[3][Performance] Query airtable by city

[kevinjcliao]

Context

• We want to make the website faster. Before thinking about caching, one thing we can do is reduce the payload that we are requesting from AirTable.

This PR

• Rather than requesting data from all hospitals at once, this PR queries AirTable by each city instead.

Performance Measurement

Methodology

By inserting time.time() calls before and after each request, we can time how long it's taking for our server to respond to requests to /government_paid_hospitals.

# pyre-fixme[56]: Decorator async types are not type-checked.
@app.route("/government_paid_hospitals")
async def government_paid_hospitals() -> wrappers.Response:
    start_time = time.time()
    city: str = request.args.get("city")
    data = await get_hospitals_from_airtable("", city)
    response = app.response_class(
        response=json.dumps(data),
        status=200,
        mimetype="application/json",
    )
    end_time = time.time()
    print(f"Execution time {end_time - start_time}")
    return response

Time (seconds)

	Test 1	Test 2	Test 3	Test 4	Test 5	Average
Before	13.18s	13.04s	14.93s	13.02	13.72	13.59
After	4.83s	6.08s	4.25s	2.72	2.95	4.17s

Total request response time improvement: 13.59-4.17 = 9.42 seconds. (3.25x speedup)

Test Plan

• [X] yarn lint, yarn tc and yarn test are passing

• [X] Website loads successfully.

• [X] Website can switch between cities

https://user-images.githubusercontent.com/8745371/126029568-e06820be-36a1-4cb1-a3cf-283b82f2c607.mp4

• [X] Website can switch between counties

https://user-images.githubusercontent.com/8745371/126029606-40ba0e7b-5c50-4894-83a7-aef7d461a588.mp4