Document ScanCode JSON ouput format

[AvishrantsSh]

The format in which scancode.io accepts the input from pipelines is not very clear. Right now, scancode.io is built to support output generated by scancode-toolkit. Also, it seems that scancode.io uses just a handful of those results, making it even harder to figure out. Besides this, other tools and pipelines may or may not be that descriptive in their output and some sort of data marshalling may be required.
So, I think that we should decide on a uniform and consistent structure that will be acceptible by scancode.io, which can serve as a template for other pipelines.

[AvishrantsSh]

A general outline of the structure can be

{
    "path": "<-path to file->",
    "type": "<-type->",
    "licenses": [
        {
            "key": "<-license key->",
            "score": "<-score-> (maybe optional?)",
            "name": "<-license name->",
            "category": "<-license category->",
            "start_line": "<-start_line->",
            "end_line": "<-end_line->"
        }
    ],
    "license_expressions": [
        "<--can be obtained from licenses too-->"
    ],
    "copyrights": [
        {
            "value": "<-copyright details->",
            "start_line": "<-start_line->",
            "end_line": "<-end_line->"
        }
    ],
    "holders": [
        {
            "value": "First",
            "start_line": "<-start_line-> (maybe optional?)",
            "end_line": "<-end_line-> (maybe optional?)"
        }
    ],
    "scan_errors": []
}

[pombredanne]

That's a good point, but that's more of a ScanCode-TK issue where we should document the format (as discussed in the AboutCode weekly this week).
So I am going top move this ticket to the SCTK.

[AyanSinhaMahapatra]

See also #2008 #2425