Optional protocol_number parameter in experiment description

[k1o0]

On rare occasions the spacer number doesn't match the task position in the experiment description file.
We can make extraction easier in the future by first checking for a protocol_number key in each task dict. Relevant code:

ibllib/ibllib/pipes/dynamic_pipeline.py

Lines 235 to 244 in 9a9ff73

	task_protocols = acquisition_description.get('tasks', [])
	for i, (protocol, task_info) in enumerate(chain(*map(dict.items, task_protocols))):
	collection = task_info.get('collection', f'raw_task_data_{i:02}')
	task_kwargs = {'protocol': protocol, 'collection': collection}
	# For now the order of protocols in the list will take precedence. If collections are numbered,
	# check that the numbers match the order. This may change in the future.
	if re.match(r'^raw_task_data_\d{2}$', collection):
	task_kwargs['protocol_number'] = i
	if int(collection.split('_')[-1]) != i:
	_logger.warning('Number in collection name does not match task order')

This was partially implemented already in the Task base class:

ibllib/ibllib/io/session_params.py

Lines 372 to 376 in 9a9ff73

	protocols = sess_params.get('tasks', [])
	if task_protocol is None: # Return set of all task numbers
	numbers = (next(iter(x.values()), {}).get('protocol_number') for x in protocols)
	numbers = list(map(int, filter(lambda x: x is not None, numbers)))
	return (next(iter(numbers)) if len(numbers) == 1 else numbers) or None