A baseline implementation using Python/FastAPI to handle CRUD operations against a Mongo database.

One key challenge is that Mongo's default primary key fields are named _id but FastAPI/Pydantic does not permit field names in a model to begin with underscores. There was much discussion of this problem in online forums, and several partially-described solutions; thankfully the official Mongo documentation was recently updated to describe a proper solution in full.

This demonstration puts it all together to create a base repository that can be used to quickly spin up APIs in the future.

• Perform Create, Read, Update and Delete on a model stored in a MongoDB collection.
• Validate fields using standard Pydantic model definitions.
• Connect to a Mongo database defined in .env.
• Authenticate all endpoints against a single secret Bearer Token defined in .env.
• Internally map the Mongo _id field to id in code and API calls.
• Cleanly separate code for each endpoint's router and model to enable easy future expansion.
• Automatically generate complete endpoint documentation via SwaggerUI.

This application has been tested with the following stack pre-installed:

• Ubuntu 20
• Python 3.6
• Mongo 4.4.15
• python3-pip

• Clone this repo and cd into the fastapi-mongo-demo folder.
• Run pip install -r requirements.txt to install the required Python modules.

• Copy or rename the file .env-sample in the application base folder to .env.
• Edit the file .env in the application base folder.
  • Change the MONGO_URI value to be the connection string for a running Mongo database.
  • Change the MONGO_DBNAME value to be the desired name of your Mongo database.
  • Change the API_KEY value to be a strong random token string known only to you.
  • Save the file changes.

• From the application base folder, run the Uvicorn server:

  python3 -m uvicorn main:app --reload --host="0.0.0.0" --port=8000

• In a browser visit http://<server_ip_address>/docs where <server_ip_address> is the IP address of the server where this application is installed.

• Confirm you see documentation describing the avaiable API endpoints.

• This demo provides CRUD endpoints for an endpoint called /posts for theoretical posts in a blog, which has the following model definition:

   id: ObjectID
   uri: str
   title: str
   excerpt: str
   body: str
  

• To create an additional API endpoint, copy and modify /models/pydantic/post.py and /routers/post.py and use these as a base to define a model and routes specific to your new endpoint.

Optionally, you may run a test script called located in the root folder of the application by executing this on the command line: python3 tests.py

This script runs a sequence of GET, POST, PUT and DELETE calls against the /posts endpoint and ensures the expected values are returned for each.