Exercise/activity data has become a prolific resource, but applying any kind of sophisticated analyses is made difficult by the variety of file formats. This python library is intended to munge a number of these formats and present the data in a predictable and useable form. Moreover, the API is both closely intertwined with, and an extension of, the awesome Pandas library.
Please note this package is still very much an alpha release, so breaking changes are likely.
The package is available on PyPI:
$ pip install activityio
There is a read function at the top-level of activityio that dispatches the appropriate reader based on file extension:
>>> import activityio as aio
>>> data = aio.read('example.srm')NOTE substitute 'example.srm' with a path to your own activity file.
But you can also call sub-packages directly:
>>> from activityio import srm
>>> data = srm.read('example.srm')data in the above example is a subclass of the pandas.DataFrame and provides some neat additional functionality. Most notably, certain columns are "magic" in that they return specific pandas.Series subclasses. These subclasses make unit-switching easy, and provide other useful methods:
>>> type(data)
<class 'activityio._types.activitydata.ActivityData'>>>> data.head(5)
temp lap dist alt cad pwr speed hr
time
00:00:00 26.1 1 1.027 67 0 0 1.027 71
00:00:01 26.1 1 2.721 67 0 0 1.694 71
00:00:02 26.2 1 4.415 67 0 0 1.694 71
00:00:03 26.2 1 6.331 67 0 0 1.916 71
00:00:04 26.2 1 8.469 67 0 0 2.138 75>>> data.normpwr()
249.54104255943844>>> type(data.speed)
<class 'activityio._types.columns.Speed'>>>> data.speed.base_unit
'm/s'
>>> data.speed.kph.mean() # use a different unit
38.485063801685477>>> data.dist.base_unit
'm'
>>> data.dist.miles[-1]
134.78580023361226>>> data.alt.base_unit
'm'
>>> data.alt.ascent.sum()
1898.0
```But NOTE you lose this functionality if you go changing column names
>>> data = data.rename(columns={'alt': 'altitude'})
>>> type(data.altitude)
<class 'pandas.core.series.Series'>The main package is composed of sub-packages that contain the reading logic for the file format after which they're named. (e.g. activityio.fit is for parsing ANT/Garmin FIT files.)
The ultimate logic is defined in a _reading module, which provides two functions: gen_records and read_and_format.
gen_recordsis a generator function for iterating over the data-points in a file. The rows of the data table if you like. A "record" is a dictionary object.read_and_formatuses the above generator to return anActivityDataobject.
read_and_format is available at the top-level of a sub-package aliased as read; so reading in a file looks like srm.read('path_to_file.srm'). gen_records is imported under the same name.
There are also some useful tools provided in module by the same name.