A python package to do several analysis tasks on PhotonStream data, as proposed by the FACT project. This package provides functions for FACT data safed in the PhotonStream format. An open data sample of Crab observations can be found here.
This package contains functions for
- image cleaning via DBSCAN
- image cleaning via pixel thresholds as is defined here
- generating the usual Hillas features
- generating meta data
It can be used to generate this features and will return a dictionary containing those.
Installation can be done via git and pip:
$ git clone [email protected]:KevSed/FeatureStream.git
$ cd FeatureStream/
$ pip install .
Importing the package, loading a PhotonStream data file and cleaning an event using DBSCAN looks like this:
import photon_stream as ps
from feature_stream import phs2image, cleaning
# read in a data file
reader = ps.EventListReader(data_file)
# get a first event
event = next(reader)
# get the list of arrival times per photon
lol = event.photon_stream.list_of_lists
# generate an image (number of photons for each of the 1440 pixlels in FACT)
image = phs2image(lol)
# do the DBSCAN cleaning and return a boolean array of all cleaned pixels
cleaned_pixel = cleaning(image, lol)
# same as above, but using the facttools cleaning
cleaned_pixel = facttools_cleaning(image, lol)Generating all available features for a data or Monte Carlo file and saving them in a data frame:
from feature_stream import gen_features
# Generating meta data and Hillas features and saving to data frame df
df = gen_features(data_file)