Skip to content

Latest commit

 

History

History
327 lines (264 loc) · 15 KB

CONTRIBUTING.md

File metadata and controls

327 lines (264 loc) · 15 KB
Raw

Contributing to the MiCall Project

If you like this project and want to make it better, help out. You could report a bug, or pitch in with some development work.

Bug Reports and Enhancement Requests

Please create issue descriptions on GitHub. Be as specific as possible. Which version are you using? What did you do? What did you expect to happen? Are you planning to submit your own fix in a pull request?

Development

The easiest way to start developing MiCall is by using DevContainers.

  1. Open Project:

    • If you're using Visual Studio Code on your local machine, open the MiCall project folder and select the "Reopen in Container" prompt to initialize the DevContainer environment. Make sure you have the necessary DevContainer extension installed beforehand, available here.
    • For a web-based development environment, you can develop directly on GitHub using GitHub Codespaces by navigating to the MiCall repository on GitHub and selecting "Code" > "Open with Codespaces" to launch a pre-configured environment.

  2. Dependency Installation: All required dependencies will be automatically installed whether you are using a local DevContainer or GitHub Codespace.

  3. Interpreter Selection:

    • Following the setup, if you encounter dependency errors, manually select the Python interpreter version 3.8. This is necessary because the container includes multiple Python versions, and the default selection might not be appropriate.
    • This issue is tracked and further details can be found here.

  4. Verification: To ensure that the environment is correctly configured, execute pytest within the DevContainer or Codespace. All tests should pass, indicating that the setup is successful.

Local install

To see how all the tools should be installed, follow the steps in Dockerfile and dev.dockerfile. If you prefer, you can run your development environment under docker, as described in dev.dockerfile. The same installation steps are also listed in the Singularity file. The docker file runs with debian, and the singularity file runs with CentOS.

If you want to see what's currently being worked on, check out the active tasks in our milestones.

Python

Check that Python is already installed.

python --version

We have tested with Python 3.11.

BaseSpace

Set up the native apps virtual machine. Make sure you have a developer account on illumina.com. The first time you run this, you will have to log in to your account using

sudo docker login docker.illumina.com

Test data

If you want to run micall/monitor/micall_watcher.py, you have to set up data folders for raw data and for the working folders. You'll also need to set up the QAI project and the MiseqQCReport so you can download QC data and upload results.

  1. Create a data folder somewhere on your workstation, like ~/data. Create subdirectories called miseq and RAW_DATA. Add folders RAW_DATA/MiSeq/runs.
  2. Connect to the shared drive using CIFS and mount smb://192.168.68.144/RAW_DATA as /media/RAW_DATA.
  3. Navigate down to /media/RAW_DATA/MiSeq/runs, pick a recent folder, and make sure it has a file named needsprocessing.
  4. Copy SampleSheet.csv to a sample run folder under your local RAW_DATA/MiSeq/runs folder.
  5. Navigate down to Data\Intensities\BaseCalls, and copy a few of the .fastq.gz files to your sample run folder under Data/Intensities/BaseCalls.
  6. Copy the Interop folder and the files RunInfo.xml and runParameters.xml.
  7. Open settings.py for editing.
  8. Point home at your local data/miseq folder.
  9. Point rawdata_mount at your local RAW_DATA folder.
  10. Set the Oracle connection information to a test database where you can upload sequence data. (https://git-int.cfenet.ubc.ca/wscott/oracleserver)
  11. Run the Ruby console for QAI and LabMiseqRun.import('01-Jan-2000') for the date of your sample run.
  12. Upload the projects to a micall pipelines in QAI, use micall.utils.projects_upload to create a new pipeline in QAI
  13. Run micall/monitor/micall_watcher.py, it does need arguments. Look up the container app ids from Kive, check the Kive server URL and ports as well as QAI server and port

