JSON generator

Script for generating BioBB JSON Schemas for every package. Passing the name of the package and the absolute route to it, all the JSON files will be automatically generated from the documentation of every package's module.

Additionally, this script also generates the sample config JSON files for each package's module.

Execution steps

Step 1: activate environment

Activate the environment where the BioBB package is loaded:

conda activate biobb_env

Step 2: execute script

Run the python script passing the BioBB package to be parsed and the folder where the JSON files will be saved:

python3 json_generator.py -p biobb_package -o path/to/biobb_package/biobb_package

python3 json_generator.py --package biobb_package --output path/to/biobb_package/biobb_package

Files structure

The structure of a biobb package must be:

• biobb_package/
  • biobb_package/
    • __init__.py
    • block/
      • __init__.py
      • module 1
      • module 2
    • docs/
    • json_schemas/
    • test/
      • data/
      • reference/
        • config/
        • block 1/
        • block 2/
      • unitests/
      • conf.yml
  • .gitignore
  • Dockerfile
  • LICENSE
  • README.md
  • setup.py
  • Singularity.latest

__init__.py files

The __init__.py file in the first level must have the following structure:

name = "biobb_package"
__all__ = ["block1", "block2", ..., "blockn"]

The __init__.py file in the second level must have the following structure:

name = "block1"
__all__ = ["module1", "module2", ..., "modulen"]

In the __all__ list we have to put all the modules for which we want to generate a JSON schema.

json_schemas folder

The json_schemas folder must exist before executing the script. The file biobb_package.json won't be affected by the script's execution.

config folder

The config folder must exist before executing the script. All the JSON config files will be saved in this folder.

Docs specifications

All the docs must be written in correct Google Style Python Docstrings format and they must be properly indented. Example of documentation:

Description of the module.

Args:
    arg_name1 (arg_type): Description for argument 1. File type: input / output. `Sample file <url_to_sample_file1>`_. Accepted formats: format1, format2, format3. 
    arg_name2 (arg_type) (Optional): Description for argument 2. File type: input / output. `Sample file <url_to_sample_file2>`_. Accepted formats: format1, format2.
    properties (dic):
        * **property1** (*prop_type*) - (property1_default) Property 1 description.
        * **property2** (*prop_type*) - (property2_default) Property 2 description. Values: value1, value2, value3.
        * **property3** (*dic*) - (None) Property 3 description.
            * **parameter1** (*param_type*) - (parameter1_default) Parameter 1 description. Values: value1, value2, value3.

Arguments

The arguments must have the next format:

arg_name (arg_type) (Optional): Description. File type: input / output. `Sample file <url_to_sample_file1>`_. Accepted formats: format1, format2, format3.

The argument type must be between parenthesis. Argument types: str (string), dic (dictionary).

If the argument is optional, the (Optional) expression must be right next the argument type separated by a single space.

Description is a string descriving the functionality of the argument.

File type must be input or output. Although it's highly recommended to name the input files as input_something and the output files as output_something, with this parameter we ensure which type the file is.

Sample file is a link to a sample file for this argument. It's recommended to put the file used for unitest integrated in the same repository:

• biobb_package/biobb_package/test/data/block/input_file: for input files.
• biobb_package/biobb_package/test/reference/block/reference_file: for output files.

The format of the Sample file link must be in reStructuredText (RST) format: Sample file <url_to_sample_file>`_

If there are formats they must be a list preceded by the Accepted formats: expression.

Properties

The properties must have the next format:

* **property** (*prop_type*) - (property_default) Property description. Values: value1, value2, value3.

The property name must be a list item in markdown bold (between double asterisk).

The property type must be between parenthesis. Property types: str (string), dic (dictionary), int (integer), float, bool (boolean).

The property default must be between parenthesis. If it's a text it's highly recommendable to put it between double quotes instead of single ones.

If there are values they must be a list preceded by the Values: expression.

Parameters

The parameters must have the next format:

* **parameter** (*param_type*) - (parameter_default) Parameter description. Values: value1, value2, value3.

The parameter name must be a list item in markdown bold (between double asterisk).

The parameter type must be between parenthesis. Parameter types: str (string), dic (dictionary), int (integer), float, bool (boolean).

The parameter default must be between parenthesis. If it's a text it's highly recommendable to put it between double quotes instead of single ones.

If there are values they must be a list preceded by the Values: expression.

Configuration files

The creation of the configuration files is automatic and the data is taken from the path/to/biobb/package/test/conf.yml file. The script will generate a JSON config file for each module with properties defined in its parameters.

Credits

Genís Bayarri, Pau Andrio, Adam Hospital, Josep Lluis Gelpí.

Molecular Modeling and Bioinformatics Group. IRB Barcelona.