Python CLI to create a JSON diff using the patchdiff package

json-diff.py

#!/usr/bin/env python

import rich_click as click
import json
from pathlib import Path
from dotenv import load_dotenv
from sys import stdout
from patchdiff import diff as patch_diff

load_dotenv()

def load_jsonl(jsonl_file):
    return [json.loads(line) for line in jsonl_file]

def key(model):
    return f"{model['model']['name']}-{model['location']}"

def index(models):
    return dict(map(lambda model: (key(model), model), models))

def make_patch(previous, next):
    ops, _ = patch_diff(previous, next)

    # Convert Path objects to strings so that they can be JSON dumped
    for op in ops:
        op["path"] = str(op["path"])

    return ops if len(ops) > 0 else None

def diff(previous, next):
    a = set(previous.keys())
    b = set(next.keys())
    common = a & b
    removed = a - b
    added = b - a
    common_models = [(key, make_patch(previous[key], next[key]), previous[key]) for key in common]
    return {
        'updates': dict([(key, {"previous": prev, "changes": diff}) for (key, diff, prev) in common_models if diff]),
        'removed': {key: previous[key] for key in removed},
        'added': {key: next[key] for key in added}
    }

@click.command()
@click.option('--previous-file', '-p', help='Previous file', type=click.File('r'))
@click.option('--next-file', '-n', help='Next file', type=click.File('r'))
@click.option('--diff-file', '-d', help='Diff file', type=click.File('w'), default=stdout)
def report(previous_file, next_file, diff_file):
    click.echo(f"Creating diff report between {previous_file.name} and {next_file.name}")
    previous = index(load_jsonl(previous_file))
    next = index(load_jsonl(next_file))
    d = diff(previous, next)

    json.dump(d, diff_file, indent=2)

if __name__ == '__main__':
    report()