Simplify attribution when using nexB-specific inventories

[JonoYang]

I am using the develop branch of aboutcode-toolkit.

It would be nice to have aboutcode toolkit generate attribution from a nexB inventory workbook without the use of vartext or aliases.

When I run about attrib <path to inventory workbook> <path to attribution report>, I get these errors:

Command completed with 2 errors or warnings.
ERROR: Field name: ['issue ref#', 'resource path', 'resource name', 'item ref', 'item type', 'component name', 'component version', 'package type', 'package namespace', 'package name', 'package version', 'license category(ies)', 'license name(s)', 'concluded license expression', 'concluded copyright holder', 'internal use only', 'analysis notes', 'homepage url', 'download url', 'license url', 'detected license expression', 'detected copyright'] contains illegal name characters (or empty spaces) and is ignored.
CRITICAL: Field name is required and empty

My workaround is to create another xlsx sheet where I:

• take the component name and package names and put them into the same column named name
• collect the component and package versions into the version column
• take the the concluded license expressions and rename the column license_expression
• take the concluded holders and rename the column copyright

[chinyeungli]

Here are couple of items to address:

1. For the renaming/key mapping process, we can use the transform utility to help (https://aboutcode-toolkit.readthedocs.io/en/latest/reference.html#transform)
2. The errors shown above (Both ERROR and CRITICAL) are rules in the SPEC:
   • ABCTK does not allow special characters (including space)
   • name is the "Required" field
3. You brought up a good sample of potential issue as AbcTK only has 1 name field while the input may contains more than 1 name field such as Component Name and Package Name. We cannot do something like the following config file with the transform utility

field_renamings:
	about_resource : 'Directory / Filename'
	name : `Component Name`
	name: 'Package Name'

because whatever come last will be used for the mapping. As a result, from the above config, the name will use the Package Name and whenever the Package Name is empty, it'll cause error as name is a required field. There are 2 solutions:

• As suggested in the ticket, manually created a new column called name and then copy the value from Component Name or Package Name so that user have the full control or what is the name to use
• Implement some logic to choose which should be the name used if there are 2 name (same field name) in the config file.

Feedback and sugggestions are welcome.

[chinyeungli]

I think the transform is the way for this issue. That's to create a sepecifc transform config file for the field renaming for the nexB-specific inventories.