Service to create text embeddings via Tensorflow Universal Sentence Encoder
To manual run embedder service:
- Unzip universal-sentence-encoder_4.tar.gz to
models/1 - Build docker image with
docker build -t embedder embedder - Run docker image with
docker run -t -p 5000:5000 -v "{PROJECT_PWD}/models/1:/models/use" -t embedder
You can find the Swagger docs on http://localhost:5000/apidocs
To create embedding for the sentence "How many Hindu nations are in the world?" you can run command:
curl -X POST "http://localhost:5000/v1/embedder/generate" -H "accept: application/json"\
-H "Content-Type: application/json" -d "{ \"text\": \"How many Hindu nations are in the world?\"}"
Also, probes are available:
- /readiness - Readiness check endpoint. Make test response to model.
- /liveness - Health check endpoint
Streamlit application to find nearest StackOverflow question.
The application needs ElasticSearch index to search by.
To install ElasticSearch locally in single node mode use:
docker run --name es01 -p 9200:9200 -p 9300:9300 -e "discovery.type=single-node" -t elasticsearch:8.4.3
Specify user credentials in .env file and create the index from pre-computed embeddings with the command:
python src/index/indexer_elastic.py --config_path params.yaml
Run streamlit app with streamlit run search_app/search_companies.py
You can modify application params in search_app/params.yaml if you need