v2_dataset – Python object-relational mapping (ORM) and tools for the V2 Dataset

• Documentation
• Issue tracker
• Repository contents
• History of changes
• Contribution/development guide

This repository contains a Python package based on the SQLAlchemy object-relational mapping (ORM) framework for accessing the V2 Dataset's curated information. It also contains a minimal command-line interface that supports the creation of SQLite databases to hold curated information and its manipulation through a set of parsing routines that are being continuously developed.

The curated information itself is stored in the contents repository, together with information on how to obtain the recording data (which is not open-sourced, as of now). The contents repository should be cloned separately and linked from your working copy, as explained below.

Installation and setup

The library may be installed with

pip install lcg-neuro-v2-dataset-orm

Additional steps:

1. Clone the V2 Dataset's contents repository to any directory on your computer, e.g.:

   git clone git@gitlab.com:lcg/neuro/v2/dataset/contents /data/v2-dataset
   

   Make sure to check out a compatible version of the contents repository. Usually, the latest version of the library will be compatible with the latest version of the dataset.

2. The library must be able to locate the dataset directory, which can be done in one of two ways:

   • If Define a V2_DATASET_DIR environment variable pointing to the directory where you cloned the contents repository (you may do it with a .envrc file), or
   • Create a symlink named dataset (it is Git-ignored in this repository) in your working copy or library installation directory (/lib/pythonx.y/site-packages/lcg-neuro-v2-dataset-orm) pointing to the contents repository copy.

Note: You may need g++ or clang to compile some dependent libraries's extensions.

Development notes

Git LFS and cloning

Cloning may take a while because we use Git LFS to track larger files that are included mainly for testing. After the initial cloning, further pushing/pulling/fetching should only take as long as needed to download missing files, which we intend to keep at a minimum over this project's lifetime.

Important: you must install Git LFS before cloning.

— Powered by GitLab CI
— Created by Pedro Asad <pasad@lcg.ufrj.br> using cookiecutter and @pedroasad.com/templates/python/python/app-1.0.0