Dataflow Transfer

A Python application for automating the transfer of sequencing data.

Overview

Dataflow Transfer monitors sequencing run directories and orchestrates the transfer of sequencing data via rsync. It supports multiple sequencer types (Illumina, Oxford Nanopore and Element), tracks transfer progress in a CouchDB-based status database, and handles both continuous and final transfer phases during and after sequencing completion.

Supported Sequencers

• Illumina: NextSeq, MiSeqi100, NovaSeqXPlus (WIP), MiSeq (WIP)
• Oxford Nanopore (ONT): PromethION (WIP), MinION (WIP)
• Element: AVITI (WIP)

Installation

Requirements

• Python 3.11+
• Dependencies listed in requirements.txt:
  • PyYAML
  • click
  • xmltodict
  • ibmcloudant
• run-one

Setup suggestions

1. Clone the repository:

git clone <repository-url>
cd dataflow_transfer

2. Install the package:

pip install -e .

Or with development dependencies:

pip install -e ".[dev]"

Usage

Command Line Interface

dataflow_transfer [OPTIONS]

Options

• -c, --config-file PATH: Path to configuration YAML file. Defaults to ~/.df_transfer/df_transfer.yaml. Can also be set via TRANSFER_CONFIG environment variable.
• -r, --run RUN_ID: Transfer a specific run (e.g., 20250528_LH00217_0219_A22TT52LT4). Requires --sequencer.
• -s, --sequencer TYPE: Sequencer type of the run (e.g., NovaSeqXPlus, MiSeq, AVITI). Required with --run.
• --version: Show version and exit.

Examples

# Transfer all runs (uses configuration for sequencing directories)
dataflow_transfer

# Transfer a specific run
dataflow_transfer --run 20250528_LH00217_0219_A22TT52LT4 --sequencer NovaSeqXPlus

# Use a custom config file
dataflow_transfer --config-file /path/to/config.yaml

Configuration

Create a YAML configuration file with the following structure:

log:
  file: /path/to/dataflow_transfer.log

run_one_path: /usr/bin/run-one

transfer_details:
  user: username
  host: remote.host.com

statusdb:
  username: couchdb_user
  password: couchdb_password
  url: couchdb.host.com
  database: sequencing_runs

sequencers:
  NovaSeqXPlus:
    sequencing_path: /sequencing/NovaSeqXPlus
    miarka_destination: /Illumina/NovaSeqXPlus
    metadata_for_statusdb:
      - RunInfo.xml
      - RunParameters.xml
    ignore_folders:
      - nosync
    rsync_options:
      - --chmod=Dg+s,g+rw
  # ... additional sequencer configurations

How It Works

1. Discovery: Scans configured sequencing directories for run folders
2. Validation: Confirms run ID matches expected format for the sequencer type
3. Transfer Phases:
   • Sequencing Phase: Starts continuous background rsync transfer while sequencing is ongoing (when the final sequencing file doesn't exist). Uploads status and metadata files (specified for each sequencer type in the config with metadata_for_statusdb) to database.
   • Final Transfer: After sequencing completes (final sequencing file appears), initiates final rsync transfer and captures exit code.
   • Completion: Updates database when transfer was successful.

Status Tracking

Run status is tracked in CouchDB with events including:

Status	Meaning	Occurs when
sequencing_started	Sequencing is ongoing	A run folder exists but the final sequencing file has not been created yet
transfer_started	Intermediate transfer was initiated	Sequencing is ongoing and an rsync has been started
sequencing_finished	Sequencing has completed	A run folder exists and the final sequencing file has been created
final_transfer_started	Final sync has started	A run folder exists and the final sequencing file has been created, but the final rsync exit code file has not yet been created or contains a non-zero exit code
transferred_to_hpc	Transfer completed successfully	A run folder exists, the final sequencing file has been created, and the final rsync exit code file contains a 0 exit code

Flow chart

Key Components

• dataflow_transfer/run_classes/: Sequencer-specific run classes
• dataflow_transfer/utils/filesystem.py: File operations and rsync handling
• dataflow_transfer/utils/statusdb.py: CouchDB session management with retry logic
• dataflow_transfer/cli.py: Command-line interface

Assumptions

• Run directories are named according to sequencer-specific ID formats (defined in run classes)
• Final completion is indicated by the presence of a sequencer-specific final file (e.g., RTAComplete.txt for Illumina)
• Remote storage is accessible via rsync over SSH
• CouchDB is accessible and the database exists
• Metadata files (e.g., RunInfo.xml) are present in run directories for status database updates

Status Files

The logic of the script relies on the following status files:

• run.final_file - The final file written by each sequencing machine. Used to indicate when the sequencing has completed.
• final_rsync_exitcode - Used to indicate when the final rsync is done, so that the final rsync can be run in the background. This is especially useful for restarts after long pauses of the cronjob.

Development

Running Tests

pytest

With coverage:

pytest --cov --cov-branch

Code Quality

Run linting and formatting checks:

ruff check .
ruff format --check .

Project Structure

dataflow_transfer/
├── cli.py                 # Command-line interface
├── dataflow_transfer.py   # Main transfer orchestration
├── log/                   # Logging utilities
├── run_classes/           # Sequencer-specific run classes
├── utils/                 # Utility modules (filesystem, statusdb)
└── tests/                 # Unit tests

Adding a new sequencer

To add support for a new sequencer, add the following to dataflow_transfer:

1. Add a new class for the sequencer in one of the run classes files below. Make sure it inherits from the manufacturer class (IlluminaRun, ElementRun, ONTRun)
   • dataflow_transfer/run_classes/illumina_runs.py
   • dataflow_transfer/run_classes/element_runs.py
   • dataflow_transfer/run_classes/ont_runs.py
2. Import the new class in dataflow_transfer/run_classes/__init__.py
3. Add a test fixture for the new run in dataflow_transfer/tests/test_run_classes.py and include it in the relevant tests
4. Add a section for the sequencer in the config file