Allocation of external dependencies & extra suggestions

[JuanmaMedina]

Here is a report of some difficulties / indications that could be taken into account. I apologize in advance if they seem too obvious:

1. To properly get louvain working, the user should move to the working directory (i.e. the folder with all meta3cbox dependencies) and create a folder named tools, where louvain should be downloaded and compiled manually:

mkdir tools
cd tools
wget https://sourceforge.net/projects/louvain/files/louvain-generic.tar.gz
tar -xzvf louvain-generic.tar.gz
# Important to change the default name of the folder, otherwise it will not be recognized in the 2nd step (partition)
mv louvain-generic louvain
cd louvain
make
cd ..
rm louvain-generic.tar.gz

2. A similar procedure should be followed to allocate the HMMs that will be used in the 3rd step (annotation). From the working directory:

wget dl.pasteur.fr/fop/5eHgTGww/modele_HMM.tar.gz
tar -xzvf modele_HMM.tar.gz
rm modele_HMM.tar.gz

3. I encountered a permission error when accessing the distance file. Maybe this was just my case, but a full access to this file should be granted (chmod) before running the 2nd step (partition)

4. I encountered some errors when trying to re-run some of the individual steps, having to delete some of the previously generated files/folders, such as config_current.sh, config.sh or output. I did not really understand what was happening here, maybe I was not doing something right.

And I think this is it. I hope this is useful!

Best,
Juanma

[baudrly]

Hello, thanks again for the feedback!

To properly get louvain working, the user should move to the working directory (i.e. the folder with all meta3cbox dependencies) and create a folder named tools, where louvain should be downloaded and compiled manually
A similar procedure should be followed to allocate the HMMs that will be used in the 3rd step (annotation)

This can be done automatically by running ./meta3c.sh dependencies, I changed what it does as you suggested.

I encountered a permission error when accessing the distance file. Maybe this was just my case, but a full access to this file should be granted (chmod) before running the 2nd step (partition)

I added a quick check before it runs the distance executable.

I encountered some errors when trying to re-run some of the individual steps, having to delete some of the previously generated files/folders, such as config_current.sh, config.sh or output. I did not really understand what was happening here, maybe I was not doing something right.

Yes, there's a trade-off to be had between 'I don't want to re-run everything every time I launch the pipeline' and 'I don't want broken/incomplete intermediary files to bog down the pipeline and break it inexplicably'. Feel free to open new issues and report what you encountered. Even though these shouldn't happen normally, it can help mitigate the awkwardness for future users.

As for config issues, you can generate a fresh new config.sh file from the template by using the --reset option.