Skip to content

prepare_spikesortingview_data.py

prepare_spikesortingview_data(*, recording, sorting, segment_duration_sec, snippet_len, max_num_snippets_per_segment, channel_neighborhood_size, output_file_name)

Prepare data for the SpikeSortingView.

Source code in src/spyglass/spikesorting/v0/figurl_views/prepare_spikesortingview_data.py 
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
def prepare_spikesortingview_data(
    *,
    recording: si.BaseRecording,
    sorting: si.BaseSorting,
    segment_duration_sec: float,
    snippet_len: Tuple[int],
    max_num_snippets_per_segment: Union[int, None],
    channel_neighborhood_size: int,
    output_file_name: str,
) -> str:
    """Prepare data for the SpikeSortingView."""
    unit_ids = np.array(sorting.get_unit_ids()).astype(np.int32)
    channel_ids = np.array(recording.get_channel_ids()).astype(np.int32)
    sampling_frequency = recording.get_sampling_frequency()
    num_frames = recording.get_num_frames()
    num_frames_per_segment = math.ceil(
        segment_duration_sec * sampling_frequency
    )
    num_segments = math.ceil(num_frames / num_frames_per_segment)

    with h5py.File(output_file_name, "w") as f:
        f.create_dataset("unit_ids", data=unit_ids)
        f.create_dataset(
            "sampling_frequency",
            data=np.array([sampling_frequency]).astype(np.float32),
        )
        f.create_dataset("channel_ids", data=channel_ids)
        f.create_dataset(
            "num_frames", data=np.array([num_frames]).astype(np.int32)
        )
        channel_locations = recording.get_channel_locations()
        f.create_dataset("channel_locations", data=np.array(channel_locations))
        f.create_dataset(
            "num_segments", data=np.array([num_segments]).astype(np.int32)
        )
        f.create_dataset(
            "num_frames_per_segment",
            data=np.array([num_frames_per_segment]).astype(np.int32),
        )
        f.create_dataset(
            "snippet_len",
            data=np.array([snippet_len[0], snippet_len[1]]).astype(np.int32),
        )
        f.create_dataset(
            "max_num_snippets_per_segment",
            data=np.array([max_num_snippets_per_segment]).astype(np.int32),
        )
        f.create_dataset(
            "channel_neighborhood_size",
            data=np.array([channel_neighborhood_size]).astype(np.int32),
        )
        f.attrs["recording_object"] = json.dumps({})
        f.attrs["sorting_object"] = json.dumps({})

        # first get peak channels and channel neighborhoods
        unit_peak_channel_ids = {}
        fallback_unit_peak_channel_ids = {}
        unit_channel_neighborhoods = {}
        for iseg in range(num_segments):
            something_missing = False
            for unit_id in unit_ids:
                if str(unit_id) not in unit_peak_channel_ids:
                    something_missing = True
            if not something_missing:
                break
            logger.info(f"Initial pass: segment {iseg}")
            start_frame = iseg * num_frames_per_segment
            end_frame = min(start_frame + num_frames_per_segment, num_frames)
            start_frame_with_padding = max(start_frame - snippet_len[0], 0)
            end_frame_with_padding = min(end_frame + snippet_len[1], num_frames)
            traces_with_padding = recording.get_traces(
                start_frame=start_frame_with_padding,
                end_frame=end_frame_with_padding,
            )
            for unit_id in unit_ids:
                if str(unit_id) not in unit_peak_channel_ids:
                    spike_train = sorting.get_unit_spike_train(
                        unit_id=unit_id,
                        start_frame=start_frame,
                        end_frame=end_frame,
                    )
                    if len(spike_train) > 0:
                        values = traces_with_padding[
                            spike_train - start_frame_with_padding, :
                        ]
                        avg_value = np.mean(values, axis=0)
                        peak_channel_ind = np.argmax(np.abs(avg_value))
                        peak_channel_id = channel_ids[peak_channel_ind]
                        channel_neighborhood = get_channel_neighborhood(
                            channel_ids=channel_ids,
                            channel_locations=channel_locations,
                            peak_channel_id=peak_channel_id,
                            channel_neighborhood_size=channel_neighborhood_size,
                        )
                        if len(spike_train) >= 10:
                            unit_peak_channel_ids[str(unit_id)] = (
                                peak_channel_id
                            )
                        else:
                            fallback_unit_peak_channel_ids[str(unit_id)] = (
                                peak_channel_id
                            )
                        unit_channel_neighborhoods[str(unit_id)] = (
                            channel_neighborhood
                        )
        for unit_id in unit_ids:
            peak_channel_id = unit_peak_channel_ids.get(str(unit_id), None)
            if peak_channel_id is None:
                peak_channel_id = fallback_unit_peak_channel_ids.get(
                    str(unit_id), None
                )
            if peak_channel_id is None:
                raise Exception(
                    f"Peak channel not found for unit {unit_id}. "
                    + "This is probably because no spikes were found in any "
                    + "segment for this unit."
                )
            channel_neighborhood = unit_channel_neighborhoods[str(unit_id)]
            f.create_dataset(
                f"unit/{unit_id}/peak_channel_id",
                data=np.array([peak_channel_id]).astype(np.int32),
            )
            f.create_dataset(
                f"unit/{unit_id}/channel_neighborhood",
                data=np.array(channel_neighborhood).astype(np.int32),
            )

