BioProv - W3C-PROV provenance documents for bioinformatics

Code
Tests
Docs

BioProv is a Python library for W3C-PROV representation of bioinformatics workflows. It enables you to quickly write workflows and to describe relationships between samples, files, users and programs.

Please see the tutorials for a more detailed introduction and visit ReadTheDocs for the complete documentation.

Quickstart

>>> import bioprov as bp

# Create samples and file objects
>>> sample = bp.Sample("mysample")
>>> genome = bp.File("mysample.fasta", "genome")
>>> sample.add_files(genome)

# Create programs
>>> output = sample.files["blast_out"] = bp.File("mysample.blast.tsv", "blast_out")
>>> blastn = bp.Program("blastn",
                        params={"-query": sample.files["genome"],
                                "-db": "mydb.fasta", "-out": output}
                        )
>>> sample.add_programs(blastn)

# Run programs
>>> sample.run_programs()

# Save your project
>>> proj = bp.Project((sample,), tag="example_project")
>>> proj.to_json()

# Create PROV documents
>>> prov = bp.BioProvDocument(proj)

# Save in PROVN or graphical format
>>> prov.write_provn()  # human-readable text format
>>> prov.dot.write_pdf()  # graphical format

BioProv also has a command-line application to run preset workflows.

$ bioprov -h
usage: bioprov [-h] [--show_config | --show_db | --clear_db | -v | -l]
               {genome_annotation,blastn,kaiju} ...

BioProv command-line application. Choose a command to begin.

optional arguments:
  -h, --help            show this help message and exit
  --show_config         Show location of config file.
  --show_db             Show location of database file.
  --clear_db            Clears all records in database.
  -v, --version         Show BioProv version
  -l, --list            List Projects in the BioProv database.

workflows:
  {genome_annotation,blastn,kaiju}

BioProv is built with the Biopython and Pandas libraries.

You can import data into BioProv using Pandas objects.

# Read csv straight into BioProv
>>> samples = bp.read_csv("my_dataframe.tsv", sep="\t", sequencefile_cols="assembly")

# Alternatively, use a pandas DataFrame
>>> df = pd.read_csv("my_dataframe.tsv", sep="\t")

# [...] manipulate your df
>>> df["assembly"] = "assembly_directory/" + df["assembly"]

# Now load from your df
>>> project = bp.from_df(df, sequencefile_cols="assembly", source_file="my_dataframe.tsv")

# `samples` becomes a Project dict-like object
>>> sample1 = project['sample1']

# You can also export your sample and associated files and attributes as a dataframe
>>> project.to_csv()

Installation

# Install from pip
$ pip install bioprov

# Install from conda
$ conda install -c conda-forge -c bioconda bioprov

# Install from source
$ git clone https://github.com/vinisalazar/bioprov && cd bioprov;     # download
$ conda env create -f environment.yaml && conda activate bioprov;     # install dependencies
$ pip install . && pytest;                                            # install and test

Important! BioProv requires Prodigal to be tested. Otherwise tests will fail.

Contributions are welcome!

BioProv is in active development and no warranties are provided (please see the License).

Dependencies

BioProv requires the follow dependencies to run. Also see the setup and environment files.

• biopython
• coolname
• coveralls
• dataclasses
• pandas
• prodigal
• prov
• provstore-api
• pydot
• pytest
• pytest-cov
• tqdm
• tinydb