unitpackage

A Python library to create, interact with, and store collections of frictionless Data Packages whose tabular resources carry unit-aware field metadata. Each data package pairs a CSV file with a JSON descriptor containing schema information (column names, types, units) and arbitrary nested metadata.

Key features

• Unit-aware fields — columns carry unit metadata (e.g. V, A / m2, s) and can be rescaled to compatible units.
• Collections — load, filter, slice, and iterate over sets of data packages from local directories or remote sources.
• Metadata access — nested metadata is accessible via attribute-style or dict-style access.
• Plotting — built-in 2D plotting with plotly.
• Domain extensions — subclass Collection and Entry for domain-specific features (e.g. electrochemistry via Echemdb).

Quick example

>>> from unitpackage.collection import Collection
>>> db = Collection.from_local('./doc/files')
>>> entry = db['demo_package_cv']
>>> entry.echemdb.description
'Sample data for the unitpackage module'

Access field units and rescale data:

>>> entry.field_unit('j')
'A / m2'
>>> entry.rescale({'E' : 'mV', 'j' : 'uA / m2'}).df
          t           E             j
0  0.000000 -196.961730  43008.842162
1  0.011368 -196.393321  51408.199892
...

Plot directly from an entry:

>>> entry.plot()

Rescale an entire collection at once:

>>> rescaled = db.rescale({'E': 'mV', 'j': 'uA / m2'})
>>> rescaled[0].field_unit('E')
'mV'

Detailed usage examples, including local collection creation and metadata handling, are provided in the documentation.

Installation

This package is available on PyPI and can be installed with pip:

pip install unitpackage

The package is also available on conda-forge and can be installed with conda

conda install -c conda-forge unitpackage

Please consult our documentation for more detailed installation instructions.

License

The contents of this repository are licensed under the GNU General Public License v3.0 or, at your option, any later version.