policies description in CLI

• docstring for each policy
• remove redundant variables
• test auto policy for a "big" source tree

In D6027#155638, @vlorentz wrote:

This diff touched the part of the code that needs to be documented, so it's fine to document it in the same diff

In D6027#155615, @vlorentz wrote:

Could you document the scan policies?

commit message

better comment wording

make CI pass

make CI pass

test scan policies SWHIDs request to the backend

rebase

commit message

make pytest pass

requested changes

rebase

requested changes

requested changes

unit test

In D5899#150867, @zack wrote:

LGTM in general, but needs unit tests (in addition to the two nitpicks above about docstrings)

requested changes

commit message

closed by https://forge.softwareheritage.org/D5825

landed in https://forge.softwareheritage.org/rDMODae50e43fe091d8cc7b4cfbe0eea17565b38dbb0d

rebase

requested changes

• missing incompatibility checks
• ignore --recursive if the input object is not a directory
• print only the swhid if using --no-filename

It seems it's still missing incompatibility checks with --no-dereference, --no-filename, and --type

Actually, --no-filename could be compatible (and implemented in this diff); should i include it?

• requested changes
• get the object path based on the data contained in the node object

In D5825#148519, @zack wrote:

In D5825#148516, @DanSeraf wrote:

2. If relevant, could you implement --verify too?

Sure, if it is useful i could open another diff for it

This is not going to be easy, because one would need to pass a set of SWHIDs, mapped to individual paths, which is much harder to do on a CLI than passing a single SWHID (as in the current implementation of --verify).
The only easy way to do this would be passing a filename pointing to, e.g., a JSON formatted manifest, to be compared with computed SWHIDs, but I don't see much the point of doing that.

I'm fine with --verify being incompatible with --recursive for the time being (maybe with an appropriate error message).

It will show only the SWHID of the given directory, basically the same process as before is applied

But what was the process before? Did it ignore directory entries?

It checks only the given directories generating a from_disk.Directory object for each directory. Should it uses the same logic used for the recursive option?

1. What happens if one runs it on a directory and --recursive isn't given?

It will show only the SWHID of the given directory, basically the same process as before is applied

• minimum required version for swh-model
• treat algo_min in the same way as other cases

Already implemented in D4193

removed unnecessary conversion

rebase

rebase