compare-gobjects.py

#!/usr/bin/env python

import sys
import os
import json

paths = sys.argv[1:]

ifiles = [ file( path ) for path in paths ]

datasets = [ ifile.read() for ifile in ifiles ]

jsons = [ json.loads( data ) for data in datasets ]

allkeys = set()
for i in range( len( jsons ) ):
    js = jsons[i]
    keys = js.keys()
    allkeys = allkeys.union( set( keys ) )
    print "File: ", i
    print "Number objects: ", len( keys )
    print

print "Total number keys: ", len( allkeys )

missingkeys = set()

for i in range( len( jsons ) ):    
    js = jsons[i]
    keys = js.keys()
    print "File: ", i
    diffcount = 0
    for key in allkeys:
        if key not in keys:
            missingkeys = missingkeys.union( set( [ key ] ) )
            print key
            print
            diffcount += 1
    print "diffcount = ", diffcount
    print

print
print "Missing objects:"
for key in missingkeys:
    print "key = ", key
    for js in jsons:
        if key in js:
            print js[key]
            break
    print