subdash / AlphaVantage-API

Python interface for AlphaVantage API

AlphaVantage API Microservice

The goal here was to create a microservice for the AlphaVantage API which provides data on stocks that can be used for data-oriented analysis as well as visual analysis of stock data.

This project demonstrates creating a Python interface for a web API as well as some simple SQL integration for caching.

Discussion

1. Microservice vs. AlphaVantage API

There is a single endpoint for the AlphaVantage API which accepts various query parameters, one of which is always FUNCTION, so it works more like an RPC API than a REST one.

So you may be wondering, what does this microservice provide that the API itself does not?

Well, it provides code completion when writing queries in an IDE, validation of query params, and caching.

What is missing from this microservice? Well, not every single function has been implemented, mostly just the ones that are flagged as "high usage" on the API documentation.

2. Caching

The caching mechanism here is very simple. The service only has one function: get. Three things can happen when this function is called:

- If the url has never been queried before, a new entry is made in the cache table

- If it has been queried before and was cached recently, the data will be returned from the cache

- If it has been queried before but was not cached recently, the entry in the cache is updated

As you can see the cache performs simple CRUD operations to reduce the number requests to the AlphaVantage API.

Usage

This microservice can be used in the Python shell, in scripts, and as a module for a larger application. There is a short example script below.

#!/usr/bin/env python3
# Script example
API_KEY = os.environ["API_KEY"]
service = AlphaVantageService()
ticker = "RPM"
earnings_history_payload = {
    "function": AVFunction.EARNINGS,
    "symbol": ticker,
    "apikey": API_KEY
}
res = service.get(earnings_history_payload)
print(res)
""""
Output:
{
  "symbol": "RPM",
  "annualEarnings": [
    {
      "fiscalDateEnding": "2021-02-28",
      "reportedEPS": "2.88"
    },
    ...
  ],
  "quarterlyEarnings": [
    {
      "fiscalDateEnding": "2021-02-28",
      "reportedDate": "2021-04-07",
      "reportedEPS": "0.38",
      "estimatedEPS": "0.2907",
      "surprise": "0.0893",
      "surprisePercentage": "30.719"
    },
    ...
  ]
}"""

An example of using this module for visual stock analysis in the Python shell (reference tools/compare_earnings.py):

$ ./src/tools.py -k $(cat api_key) -t EBF --cmpearnings
Ticker: EBF
Earnings will be announced: 2021-04-19
Earnings information for last 4 quarters:
===========================================
	yearly EPS average: 0.27
===========================================
	reportedDate: 2020-12-20
	reportedEPS: 0.32
	surprisePercentage: 14.2857
===========================================
	reportedDate: 2020-09-21
	reportedEPS: 0.25
	surprisePercentage: 13.6364
===========================================
	reportedDate: 2020-06-22
	reportedEPS: 0.16
	surprisePercentage: 33.3333
===========================================
	reportedDate: 2020-04-20
	reportedEPS: 0.33
	surprisePercentage: -5.7143

Next quarter expected earnings:
	0.27
Expected earnings / last quarter earnings:
	0.84
Expected earnings / last year avg. earnings:
	1.02

Improvements

This project could be improved in many ways. The cache database schema is inefficient and does not use relations between tables as you normally would with a relational database.

More tools could be added for analysis, and converting the JSON data into stock charts would make this a lot more useful.

Support for different types of SQL databases would be a nice addition: the service could accept a config file for your local SQL database.

Unit tests could (honestly, should) be added.