Frictionless

- Local development has been migrated to using [Hatch](https://hatch.pypa.io/latest/)

- Rebased packaging on PEP 621
- Extracted experimental application/server from the codebase

- Implemented "Metadata.from_descriptor(allow_invalid=False)" (1501)

- Various architectural and standards-compatibility improvements (minor breaking changes):
- Added new Console commands:
- list
- explore
- query
- script
- convert
- publish
- Rebased Console commands on Rich (nice output in the Console)
- Fixed `extract` returning the results depends on the source type (now it's always a dictionary indexed by the resource name)
- Enforced type safety -- many tabular command will be marked as impossible for non-tabular resources if a type checker is used
- Improved `frictionless.Resource(source)` guessing abilities; if you just like to open a table resource use `frictionless.resources.TableResource(path=path)`

- Implemented Implemented `catalog/dataset/package/resource.deference` (1451)

- Various architectural and standards-compatibility improvements (minor breaking changes):
- Improved type detection mechanism (including remote descriptors)
- Added `resources` module including `File/Text/Json/TableResource`
- Deprecated `resource.type` argument -- use the classes above
- Changed `catalog.packages[]` to `catalog.datasets[].package`
- Made `resource.schema` optional (`resource.has_schema` is removed)
- Made `resource.normpath` optional (`resource.normdata` is removed)
- Standards-compatability improvements: profile, stats
- Renamed `system/plugin.select_Check/etc` to `system/plugin.select_check_class/etc`