GPAS client

The command line interface and Python API client for the GPAS mycobacterial platform. The client enables secure sample upload with client-side human read removal, and enables programmatic retrieval of analytical outputs. For the best experience, a Linux or MacOS machine with 16GB of RAM is recommended.

Install

Installing Miniconda

If a conda package manager is already installed, skip to Installing the client, otherwise:

Linux

• In a terminal console, install Miniconda, following instructions and accepting default options:

  curl -O https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
  bash Miniconda3-latest-Linux-x86_64.sh
  

MacOS

The client requires an x86_64 conda installation. If your Mac has an Apple processor, disable or delete existing arm64 conda installations before continuing.

• If your Mac has an Apple processor, using Terminal, firstly run:

  arch -x86_64 zsh
  

• Install Miniconda using Terminal, following instructions and accepting default options:

  curl -O https://repo.anaconda.com/miniconda/Miniconda3-latest-MacOSX-x86_64.sh
  bash Miniconda3-latest-MacOSX-x86_64.sh
  

Installing or updating the client

• If using a Mac with an Apple processor, using Terminal, firstly run:

  arch -x86_64 zsh
  

• Perform the installation/upgrade:

  conda create --yes -n gpas -c conda-forge -c bioconda hostile
  conda activate gpas
  pip install gpas
  

• Test:

  gpas --version
  

Usage

Ensure that the conda environment is active:

conda activate gpas

Help

Run gpas --help for an overview of CLI subcommands. For help with a specific subcommand, use e.g. gpas auth --help

Authentication (gpas auth)

You will need to authenticate the first time you use the client. Do this by running gpas auth and entering your username and password. A token will be saved automatically inside your home directory.

gpas auth
Enter your username: bede.constantinides@ndm.ox.ac.uk
Enter your password: ***************

Uploading samples (gpas upload)

The upload subcommand performs metadata validation and client-side removal of human reads for each of your samples, before uploading sequences to the GPAS platform for analysis.

gpas upload tests/data/illumina.csv

During upload, a mapping CSV is created (e.g. a5w2e8.mapping.csv) linking your local sample names with their randomly generated remote names. Keep this file safe, as it is useful for downloading and relinking results later.

A 4GB human genome index is downloaded the first time you run gpas upload. If for any reason this is interrupted, simply run the upload command again. Upload will not proceed until the index has been downloaded and passed an integrity check.

To retain the decontaminated FASTQ files uploaded to GPAS, include the optional --save flag. To perform decontamination without uploading anything, use the optional --dry-run flag.

Downloading files (gpas download)

The download subcommand retrieves the output (and/or input) files associated with a batch of samples given a mapping CSV generated during upload, or one or more sample GUIDs. When a mapping CSV is used, by default downloaded file names are prefixed with the sample names provided at upload. Otherwise downloaded files are prefixed with the sample GUID.

# Download the main reports for all samples in a5w2e8.mapping.csv
gpas download a5w2e8.mapping.csv

# Download the main and speciation reports for samples in a5w2e8.mapping.csv
gpas download a5w2e8.mapping.csv --filenames main_report.json,speciation_report.json

# Download the main report for one sample
gpas download 3bf7d6f9-c883-4273-adc0-93bb96a499f6

# Download the main report for two samples
gpas download 3bf7d6f9-c883-4273-adc0-93bb96a499f6,6f004868-096b-4587-9d50-b13e09d01882

# Save downloaded files to a specific directory
gpas download a5w2e8.mapping.csv --out-dir results

# Download input files
gpas download a5w2e8.mapping.csv --inputs --filenames ""

Querying sample metadata (gpas query)

The query subcommand fetches either the processing status (gpas query status) or raw metadata (gpas query raw) of one more samples given a mapping CSV generated during upload, or one or more sample GUIDs.

# Query the processing status of all samples in a5w2e8.mapping.csv
gpas query status a5w2e8.mapping.csv

# Query the processing status of a single sample
gpas query status 3bf7d6f9-c883-4273-adc0-93bb96a499f6

# Query all available metadata in JSON format
gpas query raw a5w2e8.mapping.csv

Support

For technical support, please open an issue or contact support@gpas.global

Development

Development install

git clone https://github.com/GlobalPathogenAnalysisService/cli.git
cd cli
conda env create -y -f environment.yml
pip install --editable '.[dev]'

Updating

git pull origin main
gpas --version

Using an alternate host

1. The stateless way (use --host with every command):

   gpas auth --host dev.portal.gpas.world
   gpas upload samples.csv --host dev.portal.gpas.world
   

2. The stateful way (no need to use --host with each command):

   export GPAS_HOST="dev.portal.gpas.world"
   

   Then, as usual:

   gpas auth
   gpas upload samples.csv
   

   To reset:

   unset GPAS_HOST
   

Using a local development server

export GPAS_HOST="localhost:8000"
export GPAS_PROTOCOL="http"

To unset:

unset GPAS_HOST
unset GPAS_PROTOCOL

Releasing a new version

pytest
# Increment version strings inside src/gpas/__init__.py, Dockerfile
git tag 0.xx.0
git push origin main --tags
flit build
flit publish  # Uploads to PyPI given appropriate permission
# Announce in Slack CLI channel
# PR gpas/gpas/settings.py with new version