Overview

Running

To run a pipeline, the run() function is used. The only required parameter is the path to the pipeline file:

>>> pipeline = nextflow.run("pipelines/my-pipeline.nf")

This will return an Execution object, which represents the pipeline execution that just took place (see below for details on this object). You can customise the execution with various options:

>>> execution = pipeline.run("my-pipeline.nf", run_path="./rundir", output_path="./outputs", params={"param1": "123"}, profiles=["docker", "test"], version="22.0.1", configs=["env.config"], timezone="UTC", report="report.html", timeline="timeline.html", dag="dag.html")

  • run_path - The location to run the pipeline from, which by default is just the current working directory.

  • output_path - The location to store the execution outputs (work etc.), which by default is the run_path.

  • params - A dictionary of parameters to pass to the pipeline as command. In the above example, this would run the pipeline with --param1=123.

  • profiles - A list of Nextflow profiles to use when running the pipeline. These are defined in the nextflow.config file, and can be used to configure things like the executor to use, or the container engine to use. In the above example, this would run the pipeline with -profile docker,test.

  • version - The version of Nextflow to use when running the pipeline. By default, the version of Nextflow installed on the system is used, but this can be overridden with this parameter.

  • configs - A list of config files to use when running the pipeline. These are merged with the config files specified in the pipeline itself, and can be used to override any of the settings in the pipeline config.

  • timezone - A timezone to pass to Nextflow - this determines the timestamps used in the log file.

  • report - A filename for a report file to generate. This will be an HTML file containing information about the pipeline execution.

  • timeline - A filename for a timeline file to generate. This will be an HTML file containing a timeline of the pipeline execution.

  • dag - A filename for a DAG file to generate. This will be an HTML file containing a DAG diagram of the pipeline execution.

Custom Runners

When you run a pipeline with nextflow.py, it will generate the command string that you would use at the command line if you were running the pipeline manually. This will be some variant of nextflow run some-pipeline.nf, and will include any parameters, profiles, versions, and config files that you passed in.

By default, nextflow.py will then run this command using the standard Python subprocess module. However, you can customise this behaviour by passing in a custom ‘runner’ function. This is a function which takes the command string and submits the job in some other way. For example, you could use a custom runner to submit the job to a cluster, or to a cloud platform.

This runner function is passed to the run() method as the runner parameter:

>>> execution = pipeline.run("my-pipeline.nf", runner=my_custom_runner)

Once the run command string has been passed to the runner, nextflow.py will wait for the pipeline to complete by watching the execution directory, and then return the Execution object as normal.

Polling

The function described above will run the pipeline and wait while it does, with the completed Execution being returned only at the end.

An alternate method is to use run_and_poll(), which returns an Execution object every few seconds representing the state of the pipeline execution at that moment in time, as a generator:

for execution in pipeline.run_and_poll(sleep=2, run_path="./rundir", params={"param1": "123"}):
    print("Processing intermediate execution")

By default, an Execution will be returned every second, but you can adjust this as required with the sleep paramater. This is useful if you want to get information about the progress of the pipeline execution as it proceeds.

Executions

An Execution represents a single execution of a pipeline. It has properties for:

  • identifier - The unique ID of that run, generated by Nextflow.

  • started - When the pipeline ran (as a Python datetime).

  • finished - When the pipeline completed (as a Python datetime).

  • duration - how long the pipeline ran for (if finished).

  • status - the status Nextflow reports on completion.

  • command - the command used to run the pipeline.

  • stdout - the stdout of the execution process.

  • stderr - the stderr of the execution process.

  • log - the full text of the log file produced.

  • return_code - the exit code of the run - usually 0 or 1.

  • path - the path to the execution directory.

It also has a process_executions property, which is a list of ProcessExecution objects. Nextflow processes data by chaining together isolated ‘processes’, and each of these has a ProcessExecution object representing its execution. These have the following properties:

  • identifier - The unique ID generated by Nextflow, of the form xx/xxxxxx.

  • process - The name of the process that spawned the process execution.

  • name - The name of this specific process execution.

  • status - the status Nextflow reports on completion.

  • stdout - the stdout of the process execution.

  • stderr - the stderr of the process execution.

  • started - When the process execution ran (as a Python datetime).

  • started - When the process execution completed (as a Python datetime).

  • duration - how long the process execution took in seconds.

  • return_code - the exit code of the process execution - usually 0 or 1.

  • path - the local path to the process execution directory.

  • full_path - the absolute path to the process execution directory.

  • bash - the bash file contents generated for the process execution.

Process executions can have various files passed to them, and will create files during their execution too. These can be obtained as follows:

>>> process_execution.input_data() # Full absolute paths
>>> process_execution.input_data(include_path=False) # Just file names
>>> process_execution.all_output_data() # Full absolute paths
>>> process_execution.all_output_data(include_path=False) # Just file names

Note

Nextflow makes a distinction between process output files which were ‘published’ via some channel, and those which weren’t. It is not possible to distinguish these once execution is complete, so nextflow.py reports all output files, not just those which are ‘published’.