Full documentation for the competition, including much of the information here, can be found online at openproblems.bio/neurips_docs/. The documentation for Viash is available at viash.io/docs.
This starter kit assumes you have Bash, Java >= 8.0, and Docker installed.
Run ./generate_submission.sh and submit your results to eval.ai by uploading the submission.zip file (easiest) or using the evalai-cli (recommended).
In order of relevance:
config.vsh.yaml Metadata of your method.
script.* A script containing your code.
generate_submission.sh A helper script for running your method on the input datasets
and generating a submission file to upload to eval.ai.
sample_data/ A sample dataset for testing and debugging your code.
nextflow.config A config file to ensure the nextflow pipeline will be able to
find your method.
.gitignore A simple gitignore file.
If successful, running the ./generate_submission.sh script will generate the following files / folders:
submission.zip The submission files to be uploaded to eval.ai.
target/docker/ A standalone command-line script for running your method
using a Docker backend.
target/nextflow/ A NextFlow module for using your method as part of a
NextFlow pipeline.
output/predictions/ The predictions made by your method.
work/ Temporary data created by the NextFlow pipeline.
.nextflow* Nextflow execution output.
You can run the code manually on the sample dataset as follows:
$ viash run config.vsh.yaml -- \
--input_mod1 sample_data/openproblems_bmmc_multiome_starter/openproblems_bmmc_multiome_starter.mod1.h5ad \
--input_mod2 sample_data/openproblems_bmmc_multiome_starter/openproblems_bmmc_multiome_starter.mod2.h5ad \
--output test_output.h5adTip: You can also omit the config.vsh.yaml in the above command.
Tip #2: Run viash run -- --help to view the command-line interface of your component.
You can adapt the script however you like. All the code between the ## VIASH START and ## VIASH END code blocks automatically
gets removed by viash and can be used for debugging your script.
Take a look at the config.vsh.yaml. It contains information on which parameters your component has, and which package dependencies
are required in order to build a Docker container for your component.
Tip: After making changes to the components dependencies, you will need to rebuild the docker container as follows:
$ viash run -- ---setup cachedbuild[notice] Running 'docker build -t method:dev /tmp/viashsetupdocker-method-tEX78c'
Tip #2: You can view the dockerfile that Viash generates from the config file using the ---dockerfile argument:
$ viash run -- ---dockerfileFROM dataintuitive/randpy:py3.8
RUN pip install --upgrade pip && \n pip install --no-cache-dir "anndata" "umap-learn"
What if running your method on the sample data works but fails when applied to the submission datasets? Given the following output:
$ ./generate_submission.sh...
[78/9f8fc2] NOTE: Process `method:method_process (dyngen_atac_disconnected_mod2)` terminated with an error exit status (127) -- Error is ignored
[20/330b8a] NOTE: Process `method:method_process (dyngen_atac_bifurcating_converging_mod2)` terminated with an error exit status (127) -- Error is ignored
Completed at: 05-Aug-2021 12:13:39
Duration : 1m 32s
CPU hours : 0.3 (100% failed)
Succeeded : 0
Ignored : 32
Failed : 32
You can still submit your solutions to eval.ai but your submission may not execute properly.
You can check out what went wrong by looking at the tag for the process that failed (e.g. 20/330b8a):
$ ls -1a work/20/330b8a52b1a71489e7c53c66ef686e/.command.begin
.command.err
.command.log
.command.out
.command.run
.command.sh
dyngen_atac_bifurcating_converging_mod2.censor_dataset.output_mod1.h5ad
dyngen_atac_bifurcating_converging_mod2.censor_dataset.output_mod2.h5ad
.exitcode
You can view the stdout and stderr of this process by viewing the .command.log and .command.err files respectively. You can also try to run your component manually by editing the VIASH START and VIASH END code block in your script (see below) and running through the code manually.
## VIASH START
$par_block_starter
input_mod1 = "work/20/330b8a52b1a71489e7c53c66ef686e/dyngen_atac_bifurcating_converging_mod2.censor_dataset.output_mod1.h5ad",
input_mod2 = "work/20/330b8a52b1a71489e7c53c66ef686e/dyngen_atac_bifurcating_converging_mod2.censor_dataset.output_mod2.h5ad",
output = "debug.h5ad",
# ... other parameters
)
## VIASH ENDTo update the starter kit to the latest release, simply download the corresponding zip file from the latest GitHub release and unzip it somewhere. After unzipping, copy any of the files which you have changed in the new directory, usually config.vsh.yaml and script.R/py.
├── LICENSE
├── Makefile <- Makefile with commands like `make data` or `make train`
├── README.md <- The top-level README for developers using this project.
├── data
│ ├── external <- Data from third party sources.
│ ├── interim <- Intermediate data that has been transformed.
│ ├── processed <- The final, canonical data sets for modeling.
│ └── raw <- The original, immutable data dump.
│
├── docs <- A default Sphinx project; see sphinx-doc.org for details
│
├── models <- Trained and serialized models, model predictions, or model summaries
│
├── notebooks <- Jupyter notebooks. Naming convention is a number (for ordering),
│ the creator's initials, and a short `-` delimited description, e.g.
│ `1.0-jqp-initial-data-exploration`.
│
├── references <- Data dictionaries, manuals, and all other explanatory materials.
│
├── reports <- Generated analysis as HTML, PDF, LaTeX, etc.
│ └── figures <- Generated graphics and figures to be used in reporting
│
├── requirements.txt <- The requirements file for reproducing the analysis environment, e.g.
│ generated with `pip freeze > requirements.txt`
│
├── setup.py <- makes project pip installable (pip install -e .) so src can be imported
├── src <- Source code for use in this project.
│ ├── __init__.py <- Makes src a Python module
│ │
│ ├── data <- Scripts to download or generate data
│ │ └── make_dataset.py
│ │
│ ├── features <- Scripts to turn raw data into features for modeling
│ │ └── build_features.py
│ │
│ ├── models <- Scripts to train models and then use trained models to make
│ │ │ predictions
│ │ ├── predict_model.py
│ │ └── train_model.py
│ │
│ └── visualization <- Scripts to create exploratory and results oriented visualizations
│ └── visualize.py
│
└── tox.ini <- tox file with settings for running tox; see tox.readthedocs.io
Project based on the cookiecutter data science project template. #cookiecutterdatascience