Metadata-Version: 2.4
Name: word-segment-grpc-client
Version: 0.1.0
Summary: Word segment grpc service
Requires-Python: >=3.12
Description-Content-Type: text/markdown
Requires-Dist: coloredlogs>=15.0.1
Requires-Dist: grpcio-tools>=1.68.1
Requires-Dist: modelscope[framework]>=1.21.1
Requires-Dist: stanza>=1.10.1
Requires-Dist: torch>=2.5.1


## Development
Make sure you have installed `uv` and `protoc` and `pre-commit`
```bash
curl -LsSf https://astral.sh/uv/install.sh | sh

brew install protobuf

pip install pre-commit
```

### Setup
We use BERT based model to do word segment job, the F1 score is better than Jieba.
When we start to develop on this project, we can directly run below command.
```bash
make prepare_dev
```
The above command will set up virtual env for Python, and install the requirements.

### Download model
We can download the model from huggingface, or Kaggle.

https://huggingface.co/datasets/mikehug24/knowledge_graph_datasets/resolve/main/nlp_models.zst


### gRPC golang client
In order to get the gRPC client, we need to make the client package public, hence, we use git submodule
to manage it. Most of the time, we don't need to update the client.

## TODO
- Rename to NLPPythonService
