re-work haddock3 installation

[rvhonorato]

This PR drastically changes the haddock3 installation.

We no longer use submodules. For fcc the code has already been aded to the project (#1019) and for fastrmsd-matrix @mgiulini gave us permission to copy the source here, thanks!

This external code is built with a CustomBuild in setup.py.

Another main pain point is CNS, here I've added logic to determine the system's arch and download a provided pre-compiled binary - keep in mind that the installation will fail for unsupported architecture. I understand there might be possible to have a workflow without CNS after #921, the installation should then be reviewed.

To test it, follow these commands:

# Make a fresh clone and checkout this branch
$ git clone https://github.com/haddocking/haddock3 haddock3-installation-test
$ cd haddock3-installation-test
$ git checkout installation

# Setup the python environment
$ python -V # make sure you are on python3.9
Python 3.9.16
$ python -m venv .venv
$ source .venv/bin/activate

# With conda:
## $ conda create -n haddock3-installation-test python=3.9
## $ conda activate haddock3-installation-test

# Install haddock3
(.venv) $ pip install .

# Run one example
(.venv) $ cd examples/docking-protein-protein
(.venv) $ haddock3 docking-protein-protein-test.cfg
# This should finish without issues 🎉 

@haddocking/haddock-developers please look at this one very carefully

[AnnaEngel98]

Tried this in a conda python 3.9 environment.
After pip install . I get the following error :

ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
matplotlib 3.9.0 requires pyparsing>=2.3.1, which is not installed.
matplotlib 3.9.0 requires numpy>=1.23, but you have numpy 1.22.2 which is incompatible.

Running the example then also does not work

[rvhonorato]

Tried this in a conda python 3.9 environment.
After pip install . I get the following error :

ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
matplotlib 3.9.0 requires pyparsing>=2.3.1, which is not installed.
matplotlib 3.9.0 requires numpy>=1.23, but you have numpy 1.22.2 which is incompatible.

Running the example then also does not work

I cannot replicate this, and seems like there's an issue with your package management, as written in the error message above - so it's probably related to that

ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.

"Works on my machine" both with native python and anaconda

[rvhonorato]

For prosperity: we found an error with the maximum path size that CNS allows - @amjjbonvin managed to patch it