Some problems validating while checking "type", "pattern" and "format"

[EwanLanegan]

Hey @Julian ! Loving the module so far. Having some problems whilst trying to validate a large JSON (the code below is an example, not the real thing). I am trying to validate JSON data against the following schema:

{	
    "$schema": "http://json-schema.org/draft-04/schema#"
    "title": "schema",
    "type": "object",
    "properties": {
        "key1": {"type": "string"}, 
        "key2": {"type": "number"}, 
        "key3": {"type": "string, "format": "email"}, 
        "key4": {"type": "string", "pattern": "[0-9]{4}"} 
    },
    "required": ["key1", "key2", "key3", "key4"],
    "additionalProperties": false
}

The problem is that the validation method is allowing any type of data to be input, and so is outputting false trues. I have read the docs but I cannot find a clear way to validate the types, patterns (especially) and formats. I think the main parts I may need advice on is FormatChecker() and TypeChecker(), but to be honest I am not sure where the problem is.

def testvalidatetypes(event, context):
    s3 = boto3.client("s3")
    bucket = "ewanstestbucket"                      #this just collects the schema from a storage system
    key = "schema2.json"
    schema = s3.get_object(Bucket=bucket,Key=key)

    
    json_data = {
        "key1": "Hello world",
        "key2": 42,
        "key3": "[email protected]",
        "key4": 42                        #this data should cause an error - should be 4 numbers, and a string
    }
    
    checker = jsonschema.FormatChecker()
    validator = jsonschema.Draft4Validator(schema, format_checker=checker)
    
    try:
        validator.validate(json_data)
        print("User data is valid.")
    except:
        print("User data is invalid!")
        for error in sorted(validator.iter_errors(json_data), key=str):
            print(error.message)

The output for this is User data is valid.
The output I need is User data is invalid! (type of error, where etc.)

I am new to this, so I have probably made a few really silly mistakes. However, if you could give a clear explanation of how to check if each type is the correct type, pattern and format, that would be great! I will test whatever you suggest immediately and get back to you.
Thanks :)

[Julian]

Hi @EwanLanegan, glad you're finding this useful!

In general I'd ask that you use StackOverflow, IRC, or Slack for support and reserve this issue tracker for bugs (or suspected bugs).

It's also extremely helpful regardless to minimize your error into a SSCCE, and there are a large number of things here that I suspect you can remove while still getting your error (at which point it may even be obvious what the issue is).

Because this is already here and there are a few obvious things:

• Your schema is invalid JSON, it's missing things like a comma after the $schema value, an end quote on one of the "strings, etc. -- wouldn't recommend you hand-author JSON :), and certainly you need to validate it afterwards
• You're passing a string into jsonschema.Draft4Validator, when it takes a schema, i.e. a dict, so you need to be calling json.loads on that (which would have pointed out the first thing)
• Never use a bare except: :)

[EwanLanegan]

@Julian Thanks for pointing these issues out to me! One question: Does the TypeChecker method need to be a parameter within the validate method, or is it a separate thing?

[Julian]

TypeChecker is new in 3.0.0 (which is imminent...). It's an argument to jsonschema.validators.extend you'd use if you wanted to extend an existing validator to customize the corresponding Python types used.