On repo, docs and Zenodo add license and citations for data we are using

small text files giving the values of each CV fold if it was able to be done. If not, maybe return a message that says CV couldn't be completed

Security issue with pickle

• #8
• #9
• #10
• #11
• #12
• #13
• #14
• #15
• #16
• #17
• #18
• #19
• #20
• #21
• #22
• #23
• #24

Potential options

• Prompt for acknowledgment (but probably not good for use cases such as job submissions)
• CLI option for always overwriting
• Create new result dir, append some suffix, warn to log about saving to a slightly different dir

create branch that contains slurm script for testing all the combinations of options

• #3
• #4

Use lowercase argument name, probably won't fix

• GenePlexus API (#45)
• Data downloader module
• GenePlexus CLI and tutorial (#41)
• Utilities
• References (#52)

After the first release is ready, create a release v1.0.0 and then deployt to PyPI

python3 setup.py -q sdist bdist_wheel 
python3 -m pip install twine
twine upload --skip-existing dist/*

@ChristopherMancuso

https://github.com/krishnanlab/GeneplexusPublic/blob/8b0fb8ceba5585dfaa5db1758b09b996e0229ad3/geneplexus/util.py#L182-L183

Security issue with pickle

Set log level

Nice lib to use: https://pypi.org/project/pystow/

Great suggestion from Pat:

set default data folder in user's home dir? at one point, I'd cd'd into the result folder, than ran the program again and was surprised to see the downloads starting over. I'd forgotten I was in a different folder. If the input data will never change across any run of PyGenePlexus, then will a user ever want to specific a different data directory? That is, could you create a data directory in the users home folder that would be re-used for all runs of geneplexus? Other software uses dot-files to store configuration or data like this, e.g. .ipython or .java or .keras Could the data directory by default be $HOME/.geneplexusdata and the user would rarely specify it? I think dot-directories for configuration would also work in Windows (but would have to check). To see existing dot-directories on your computer (or HPCC) use ls -ald $HOME/.* Another possibility is to set an env variable like $GENEPLEXUSDATA (this is a lot of work for you but would mean the user would not have to specify the data dir every time, or could run the program in different places (*e.g. home, shared, etc) without specifying a data dir.

Use lowercase function name, probably won't fix

possible add a warning or note about this is for a research project so if user needs data in many spots please copy data locally. Could add this to

• repo README
• jupyter notebook tutorial
• message when package is installed or a certain function is run

Currently, if a user wants to use GenePlexus with their custom network or gsc, they will need to set up the required custom files using the geneplexus.custom module first, before they can proceed to run the GenePlexus pipeline using the CLI.

The goal here is to make a CLI option that calls the necessary geneplexus.custom functions to set up custom files, and thus eliminates the need for one to manually prepare them.

Working notes

• --custom option -> enables preprocessing custom network/gsc data
  • Preprocessing runlong also save to ${data_dir}/custom_logs/${net}_${feature}_${gsc}.log
    • Network stats: num_nodes, num_edges
    • GSC stast: num_genesets, med_size, avg_size, std_size, max_size, min_size
• Required files
  • Edgelist_xxx.edg (custom network)
  • GSCOriginal_xxx.json (custom gsc)
• Set up custom network and gsc
  • custom.edgelist_to_node -> NodeOrder_${net}.txt
  • custom.edgelist_to_matrix -> Data_${feature}_{network}.npy
  • custom.subset_gsc_to_network -> GSC_${gsc}_${net}_GoodSets.json, GSC_${gsc}_${net}_universe.txt

make a new branch an speed test downloading all the data from Azure

change the example and defaults to match PCD example on web server

• function that takes edge list to numpy array and node order text file
• add original GSCs to downloadable files
• function that takes a network and original GSC and return a subset GSC specific to the network
• add info on exact filename and data format conventions for user added files

e.g. -z- for --zip_output

after pip install in a new environment and trying to download data from Azure got error massage "no module named requests"

https://github.com/krishnanlab/GeneplexusPublic/blob/8417ddc1815375f3eadecfccca2e59af3a6f7896/geneplexus/geneplexus.py#L25

vs.

https://github.com/krishnanlab/GeneplexusPublic/blob/8417ddc1815375f3eadecfccca2e59af3a6f7896/geneplexus/geneplexus.py#L43

• #44
• Make repo public
• Check RTD build