Add JSON filetype

[hashier]

I disable the filter for the comparison, since the key/secret would be filtered out otherwise.
Though, without the patch not every valid line inside JSON would be scanned.

# Current master

$ PYTHONPATH="" python3 -m detect_secrets scan test.json --disable-filter detect_secrets.filters.heuristic.is_not_alphanumeric_string | jq '.results'
{}

# My branch

$ PYTHONPATH="/src/detect-secrets/" python3 -m detect_secrets scan test.json --disable-filter detect_secrets.filters.heuristic.is_not_alphanumeric_string | jq '.results'
{
  "test.json": [
    {
      "type": "Secret Keyword",
      "filename": "test.json",
      "hashed_secret": "3c8567b928837d1264eec832a7e59fdcb9470f8b",
      "is_verified": false,
      "line_number": 2
    }
  ]
}

# file to be scanned, valid JSON

$ cat test.json
{
  "secret": 78193672857003048172616861202423,
}

[domanchi]

Do we need this?

I'm assuming that secrets in JSON files (found by the KeywordDetector) will be in string fields. aka.

{
    ...
    "secret": "blah"
}

By having FOLLOWED_BY_COLON_REGEX, we would also include things like numbers (which we want to ignore), and nested values (which is debatably useful, given that this is a line-based scanner).

[hashier]

Ah ok, if you want to ignore numbers then of course this isn't a rule you need.

I just followed more the "what is legal json" and added those rules.

I guess the PR without the FOLLOWED_BY_COLON_REGEX might still be useful, just to improve scanning times of JSON files.