The purpose of this repo is to use embedding and Qdrant running locally as an API
- Run the following command:
docker-compose up
Please note that the first run will take a while, as there is a few gigabytes of data to download. Also, the first upsert to Qdrant will take a while, as the model has to be downloaded and the index needs to be built.
- Open
http://localhost:5000/
in your browser - Set up your environment by clicking "Initialize Qdrant" button
- You can now use "upsert" to add new embeddings to the database, "test search" to search for the nearest neighbors of a given embedding. Please note that you should provide UUID in the ID field.
ID
- UUID of the embedding (required)phrase
- phrase to embed and upsert to Qdrant (required)
search_term
- phrase to embed and search for nearest neighbors (required)
phrase
- phrase to embed (required)
When you change your requirements.txt
file, you'll need to rebuild your Docker image to install the new Python packages.
Here are the steps:
-
Stop the running Docker containers with the following command:
docker-compose down
-
Then rebuild and start your Docker containers:
docker-compose up --build