This is an example NER service using SpaCy and was built to show Data Scientists & Analysts a quick way to turn their models into deployable services.
Tools used: FastAPI, SpaCy, Docker, and Poetry.
git clone https://github.com/JayThibs/spacy-example-deployed-service
cd spacy-example-deployed-service
docker build -t spacy-example-deployed-service .
docker run -d -p 80:80 spacy-example-deployed-service:latest
Then, head to http://localhost/docs
to see the docs and test the app by clicking Try It Out, changing content
to whatever text you would like to get NERs from with SpaCy, and hitting Execute. Scroll down to find the output entities in the Response body.
On the Swagger UI, you can Try It Out here:
Once you've executed the request, you can see the output entities in the Response body:
You may need to update the main.py
file import the BaseModel
s. You simply need to remove the period from from .models import Payload, Entities
.
Then, you will need to do:
git clone https://github.com/JayThibs/spacy-example-deployed-service
cd spacy-example-deployed-service
pip install poetry # in case you don't have it
poetry install # install dependencies
uvicorn src.main:app --reload
Now, you can head to http://localhost:8000/docs
. Go to the previous section for help with the Swagger UI.
As a way to get better at quickly deploying useful ml / data science apps, I worked through this reference video walkthrough to get going quickly for future projects.