Looking at SAM files

When you don't understand the pipeline's output, it can be helpful to look at the raw reads in a sequence viewer like Tablet. Run the micall_docker script on a run folder or a single sample, like this:

python micall/utils/micall_docker.py folder --debug_remap --all_projects --keep_scratch /path/to/run

The options tell it to write the debug files, use all projects, and save the scratch folder that holds all the debug files. Look through the scratch folders under the run folder to find the one for the sample you're interested in. The remap step writes the mapping results as debug_remapX_debug.sam and debug_remapX_debug_ref.fasta, where X is the remapping iteration number. You should be able to open an assembly in Tablet using those two files. If the SAM file contains multiple regions, you'll probably have to sort it with the micall/utils/sort_sam.py script. That same script can convert prelim.csv into a SAM file.

If you want to understand the de novo assembly process, read through the assembly page.

GitHub Web Site

Most of the time, you can change the web site content just by editing the markdown files in the docs folder. However, you may occasionally need to dig into the page templates or do more serious work. If that happens, you can test out the web site locally before publishing it.

  1. Install Ruby 2.6, preferably with Ruby Version Manager.

    rvm install 2.6
rvm use 2.6

  2. Install the gems for the web site.

    cd MiCall/docs
gem install bundler
bundle install

  3. Serve the web site.

    bundle exec jekyll serve

What changes might you want to make? The web site is based on the Bulma Clean Theme, so read through the documentation there to see if it already has the feature you want. Usually, the advanced features require you to write files in the docs/_data folder or add settings to the front matter at the top of a markdown file.

If you have to add a new feature to the web site, you can override one of the files in the theme by copying it into the docs/_includes folder, and making changes there. Consider offering it back to the theme project as a pull request, because any files you override won't get automatic improvements from the original theme project.

Releases

This section assumes you already have a working server up and running, and you just want to publish a new release. If you're setting up a new server, follow similar steps to setting up a development workstation. Follow these steps:

  1. Check that all the issues in the current milestone are closed, and make sure the code works in your development environment. Run all the unit tests as described above, process the microtest data set with micall/utils/release_test_microtest.py.

  2. Check if the kiveapi package needs a new release by looking for new commits. Make sure you tested with the latest version.

  3. Determine what version number should be used next.

  4. Use the micall/utils/projects_dump.py script for the previous version and compare projects.json to check that the projects match, or that the differences were intended. Test the micall/utils/projects_upload.py script with your updated project files in your local test QAI.

  5. Check the history of the HIV and HCV rules files in the micall/resistance folder. If they have changed, create a new display file in the docs folder and upgrade the version numbers in the genreport.yaml file and asi_algorithm.py.

  6. Check the history of the micall.alignment folder. If it has changed since the last release, then update the version number in setup.py.

  7. Update the change notes in the Singularity file, and commit those changes.

  8. Create a release on Github. Use "vX.Y.Z" as the tag, where X.Y matches the version you used in QAI. If you have to redo a release, you can create additional releases with tags vX.Y.Z.1, vX.Y.Z.2, and so on. Mark the release as pre-release until you finish deploying it.

  9. Rebuild the Singularity image, and upload it to your local Kive server. Process the microtest data.

  10. Upload the Singularity image to the Kive test server, and record the ids of the new apps.

  11. Process all the samples from test_samples.csv on the Kive test server, and run the micall_watcher service on a VirtualBox. Use the micall/utils/release_test_*.py scripts to compare the results of the new release with the previous version. Also run the internal scripts miseq_gen_results.rb and miseq_compare_results.rb to look for differences. Get the comparison signed off to begin the release process.

  12. Upload the Singularity image to the main Kive server, and record the id of the new apps.

  13. Upload the pipeline definitions to QAI, using the micall/utils/projects_upload.py script. There is no need to create the new pipeline version in QAI beforehand, the script will do this for you - just remember to update the Order by field afterwards.

