makeprov 0.4.5

An provenance tracking library for simple Python workflows

makeprov: Pythonic Provenance Tracking

This library provides a way to track file provenance in Python workflows using PROV (W3C Provenance) semantics. Decorators declare inputs and outputs, provenance is written automatically, and templated targets can be resolved on demand.

Features

• Use decorators to define rules for workflows.
• Resolve templated targets (results/{sample}.txt) via parse-style patterns.
• Support phony/meta rules for orchestration alongside file-producing rules.
• Automatically generate RDF-based provenance metadata (rdflib optional).
• Handles input and output streams.
• Integrates with Python's type hints for easy configuration.
• Outputs provenance data in TRIG format if rdflib is installed; otherwise outputs json-ld.
• Optional Snakemake CLI integration that turns --d3dag and --detailed-summary output into PROV JSON-LD artifacts ready for inclusion in Snakemake HTML reports.

Installation

You can install the module directly from PyPI:

pip install makeprov

Install the Snakemake extra if you want to use the CLI bridge:

pip install "makeprov[snakemake]"

Usage

Here’s an example of how to use this package in your Python scripts:

from makeprov import rule, InPath, OutPath, build

@rule()
def process_data(
    sample: int | None = None,
    input_file: InPath = InPath('data/{sample:d}.txt'),
    output_file: OutPath = OutPath('results/{sample:d}.txt')
):
    with input_file.open('r') as infile, output_file.open('w') as outfile:
        data = infile.read()
        outfile.write(data.upper())

if __name__ == '__main__':
    # Build a specific templated target and its prerequisites
    from makeprov import build
    build('results/1.txt')

    # Or expose rules via a command line interface
    import defopt
    defopt.run(process_data)

You can execute examples/example.py via the CLI like so:

python examples/example.py build-all

# Or set configuration through the CLI
python examples/example.py build-all --conf='{"base_iri": "http://mybaseiri.org/", "prov_dir": "my_prov_directory"}' --force --input_file input.txt --output_file final_output.txt

# Or set configuration through a TOML file
python examples/example.py build-all -c @my_config.toml

# Inspect dependency resolution without executing rules
python examples/example.py --explain results/1.txt
python examples/example.py --to-dot results/1.txt

Complex CSV-to-RDF Workflow

For a more involved scenario, see examples/complex_example.py. It creates multiple CSV files, aggregates their contents, and emits an RDF graph that is both serialized to disk and embedded into the provenance dataset because the function returns an rdflib.Graph.

@rule()
def export_totals_graph(
    totals_csv: InPath = InPath("data/region_totals.csv"),
    graph_ttl: OutPath = OutPath("data/region_totals.ttl"),
) -> Graph:
    graph = Graph()
    graph.bind("sales", SALES)

    with totals_csv.open("r", newline="") as handle:
        for row in csv.DictReader(handle):
            region_key = row["region"].lower().replace(" ", "-")
            subject = SALES[f"region/{region_key}"]

            graph.add((subject, RDF.type, SALES.RegionTotal))
            graph.add((subject, SALES.regionName, Literal(row["region"])))
            graph.add((subject, SALES.totalUnits, Literal(row["total_units"], datatype=XSD.integer)))
            graph.add((subject, SALES.totalRevenue, Literal(row["total_revenue"], datatype=XSD.decimal)))

    with graph_ttl.open("w") as handle:
        handle.write(graph.serialize(format="turtle"))

    return graph

Run the entire workflow, including CSV generation and RDF export, with:

python examples/complex_example.py build-sales-report

Bundling nested provenance and directory outputs

Rules can merge the provenance from any rules they invoke by passing merge=True to makeprov.rule. Pair this with makeprov.OutDir to declare a directory and then materialize multiple outputs beneath it while keeping them linked to a single provenance record. Use makeprov.InDir for the same tracked-directory semantics on inputs. See examples/merge_outdir_example.py for an example.

Merging is enabled by default: top-level runs start a provenance buffer and flush it once the CLI finishes, so downstream rules end up in one document unless you explicitly turn buffering off with merge=False on a rule or in the global config. Nested merges append to their parent buffer rather than writing multiple files.

Configured context and isolated sessions

examples/context_demo_example.py demonstrates pinning a base IRI, writing provenance to a dedicated directory, and running rules inside an isolated session so registries and buffers do not leak across runs:

python examples/context_demo_example.py build-all

Snakemake workflows

makeprov ships with an optional subcommand that shells out to Snakemake and converts the job DAG together with --detailed-summary metadata into a PROV document. The CLI mirrors the familiar configuration flags from makeprov.config and writes JSON-LD by default.

python -m makeprov.snakemake --prov-path prov/snakemake -- --snakefile Snakefile --nolock

Wire the resulting file into a report by marking it with Snakemake’s report() helper:

rule provenance:
    input:
        "results/word_count.txt"
    output:
        "prov/snakemake.json"
    shell:
        (
            "python -m makeprov.snakemake "
            "--prov-path prov/snakemake "
            "--out-fmt json --context --frame provenance "
            "-- "
            "--snakefile {workflow.snakefile} --nolock {input}"
        )

Using the optional --forceall-dag flag ensures that the job-level dependency edges in the provenance graph remain complete even when Snakemake skips nodes that are already up to date.

Configuration

You can customize the provenance tracking with the following options:

• base_iri (str): Base IRI for new resources
• prov_dir (str): Directory for writing PROV .json-ld or .trig files
• force (bool): Force running of dependencies
• dry_run (bool): Only check workflow, don't run anything

Scoped spans and cached downloads

Use makeprov.span(label, prov_path=None, frame=None, context=None) as a context manager or decorator to bracket a chunk of work in its own provenance buffer:

from makeprov import span

with span("model-run", prov_path="prov/models/model1"):
    run_model()

For remote resources that are cached locally, wrap the path with CachedDownload. It will lazily fetch on first access and record the source URL (and optional headers) in the provenance:

from makeprov import CachedDownload, rule

@rule()
def fetch_data(meta_json=CachedDownload("https://example.org/meta.json", "cache/meta.json")):
    with meta_json.open() as handle:
        return handle.read()

Documentation

Build the Sphinx docs (including autosummary API stubs) with the docs extra so that the CLI dependencies needed for imports are available:

pip install -e ".[docs]"
python docs/build.py

Contributing

Contributions are welcome! Please open an issue or submit a pull request.

License

This project is licensed under the MIT License - see the LICENSE file for details.