JSON to JSON diff tool

##Introduction

A python module for checking equivalence and difference of two JSONs with regex support

This repository will treat two types of JSON necessary for our purpose:

1. Regular JSON where only data is stored

2. Model JSON which will treat all keys and values as regular expressions

   • Models will be used for comparing JSON objects to A model for equivalence with regex matching
   • As JSON can be in a non-deterministic order, this is harder than doing a straight string comparison

###There are four modes of this module

1. Regular JSON comparison
2. JSON to Model comparison
3. Regular JSON diff
4. JSON to Model Diff

##1. Regular JSON comparison This simply tests whether two files contating json are the same

./json_diff.py path/to/file1.json path/to/file2.json

Returns True if equal or False otherwise.

.json_diff.py path/to/file1.json path/to/json_directory/

Returns the filename of the first match if there is one, or False otherwise.

##2. JSON to Model comparison This mode tests whether a given JSON file matches a model when regular expressions are evaluated.

./json_diff.py --mode j2m path/to/pure.json path/to/model.json

Returns True if the JSON and model match, False otherwise.

./json_diff.py --mode j2m path/to/pure.json path/to/model_directory

Returns the filename of the first match if there is one, or False otherwise.

##3. Regular JSON diff This mode computes a file diff between two JSON files.

The logic flow of this code is loosely based off of code found in https://github.com/monsur/jsoncompare

####Assuming old.json:

{
    "accounting" : [ {
        "firstName" : "John",
        "lastName"  : "Doe",
        "age"       : 23
    }, {
        "firstName" : "Mary",
        "lastName"  : "Smith",
        "age"       : 31
    } ],
    "sales" : [ {
        "firstName" : "Sally",
        "lastName"  : "Green",
        "age"       : 27
    }, {
        "firstName" : "Jim",
        "lastName"  : "Galley",
        "age"       : 41
    } ]
}

and new.json:

{
    "accounting" : [ {
        "firstName" : true,
        "lastName"  : "Doe",
        "age"       : 23
    }, {
        "firstName" : "Susan",
        "lastName"  : "Smith",
        "age"       : 31
    } ],
    "sales" : [ {
        "firstName" : "Sally",
        "lastName"  : "Green",
        "size"       : 27
    }, {
        "firstName" : "Jim",
        "age"       : 41
    } ]
}

Then

./json_diff.py -d new.json old.json

Should produce output like:

TypeDifference : accounting[0].firstName - is bool: (True), but was unicode: (John)
Changed: accounting[1].firstName to Susan from Mary
+: sales[0].size =27
-: sales[0].age=27
-: sales[1].lastName=Galley

##4. JSON to Model Diff This mode computes a file diff between a JSON file and a model with regular expressions.

Note: at this time, we only support regular expression matching on singleton values (that is values that are NOT a list or dictionary)

####Assuming new.json:

{
    "accounting" : [ {
        "firstName" : "John",
        "lastName"  : "Doe",
        "age"       : 23
    }, {
        "firstName" : "Mary",
        "lastName"  : "Smith",
        "age"       : 31
    } ],
    "sales" : [ {
        "firstName" : "Sally",
        "lastName"  : "Green",
        "age"       : 27
    }, {
        "firstName" : "Jim",
        "lastName"  : "Galley",
        "age"       : 41
    } ]
}

and model.json:

{
    "accounting" : [ {
        "lastName"  : "Doe",
        "age"       : 23
    }, {
        "firstName" : "Mary",
        "lastName"  : "Smith",
        "age"       : "[0-9]+"
    } ],
    "sales" : [ {
        "firstName" : "Sally",
        "lastName"  : "(.*)",
        "age"       : 24
    }, {
        "firstName" : "Jim",
        "lastName"  : "Galley",
        "age"       : 41
    } ]
}

Then

.json_diff.py -d --mode j2m new.json model.json

Should produce output

+: accounting[0].firstName=John
Changed: sales[0].age to 27 from 24

As you can see we find a match on numbers even though the type is 'int' (Mary Smith's age). Also, even though we changed Sally's age, we were still able to find a match on her entry with a regular expression representing her last name.