Music-Genre Dependence of Graph-Related Metrics in a Music Credits Network
We published some of the results in the article ``Citation is not Collaboration: Music-Genre Dependence of Graph-Related Metrics in a Music Credits Network'' at the 20th Sound and Music Computing Conference.
@inproceedings{clerici2023citation,
author = {Clerici, Giulia and Tiraboschi, Marco},
title = {{Citation is not Collaboration: Music-Genre Dependence of Graph-Related Metrics in a Music Credits Network}},
booktitle = {Proceedings of the 20th Sound and Music Computing Conference},
year = {2023},
series = {SMC},
address = {Stockholm, Sweden},
month = {6},
organization = {Sound and Music Computing Network},
}
The plots for the genre-wide comparison of the centrality measures can be browsed at chromaticisobar.github.io/CitationIsNotCollaborationStats.
Documentation is generated with sphinx and is available at limunimi.github.io/CitationIsNotCollaboration
Some dependencies require conda
.
The following instructions assume that you are working from the root directory of the repository
There seems to be some issues with conda when trying to install too many packages.
So we spit the installation requirements in chunks.
Create the environment
conda create -p ./venv python=3.8
Activate the enviroment
conda activate ./venv
Install packages in chunks
conda install -c chromaticisobar --file requirements-0of6.txt -y && \
conda install -c conda-forge --file requirements-1of6.txt -y && \
conda install -c conda-forge --file requirements-2of6.txt -y && \
conda install -c conda-forge --file requirements-3of6.txt -y && \
conda install -c conda-forge --file requirements-4of6.txt -y && \
conda install -c conda-forge --file requirements-5of6.txt -y
You can add extra development dependencies using the requirements files.
conda install --file <FILE> -c conda-forge
test-requirements.txt
: Tests dependenciesdocs-requirements.txt
: Docs generation dependenciesstyle-requirements.txt
: Style check dependenciesnotebooks-requirements.txt
: Notebooks dependencies
To add the featgraph
package, run
conda develop .
Interactive notebooks are in the notebooks folder
Some functionalities are available to the command-line
You can convert the original pickled dataset in BVGraph binary files and plain-text metadata with
python -m featgraph.conversion
usage: python -m featgraph.conversion [-h] [--jvm-path PATH] [-l LEVEL] [--tqdm] adjacency_path metadata_path dest_path
Convert original pickled dataset into text and BVGraph files
positional arguments:
adjacency_path The path of the adjacency lists pickle file
metadata_path The path of the metadata pickle file
dest_path The destination base path for the BVGraph and text files
optional arguments:
-h, --help show this help message and exit
--jvm-path PATH The Java virtual machine full path
-l LEVEL, --log-level LEVEL
The logging level. Default is 'INFO'
--tqdm Use tqdm progress bar (you should install tqdm for this)