Improve identification of conda package files

[mjherzog]

Working with SCTK v32.3.1 (running in SCIO v34.9.3), SCTK does not currently identify the installed files for a conda package in the Resources for_packages field. This data seems to be readily available in a set of .json files located under /conda-meta/ directory where conda is installed - typically opt/conda. The file names are in the format <package name>-<package-version>.json
This pattern is present for both Anaconda and miniconda distributions.

[AyanSinhaMahapatra]

Before scanning a docker image: docker://continuumio/miniconda3 resulted in:

218 packages
388 dependencies
18355 files

• 9213 files in a package
• 9142 files not in a package

With #4089 above:

295 packages
388 dependencies
18355 files

• 16469 files in a package
• 1886 files not in a package

So we should do much better conda resource assigning with this PR merged and released to SCIO

[mjherzog]

Excellent

[simrancharde]

Before scanning a docker image: docker://continuumio/miniconda3 resulted in:

218 packages 388 dependencies 18355 files

• 9213 files in a package
• 9142 files not in a package

With #4089 above:

295 packages 388 dependencies 18355 files

• 16469 files in a package
• 1886 files not in a package

So we should do much better conda resource assigning with this PR merged and released to SCIO

how to solve #4083 issue? we need to add scanner.py fike for_package and then parse it in json file ? anyrhing else

[AyanSinhaMahapatra]

@simrancharde thanks for your interest, but this already has a fix at #4089, could you check out our open good first issues instead: https://github.com/aboutcode-org/scancode-toolkit/issues?q=is%3Aissue%20state%3Aopen%20label%3A%22good%20first%20issue%22, this is where we need help mostly.