contributing more actively >>> around *May 20* and would be truly grateful for your *guidance and >>> mentorship* throughout the process. >>> >>> Please let me know if you’re open to supporting this contribution and if >>> there are any specific expectations or directions I should follow to >> align >>> well with the project. >>> >>> Looking forward to your response. >>> >>> Warm regards, >>> *Jobin Sabu* >> >>
