pfa-validator

A small python program that validates that a PFA document is:

• Syntaxically correct JSON
• Syntaxically correct PFA
• Semantically correct PFA

This project does not perform cross-validation on the PFA document.

Usage

The program can be run as a Python program or as a Docker container.

It may fetch the PFA document to validate from the file system or from a PostgreSQL database.

• When fetching the PFA document from the file system, the program expects a PFA_PATH environment variable that contains the path to the PFA document to validate.

• When fetching the PFA document from a PostgreSQL database, the program expects the following environment variables:

  • INPUT_METHOD that must be set to POSTGRESQL
  • the database's credentials: DB_HOST, DB_PORT, DB_NAME, DB_USER and DB_PASSWORD
  • the parameters of the query to perform: DB_TABLE DB_COLUMN
  • the job ID (used to check the job_id column and get the right PFA document): JOB_ID

In addition, the program will also need the following environment variables to be set up in order to get some validation data: FEATURES_DB_HOST, FEATURES_DB_PORT, FEATURES_DB_NAME, FEATURES_DB_USER, FEATURES_DB_PASSWORD and FEATURES_DB_TABLE

Examples

Validate a PFA file

1. Build the image: docker build -t pfa-validator .
2. Run a container based on that image:

    docker run --volume $(pwd)/data:/data \
    --env PFA_PATH="/data/example_01_valid/model.pfa" \
    --env FEATURES_DB_HOST=db \
    --env FEATURES_DB_PORT=5432 \
    --env FEATURES_DB_NAME=sample \
    --env FEATURES_DB_USER=sample \
    --env FEATURES_DB_PASSWORD=... \
    --env FEATURES_DB_TABLE=features \
    hbpmip/pfa-validator:0.10.1-2

Validate a PFA PostgreSQL column

1. Build the image: docker build -t pfa-validator .
2. Run a container based on that image:

  docker run \
    --env INPUT_METHOD=POSTGRESQL \
    --env JOB_ID=1
    --env DB_HOST=172.20.0.2 \
    --env DB_PORT=5432 \
    --env DB_NAME=woken \
    --env DB_USER=woken \
    --env DB_PASSWORD=... \
    --env DB_TABLE=job_result \
    --env DB_COLUMN=data \
    --env FEATURES_DB_HOST=db \
    --env FEATURES_DB_PORT=5432 \
    --env FEATURES_DB_NAME=sample \
    --env FEATURES_DB_USER=sample \
    --env FEATURES_DB_PASSWORD=... \
    --env FEATURES_DB_TABLE=features \
    hbpmip/pfa-validator:0.10.1-2

NOTE: If you don't want to use Docker, you can install the dependencies with: pip install -r requirements.txt and run this program using Python2. The environment variables still have to be set up.

TODO

• Allow validation of YAML formatted PFA document

Acknowledgements

This work has been funded by the European Union Seventh Framework Program (FP7/2007­2013) under grant agreement no. 604102 (HBP)

This work is part of SP8 of the Human Brain Project (SGA1).