kreczko / fast-curator

Curate your datasets and metadata in a plain-text, machine-readable, version-controllable way

fast-curator

Create, read and write dictionary descriptions of input datasets to process. Currently all datasets are expected to be built from sets of ROOT Trees.

Installing

pip install --user fast-curator

Usage

# Local files:
fast_curator -o output_file_list.txt -t tree_name -d dataset_name --mc input/files/*root

# Single XROOTD files:
fast_curator -o output_file_list.txt --mc root://my.domain.with.files://input/files/one_file.root

# XROOTD files with several globs
fast_curator -o output_file_list.txt --mc root://my.domain.with.files://inp*/files/*.root

Notes: 1. If the command is called multiple times with the same output file (using the -o option), the additional files specified will be appended to the output file. 2. Arbitrary meta-data (such as cross-section, data quality, generator precision, etc) can be added to each dataset with the -m option.

For more guidance try the built-in help:

fast_curator --help

Reading dataset files back

import fast_curator
datasets = fast_curator.read.from_yaml("my_dataset_file.yml")

Will return a list of datasets with the default section applied to each dataset.

Further Documentation

Is on its way...