Fix loading data from the sync stream in spikeglx

[h-mayorquin]

After this PR on neo:
NeuralEnsemble/python-neo#1683

the sync channel is separated into its own stream. However, if someone wants to load the stream-sync there is an error:

from pathlib import Path

from spikeinterface.extractors import SpikeGLXRecordingExtractor

recording = SpikeGLXRecordingExtractor(folder_path=folder_path, stream_name="imec0.ap-SYNC")
recording


KeyError: 'imec0.ap-SYNC'
---------------------------------------------------------------------------
KeyError                                  Traceback (most recent call last)
Cell In[4], line 5
      1 from pathlib import Path
      3 from spikeinterface.extractors import SpikeGLXRecordingExtractor
----> 5 recording = SpikeGLXRecordingExtractor(folder_path=folder_path, stream_name="imec0.ap-SYNC")
      6 recording

File ~/development/spikeinterface/src/spikeinterface/extractors/neoextractors/spikeglx.py:80, in SpikeGLXRecordingExtractor.__init__(self, folder_path, load_sync_channel, stream_id, stream_name, all_annotations, use_names_as_ids)
     78 # Checks if the probe information is available and adds location, shanks and sample shift if available.
     79 signals_info_dict = {e["stream_name"]: e for e in self.neo_reader.signals_info_list}
---> 80 meta_filename = signals_info_dict[self.stream_id]["meta_file"]
     81 ap_meta_filename = meta_filename.replace(".lf", ".ap") if "lf" in self.stream_id else meta_filename
     82 ap_meta_file_exists = Path(ap_meta_filename).exists()

KeyError: 'imec0.ap-SYNC'

The reason for this is that stream id is built on the fly for the synch data and therefore is not available in the neo_reader.signals_info_list so this fails.

This PR fixes it by just avoiding this lookup when the sync stream is requested. It also adds a test and removes a test for loading the sync channel as that feature is going to be removed.

[samuelgarcia]

Merci.