Python NOAA Unit Converter for Oil Spills
This repo provides a python package for doing physical unit conversion. It includes the odd units (API gravity, etc) that are used for oil spill response and planning, but not the odd units that other fields may require.
It also includes many common units for general use.
NOTE: NUCOS does NOT properly manage physical quantities -- rather it is helpful for working with units in the sloppy way humans often do. For example "pounds" is considered a unit of mass and can be converted, e.g. into kilograms -- this IS NOT PHYSICALLY correct, but it is useful.
It also handles "units" that are strictly dimensionless -- like mass fraction: e.g. g/kg. This is useful as you can convert to other dimensionless "units" like percent. But it does make the distinction between, e.g. mass fraction and volume fraction, so you can't convert between g/kg and ml/l, for example.
There are also a few utilities that are not strictly unit conversion:
- Converting latitude/longitude to/from degrees, degrees minutes seconds, etc. (and formatting as Unicode objects)
NOTE: lat-long parsing and formatting is also available in the lat-long parser project
- Converting to/from oil mass units to/from volume units: e.g. tons to barrels.
As of version 3, PyNUCOS is on PyPi and conda-forge, and of course, can be installed from source.
From PyPi:
pip install pynucos
From conda-forge:
conda install pynucos -c conda-forge
(the -c conda-forge is optional if you've already got the conda-forge channel)
From Source:
Get the source from gitHub:
https://github.com/NOAA-ORR-ERD/PyNUCOS
Then the usual:
pip install ./
There are many unit conversion codes out there, but none that easily support the strange units used in oil spill response (and the petroleum industry in general), such as API Gravity and conversion of amount of oil from mass to volume.
This code is used as the core lib for a desktop unit conversion application:
https://github.com/NOAA-ORR-ERD/wxNUCOS
It is available as an installable binary from:
The code is also used in the NOAA Oil Spill modeling tools:
Available on gitHub here:
https://github.com/NOAA-ORR-ERD
There is also a Javascript version available for use in browser Client-side applications:
https://github.com/NOAA-ORR-ERD/jsNUCOS
Most of the primary functionality is available with a single function:
In [7]: from nucos import convert
In [8]: convert('gal', 'liter', 1.0)
Out[8]: 3.7854118
However, some unit names can have different meanings, e.g. fluid ounce and weight ounce, so can not be converted:
In [9]: convert('oz', 'ml', 1.0)
---------------------------------------------------------------------------
UnitConversionError Traceback (most recent call last)
<ipython-input-9-86edffc0a76a> in <module>
----> 1 convert('oz', 'ml', 1.0)
~/Hazmat/ERD-PythonPackages/PyNUCOS/nucos/unit_conversion.py in convert(unit1, unit2, value, unit_type)
464
465 if unit_type != unit_type2:
--> 466 raise UnitConversionError("Cannot convert {0} to {1}"
467 .format(unit1, unit2))
468
UnitConversionError: Cannot convert oz to ml
To be more clearly specified, the unit type can be passed as the first argument:
In [10]: convert('volume', 'oz', 'ml', 1.0)
Out[10]: 29.57353
In [16]: convert('mass', 'oz', 'gram', 1.0)
Out[16]: 28.349523
There are functions for converting latitude and longitude to/from various formats.
Pass ustring=True to get a Unicode formatted string version.
In [24]: from nucos import LatLongConverter In [25]: LatLongConverter.ToDecDeg(-45, 34, 12) Out[25]: -45.57 In [26]: LatLongConverter.ToDecDeg(-45, 34, 12, ustring=True) Out[26]: '-45.570000°' In [27]: LatLongConverter.ToDegMin(-45.57) Out[27]: (-45.0, 34.2) In [28]: LatLongConverter.ToDegMin(-45.57, ustring=True) Out[28]: "-45° 34.200'" In [29]: LatLongConverter.ToDegMinSec(-45.57) Out[29]: (-45.0, 34, 12.0) In [30]: LatLongConverter.ToDegMinSec(-45.57, ustring=True) Out[30]: '-45° 34\' 12.00"'
Unit names are simple strings, and there are a lot of synonyms, both in ascii and Unicode formats.
The full list of units and names is in the NUCOS_unit_list.rst file.
You can programmatically access the unit types, unit names, etc, via assorted utility functions:
# get the names for a given unit type
nucos.get_supported_names('mass')
Out[15]:
['kilogram',
'kg',
'kilograms',
'lb',
'pounds',
'lbs',
'g',
'grams',
'mg',
'µg',
'ug',
'tons',
'uston',
'tonne',
'tonnes',
'metric ton',
'metric tons',
'mt',
'slugs',
'oz',
'ounces',
'ukton',
'long ton']
# all the available types
In [3]: nucos.get_unit_types()
Out[3]:
['Length',
'Oil Concentration',
'Area',
'Volume',
'Temperature',
'Delta Temperature',
'Mass',
'Time',
'Velocity',
'Discharge',
'Mass Discharge',
'Density',
'Kinematic Viscosity',
'Dynamic Viscosity',
'Interfacial Tension',
'Pressure',
'Concentration In Water',
'Concentration',
'Dimensionless',
'Mass Fraction',
'Volume Fraction',
'Angular Measure',
'Angular Velocity']
# Get the primary (spelled out) name for a unit
In [5]: nucos.get_primary_name('mg')
Out[5]: 'milligram'
# Get the abbreviation (short form) for a unit:
In [7]: nucos.get_abbreviation('microgram')
Out[7]: 'µg'
Added a couple utilities for working with the names.
Added some more synonyms for meter per second.
Fixed a bug with GetUnitNames() and capitalization.
Added code to get valid unit names for a given unit_type
Added code to get valid unit names for a given unit_type
The first release on PyPi -- major change in this release is the top-level package name is now nucos -- it used to be unit_conversion. The unit_conversion name is still there, but should raise a DeprecationWarning
If you have any suggestions for improvements, bug fixes, etc, please post an issue on GitHub:
https://github.com/NOAA-ORR-ERD/PyNUCOS
Or better yet, make a Pull Request.
there is a fairly comprehensive set of tests in:
nucos/tests
they can be run with pytest -- either directly:
pytest nucos/tests
or on the installed package:
pytest --pyargs nucos
Any additional units should be added to:
nucos/unit_data.py
It's a big nested dict with units, conversion factors and synonyms all there. Hopefully, it's self describing :-)
Please add a test if you add a new unit.
Be sure to run the tests after making any changes -- that will catch errors in the format, duplicate names, etc.
Minor changes can be done directly in the main branch.
Checklist for a new release:
- Make sure the tests all pass, of course
- make sure the version is properly set: in
nucos/__init__.py - Make sure that
NUCOS_unit_list.rsthas been updated -- it should happen when the tests are run. - Push to GitHub
- Make a release on GitHub -- follow the tag convention already there.
- Increment the __version__ in the main branch after making the release.
- Push to PyPi:
-
conda install twine build(you can pip install those, too) -python -m build-twine upload dist/*
NOTE: At this point, only Chris Barker has permissions on PyPi to do this -- so either ask him to do it, or ask him for permissions.
- Check conda-forge for a build: conda-forge should detect that a release was made on GitHub, and then build a new package -- you can check the progress here:
https://github.com/conda-forge/pynucos-feedstock
ChrisBarker-NOAA and JamesMakela-NOAA have permissions on the feedstock.