TUTORIAL_4.md

back to intro

Creating the API app

This part builds a simple FastAPI app with two endpoints that interface with the async workers and also the database.

Our FastAPI app will have two endpoints:

• GET / - returns the latest data record
• POST /refresh - triggers a fetch of new data in the background

The api base brick

We can do everything we want within one brick. We will make this a base (as it is dealing with the outside world) and call it api.

poetry poly create base --name api

Our code follows standard FastAPI patterns. We import two other bricks: the database connection and the component that can start fetch jobs.

# bases/mousetrap/api/core.py

import fastapi
from mousetrap.mongo_connection import open_db
from mousetrap.start_fetch_worker import start_fetch_job

app = fastapi.FastAPI()


@app.get("/")
def get_latest_data_record() -> dict:
    """Gets the latest processed data record."""

    db = open_db()

    record = db.results.find_one(sort=[("_id", -1)])

    if record:
        # Drop the Mongo ID field
        if "_id" in record:
            del record["_id"]
        return record

    else:
        # Empty database
        return {}


@app.post("/refresh")
def refresh_data() -> dict:
    """Starts a background data refresh job."""

    start_fetch_job()

    return {"status": "ok"}

As with the other components, we also need to expose the app object in the __init__.py file.

And we need to add this base component to the root pyproject.toml file.

Building the Poetry package

We now have all component and base bricks we need.

We now have the following packages in our root pyproject.toml file:

packages = [
    {include = "mousetrap/job_queue", from = "bases"},
    {include = "mousetrap/mongo_connection", from = "bases"},
    {include = "mousetrap/api", from = "bases"},
    {include = "mousetrap/fetch_data", from = "components"},
    {include = "mousetrap/process_data", from = "components"},
    {include = "mousetrap/start_fetch_worker", from = "components"},
    {include = "mousetrap/run_fetch_worker", from = "components"},
    {include = "mousetrap/run_processing_worker", from = "components"},
]

Using the CLI

We can use the CLI to confirm that our bricks seem to be connected correctly by using

poetry poly info

The output of this is a grid showing the bricks used in the different projects.

Everything looks correct.

---

Next we will assemble the bricks together into project bricks that would be deployed as Docker containers in our setup.

next