Update to upstream version '6.6.2'
with Debian dir e5c2f227d7e6087b3152d6160e9502f4199a6105

Previously when looking up data by key in an ImmutableDict, the inner
tuple storing keys and values was iterated until finding the requested
key.

This is not really efficient when the ImmutableDict contains a lot of
entries, typically for an origin snapshot containing a lot of branches.

So use an inner dictionary to speedup look up by key operations and
improve loader performances.

This fixes python 3.7 support due to poetry, a dependency of isort, that
removed support for that Python version in a recent release.

In order to remove warnings about /apidoc/*.rst files being included
multiple times in toc when building full swh documentation, prefer to
include module indices only when building standalone package documentation.

Also include them the proper sphinx way.

Related to T4496

Update to upstream version '6.6.1'
with Debian dir 5264b8650def64f97ea2b7aae15cadc2dc496623

There were two issues that was preventing to browse some SWHIDs given as
examples in that documentation:

- Some sphinx links were broken in rDMODe1c3fe80731226618616117dfd67a95f3d365645

- A SWHID with ';' in its path qualifier was correctly percent escaped but
  when used as URL argument an extra percent escaping is required as HTTP
  server will unescape URL arguments and thus break SWHID percent escaping.

Closes T4721

from_disk.Content object created for a symlink was missing path info
so ensure to add it for consistency with from_disk.Content object
created for a regular file.

- pre-commit from 4.1.0 to 4.3.0,
- codespell from 2.2.1 to 2.2.2,
- black from 22.3.0 to 22.10.0 and
- flake8 from 4.0.1 to 5.0.4. Also freeze flake8 dependencies.

Also change flake8's repo config to github (the gitlab mirror
being outdated).

Update to upstream version '6.6.0'
with Debian dir 18c244cdb4270a0e59b359f3522dd0a7f1bf55f8

When using attr < 21.3.0, adding field transformer breaks attrs
integration with hypothesis, because attributes transformed with
such function are not casted to generated AttrsClass, but remains
just an list of attributes. This causes error in hypothesis by
raising an AttributeError.

As we use attr 21.2.0 in production and when building debian buster
package, add a workaround for that issue as explained here:
https://github.com/python-attrs/attrs/issues/821.

Previously the MerkleNode.collect method was returning a dict whose keys
are node types and values dict of {<node_hash>: <node_data>}.

In order to give more flexibility to client code for the processing of
collected nodes, prefer to simply return a set of MerkleNode.

As a consequence, MerkleNode objects need to be hashable by Python so
the __hash__ method has also been implemented.

Closes T4633

Update to upstream version '6.5.1'
with Debian dir a243e7419a0f035452acb893131c246c9f49946a

It exists use cases where sha512 checksums need to be computed (content
integrity checks for instances) so add sha512 in the list of hashing
algorithms supported by the MultiHash class.

The pattern were validated from $PWD and later applied on path relative
to `root_path`. So we shuffle a bit of code to test them againt
root_path. We make the absolute pattern relative in the same go.

This code is coming from swh-scanner and should probably get an
overhaul, how ever for now we start with making it no broken.

Update to upstream version '6.5.0'
with Debian dir 58d5f805181a3d5452f6da0fec2d4dd6cec629b6

This reduce the number of function call and should be faster.

The mashup of blind optimisation in the previous changeset yield some
interesting results in total.

It would be insightful to measure them individually, but that would
take more time than we currently have.

When testing all the validator changes on our previous "benchmark" we
see quite interesting improvement.

    swh loader run mercurial https://foss.heptapod.net/mercurial/mercurial-devel directory=/data/repos/mercurial-devel

= Median time of 3 run =
base:   17 minutes 48 seconds
before: 11 minutes 50 seconds
after:  11 minutes 11 seconds

On a profile of the same run, the `to_model` call of the from_disk's `Directory` class took the following percentage:
base:   43%
before: 15%
after:  11%

(This commit is actually doing two things /o\)

- we inline the type-checking in the custom validator to reduce the
  number of function call.

- we optimize some of the custom validator by skipping the creation of
  intermediate tuples.

Since all `generic_type_validator` are optimized away, the code will no
longer be called. So we remove that code to avoid any drifting.

A nice "exception" is provided in case this start getting called again
in the future.

Since try/except context are known to be expensive in Python, it seems
useful to remove them.