Python: using JSON file for increasing program configurability

As the experience tells us, wrong decisions in program design and life will usually bite back hard. In order to avoid the most obvious traps leading into horrific maintenance problems, we should always design our programs to be free of any hard-coded parameters. By using configuration scheme presented in this post, flexible programs, which can use any desired set of input configurations, can be created. This means we can (as an example) execute a specific program (for valuing batch of transactions) several times, but using different set of configurations (different set of market data) for each execution. All example files can be downloaded from my GitHub page.

In this very simple example, Python program will just print a set of market and fixings data from CSV files, based on a given set of configurations. In order to keep this example program short and sweet, our JSON configuration file has only two configurations: directory addresses for market and fixings data CSV files, as follows.

{
  "MARKETDATA":"/home/mikejuniperhill/Market.csv",
  "FIXINGSDATA":"/home/mikejuniperhill/Fixings.csv"
}

Directory address of this configuration file will be given as a command line argument for the program. Based on this given configuration, program will then read configured data from files to be used in program. 

Example program is shown below. In the first stage, program will create configurations object. Technically, this object is just a wrapper for dictionary data structure. Any specific configuration can be accessed by using Python version of index operator overloading. After this, program reads data from configured CSV files into DataFrame objects and prints their contents to terminal.

import json
import sys
import pandas

# class for hosting configurations
class Configurations:
    inner = {}
    # read JSON configuration file to dictionary
    def __init__(self, filePathName):
        self.inner = json.load(open(filePathName))
        self.inner = {k.upper(): v for k, v in self.inner.items()}
    # return value for a given configuration key
    # 'overload' indexing operator
    def __getitem__(self, key):
        return self.inner[key.upper()]

# configuration file string is command line argument
configurationsFilePathName = sys.argv[1]

# create configurations object
config = Configurations(configurationsFilePathName)

# create market data based on configuration
market = pandas.read_csv(config['MarketData'])
print('EUR swap curve:')
print(market.head())

# create fixings data based on configuration
fixings = pandas.read_csv(config['FixingsData'])
print('6M Euribor fixings:')
print(fixings.head())

Handy tool for constructing and testing syntactic correctness of any JSON file can be found in here. Finally, thanks for reading.
-Mike