Support storing scancode-config.yml files in the scanned codebase

[pombredanne]

We need to allow storing scancode-config.yml files in the scanned codebase and used when running a pipeline.

There are many small edge cases to support:

• when there multiple scancode-config.yml in multiple input archives (multiple "codebases")
• when there are multiple scancode-config.yml inside an input archive
• when there is a direct input scancode-config.yml and other scancode-config.yml in input archives

For now we should support only these cases:

1. If there is a direct input scancode-config.yml file provided, it should be taken and have precedence and all other scancode-config.yml ignored.
2. If there is a single archive input and it has a scancode-config.yml file at the root, it should be used
3. If there are any other scancode-config.yml files elsewhere or more than one file in a single or multiple inputs, we should just report an error, and also in any other cases.

See also:

• Enhance and create detailed documentation for scancode-config.yml #1197

[pombredanne]

And also update the doc accordingly

[pombredanne]

@tdruez and I had a chat earlier today to discuss the topic of recursive extraction which is a config option.
If the scancode-config.yml is inside an input archive we would to extract it first to determine how we should extract. This is chicken and egg problem.

recursive extraction is the default and rarely to never makes sense to disable, therefore we are going remove recursive extraction as a scancode-config.yml option, and make this issue go away. recursive extraction will always take place.

Therefore we are going to have this revised and refined behavior:

1. If there is a direct input scancode-config.yml file provided, it should have precedence and any other scancode-config.yml found in the codebase are ignored. Same if the config options were provided in the UI.
2. Then if there is a single archive input and it has a scancode-config.yml file at the root, it should be used.
3. Then if there are any other scancode-config.yml files elsewhere or more than one file in a single or multiple inputs, we should just report a warning stating that we ignore these config files.

[tdruez]

Implemented in #1243