Improve unitpackage functionality

[DunklesArchipel]

Currently, unitpackage is limited to CSV data, while frictionless also accepts other formats.
Also, CSVs are loaded as pandas DataFrames upon loading the collection, which is not convenient, when you only want to load and filter datapackges.

The changes are pushed to [resource_entries](the https://github.com/echemdb/unitpackage/tree/resource_entries) branch. Individual pull requests to the branch can fail. In the first iteration, most of the functionality of the unitpackage module should work the same as before.

Mandatory changes

• change module structure (see Collection and entries consist of resources instead of packages #59)
• adapt to pixi (as in Add pixi setup to replace environment.yml svgdigitizer#220)
• clean code (see Collection and entries consist of resources instead of packages #59)
• adapat doctsrings (see Collection and entries consist of resources instead of packages #59)
• adapt documentation(see (see Adapt documentation to changes of the "entry_to_resource" branch #61)
• include staged PRs (such as Add extending existing dfs with columns and units #55)
• Add news for these changes (see Collection and entries consist of resources instead of packages #59)

Further steps

• allow loading different file formats
• Create Collection from either Data Packages within a folder by providing the folder path or directly from a single file containing a Data Package with multiple resources(JSON). (see Collections are created from Data Packages #72) has now collection.from_local(datadir) or collection.from_local_file(filename -> JSON)