ONE light API proposal

onelight.py

from onelib import one

"""
Different ONE backends are available, HTTP, figshare, etc.
One has to implement the following functions to create a new ONE backend:

* list_all_files(): return a list of relative file paths for ALL available files.

* search(...): return a list of dset_ids (by default, a dset_id is a relative file path).
  The default implementation calls list_all_files(), and performs the search directly on that list.
  More   sophisticated backends (like the alyx one) could use a REST API instead, which will be 
  more efficient when there is a huge number of files.

* download_file(dset_id): download a file and return a local file path

"""

# Set the configuration.
# The following is the default.
# one.set_download_dir('~/.one/data/{lab}/Subjects/{subject}/{date}/{session}/alf/{filename}')
# Use this to download all files directly in the current directory instead.
one.set_download_dir('.')

# Find all sessions that have both spikes.times and spikes.clusters
sessions = one.search(['spikes.times', 'spikes.clusters'])

# First argument is dataset_types. Other filters available as keyword arguments : subjects, lab, date_range, number.

# Get the first session, or do a loop.
session = sessions[0]

# OPTION 1: load object

# Load the spikes object of that first session. Files are first saved in the download directory, which acts as a cache.
spikes = one.load_object(session, 'spikes')  # Load all default dataset types for that object, at least spikes.times and spikes.clusters. There is an optional dataset_types keyword arguments to override that default list.

# Raster plot
plt.plot(spikes.times, spikes.clusters, 'o')

# OPTION 2: load datasets

spike_times = one.load_dataset(session, 'spikes.times')  # first dataset with that dataset type
spike_clusters = one.load_dataset(session, 'spikes.clusters')