Wrapper around np.array that can use keys to access the first axis. Also has convenience functions for data formatting and recording.
Similar in usage to a {'x':np.array(~), 'y':np.array(~)} or np.recarray:
without sacrificing numpy indexing/slicing features.
Wrapper around np.array that can use keys to access first axis.
DictArray(data=None, **kwargs) -> np.array(~data, **kwargs)
data: can be any of `None`, `DictArray`, `dict of lists`, `dict of values`,
`list of lists`, `list of dicts`, `list of values`, `np.recarray`, `np.ndarray`
kwargs: passed to np.array() after processing `data`
NOTE: any `list` in above formats can be a sequence with __iter__
Methods:
.append(*data, **kwargs): add data to the array, where `data` fits format
above, OR keywords can be used
.keys() ~= dict().keys()
.items() ~= dict().items()
.values() ~= dict().values()
.as_dict() returns dict version of .items()
CAUTION: keys can be any valid dict key, except the types
tuple/slice/list/int/np.ndarray that have no custom handling.
All np.array attributes are implicitly delegated to DictArray().array
obj = DictArray([[0,2,4],[1,3,5]])
obj['x'] == obj[0] == obj[0,:] == [0,2,4]
obj['x',1] == obj[0,1] == 2
obj['y'] == obj[1] == obj[1,:] == [1,3,5]
obj['y',1] == obj[1,1] == 3
x, y = object(), object()
obj = DictArray({x:[0], y:[1]})
obj[x] == obj[0] == obj[0,:] == [0]
obj[x,0] == obj[0,0] == 0
obj[y] == obj[1] == obj[1,:] == [1]
obj[y,0] == obj[1,0] == 1
DictArray with data recording functions.
.set_record_file(file, format)
set a file/object/function used to record data points.
`file`:
str -> filepath -> .write(msg)
io.IOBase -> uses .write(msg)
logging.Logger -> uses Logger.info(msg)
logging.Handler -> uses Handler.handle(logging.getLogRecordFactory(msg))
callable -> calls, passing the new data point, unformatted
`format`: str: 'csv', 'dict', or 'list'
.append # extended to record new data when applicable
.start_recording() # begin recording appended data
.stop_recording() # stop recording appended data