  14. Stop the micall_watcher service on the main Kive server after you check that it's not processing any important runs.

    ssh user@server
tail /var/log/micall/micall.log
sudo systemctl stop micall_watcher

  15. Get the code from Github into the server's environment.

    ssh user@server
cd /usr/local/share/MiCall
git fetch
git checkout tags/vX.Y.Z

  16. Look for changes in micall/monitor/micall_watcher.py's parse_args() function. Either look at the blame annotations at the link above, or review the changes in the new release. If there are new or changed settings, adjust the configuration in /etc/systemd/system/micall_watcher.service or /etc/micall/micall.conf.

  17. Update the container app ids and pipeline version number in /etc/systemd/system/micall_watcher.service. If you change the configuration, reload it:

    sudo systemctl daemon-reload

  18. Check that the kiveapi package is the same version you tested with. If not, do a Kive release first.

    cd /usr/local/share/Kive
/usr/local/share/venv-micall/bin/pip show kiveapi
cat api/setup.py

  19. Start the micall_watcher service, and tail the log to see that it begins processing all the runs with the new version of the pipeline.

    sudo systemctl start micall_watcher
sudo systemctl status micall_watcher
tail -f /var/log/micall/micall.log

    If the log doesn't help, look in /var/log/messages on CentOS or /var/log/syslog on Ubuntu.

  20. Launch the basespace virtual machine (see BaseSpace section above) and copy MiCall source files into it. The easiest way to copy is via scp:

    scp -P 2222 /path/to/micall/on/host basespace@localhost:MiCall
# (password is "basespace")

    Then login to virtual machine and build the docker image:

    ssh basespace@localhost -p2222     # (password is "basespace")
sudo python3 MiCall/micall/utils/docker_build.py -t vX.Y.Z --nopush

    The script is able to push the docker image to the illumina repo and launch spacedock as well, but that is currently broken because of the old docker version in the VM. If this is ever updated, or we build our own VM, you won't have to do these steps manually anymore and can remove the --nopush.

  21. Tag the same docker image and push it to docker hub and illumina. Unfortunately, the old version of docker that comes with the basespace virtual machine can't log in to docker hub or illumina, so you'll have to save it to a tar file and load that into your host system's version of docker. Before pushing it anywhere, check that the docker image works by running the microtests. If the docker push fails with mysterious error messages (access to the resource is denied), try docker logout and docker login again, and make sure you are on the owner team of cfelab on docker hub.

    ssh basespace@localhost -p2222  # password is "basespace"
sudo su
sudo docker save cfelab/micall:vX.Y.Z > micall-vX.Y.Z.tar
exit  # Exit the root shell.
exit  # Exit the virtual machine.
sudo docker load < micall-vX.Y.Z.tar
sudo docker login docker.illumina.com
sudo docker tag docker.illumina.com/cfe_lab/micall:vX.Y.Z cfelab/micall:vX.Y.Z
sudo docker push docker.illumina.com/cfe_lab/micall:vX.Y.Z
rm micall-vX.Y.Z.tar

  22. Duplicate the MiCall form in the revisions section of the form builder, then edit the callbacks.js in the form builder itself, and add the :vX.Y.Z tag to the containerImageId field. In My Apps, create a new version of the App with the new version number. Record the new agent ID (click the arrow on the bottom right of the form builder).

  23. Launch the spacedock version by running this in your basespace VM:

    sudo spacedock -a [agent ID] -m https://mission.basespace.illumina.com

  24. Check that the new MiCall version works as expected by processing some of the microtests in BaseSpace.

  25. Activate the new revisions in the form builder and the report builder.

  26. Submit the new revision of the MiCall app for review, and ask our contact at Illumina to set the price to zero for the list of test users.

  27. Send an e-mail to users describing the major changes in the release.

  28. Close the milestone for this release, create one for the next release, and decide which issues you will include in that milestone.

  29. When the release is stable, check that it's included on the Zenodo page. If you included more than one tag in the same release, the new tags have not triggered Zenodo versions. Edit the release on GitHub, copy the description text, update the release, then click the Delete button. Then create a new release with the same description, and that will trigger a Zenodo version.