Python API

The arelle.api module is the supported method for integrating Arelle into other Python applications.

Warning

Arelle uses shared global state (PackageManager, PluginManager) which is NOT thread-safe. Only ONE Session can run at a time across the entire process.

Safe usage:

• Use one Session at a time per process

• Use a process pool instead of thread pool for parallelism

Unsafe usage:

• Running multiple Sessions concurrently in any threads

• Threading.Thread with Session.run()

Session

The Arelle Python API provides Session to run Arelle and access output. You can import it with:

from arelle.api.Session import Session

From there you can configure the session, run Arelle, and retrieve the generated models and logs (see examples below).

Examples

Creating an iXBRL Viewer

with open(samples_zip_path, 'rb') as stream:
    options = RuntimeOptions(
        entrypointFile=str(target_path),
        internetConnectivity='offline' if arelle_offline else 'online',
        keepOpen=True,
        logFile="logToStructuredMessage",
        logFormat="[%(messageCode)s] %(message)s - %(file)s",
        logPropagate=False,
        pluginOptions={
            'saveViewerDest': str(viewer_path),
            'viewer_feature_review': True,
        },
        plugins='ixbrl-viewer',
    )
    # Plugin default options haven't been applied yet.
    assert not hasattr(options, 'viewerURL')
    with Session() as session:
        session.run(
            options,
            sourceZipStream=stream,
            logFilters=[log_filter],
        )
        # Plugin default options were applied.
        assert hasattr(options, "viewerURL")
        assert options.viewerURL.endswith("ixbrlviewer.js")
        log_xml = session.get_logs('xml')

Querying a Model

options = RuntimeOptions(
    entrypointFile=str(report_zip_path),
    disclosureSystemName='esef-2024',
    internetConnectivity='offline',
    keepOpen=True,
    logFile=str(arelle_log_file),
    logFormat="[%(messageCode)s] %(message)s - %(file)s",
    packages=package_paths,
    plugins='validate/ESEF',
    validate=True,
)
target_qname = qname('https://xbrl.ifrs.org/taxonomy/2022-03-24/ifrs-full', 'Equity')
with Session() as session:
    session.run(options)
    model_xbrls = session.get_models()
    for model_xbrl in model_xbrls:
        equity_facts: set[ModelFact] = model_xbrl.factsByQname.get(target_qname, set())
        for equity_fact in equity_facts:
            print(f'Found Equity fact with value: {equity_fact.xValue}')
    log_xml = session.get_logs('xml')

Using a Validation Plugin

options = RuntimeOptions(
    entrypointFile=str(report_zip_path),
    disclosureSystemName='esef-2024',
    internetConnectivity='offline',
    logFile=str(arelle_log_file),
    logFormat="[%(messageCode)s] %(message)s - %(file)s",
    packages=package_paths,
    parameters="authority=SE",
    plugins='validate/ESEF',
    validate=True,
)
with Session() as session:
    assert not hasattr(options, "saveTargetInstance"), "validate/ESEF plugin dependency inlineXbrlDocumentSet CLI args should not be loaded yet."
    session.run(options)
    log_xml = session.get_logs('xml')
    assert hasattr(options, "saveTargetInstance"), "validate/ESEF plugin dependency inlineXbrlDocumentSet CLI args were not loaded."