        for iseg in range(num_segments):
            logger.info(f"Segment {iseg} of {num_segments}")
            start_frame = iseg * num_frames_per_segment
            end_frame = min(start_frame + num_frames_per_segment, num_frames)
            start_frame_with_padding = max(start_frame - snippet_len[0], 0)
            end_frame_with_padding = min(end_frame + snippet_len[1], num_frames)
            traces_with_padding = recording.get_traces(
                start_frame=start_frame_with_padding,
                end_frame=end_frame_with_padding,
            )
            traces_sample = traces_with_padding[
                start_frame
                - start_frame_with_padding : start_frame
                - start_frame_with_padding
                + int(sampling_frequency * 1),
                :,
            ]
            f.create_dataset(
                f"segment/{iseg}/traces_sample", data=traces_sample
            )
            all_subsampled_spike_trains = []
            for unit_id in unit_ids:
                peak_channel_id = unit_peak_channel_ids.get(str(unit_id), None)
                if peak_channel_id is None:
                    peak_channel_id = fallback_unit_peak_channel_ids.get(
                        str(unit_id), None
                    )
                if peak_channel_id is None:
                    raise Exception(
                        f"Peak channel not found for unit {unit_id}. "
                        + "This is probably because no spikes were found in any"
                        + " segment for this unit."
                    )
                spike_train = sorting.get_unit_spike_train(
                    unit_id=unit_id,
                    start_frame=start_frame,
                    end_frame=end_frame,
                ).astype(np.int32)
                f.create_dataset(
                    f"segment/{iseg}/unit/{unit_id}/spike_train",
                    data=spike_train,
                )
                channel_neighborhood = unit_channel_neighborhoods[str(unit_id)]
                peak_channel_ind = channel_ids.tolist().index(peak_channel_id)
                if len(spike_train) > 0:
                    spike_amplitudes = traces_with_padding[
                        spike_train - start_frame_with_padding, peak_channel_ind
                    ]
                    f.create_dataset(
                        f"segment/{iseg}/unit/{unit_id}/spike_amplitudes",
                        data=spike_amplitudes,
                    )
                else:
                    spike_amplitudes = np.array([], dtype=np.int32)
                if len(spike_train) > max_num_snippets_per_segment:
                    subsampled_spike_train = subsample(
                        spike_train, max_num_snippets_per_segment
                    )
                else:
                    subsampled_spike_train = spike_train
                f.create_dataset(
                    f"segment/{iseg}/unit/{unit_id}/subsampled_spike_train",
                    data=subsampled_spike_train,
                )
                all_subsampled_spike_trains.append(subsampled_spike_train)
            subsampled_spike_trains_concat = np.concatenate(
                all_subsampled_spike_trains, dtype=np.int32
            )
            # logger.info('Extracting spike snippets')
            spike_snippets_concat = extract_spike_snippets(
                traces=traces_with_padding,
                times=subsampled_spike_trains_concat - start_frame_with_padding,
                snippet_len=snippet_len,
            )
            # logger.info('Collecting spike snippets')
            index = 0
            for ii, unit_id in enumerate(unit_ids):
                channel_neighborhood = unit_channel_neighborhoods[str(unit_id)]
                channel_neighborhood_indices = [
                    channel_ids.tolist().index(ch_id)
                    for ch_id in channel_neighborhood
                ]
                num = len(all_subsampled_spike_trains[ii])
                spike_snippets = spike_snippets_concat[
                    index : index + num, :, channel_neighborhood_indices
                ]
                index = index + num
                f.create_dataset(
                    f"segment/{iseg}/unit/{unit_id}/subsampled_spike_snippets",
                    data=spike_snippets,
                )

get_channel_neighborhood(*, channel_ids, channel_locations, peak_channel_id, channel_neighborhood_size)

Return the channel neighborhood for a peak channel.

Source code in src/spyglass/spikesorting/v0/figurl_views/prepare_spikesortingview_data.py 
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
def get_channel_neighborhood(
    *,
    channel_ids: np.array,
    channel_locations: np.ndarray,
    peak_channel_id: int,
    channel_neighborhood_size: int,
):
    """Return the channel neighborhood for a peak channel."""
    channel_locations_by_id = {}
    for ii, channel_id in enumerate(channel_ids):
        channel_locations_by_id[channel_id] = channel_locations[ii]
    peak_location = channel_locations_by_id[int(peak_channel_id)]
    distances = []
    for channel_id in channel_ids:
        loc = channel_locations_by_id[int(channel_id)]
        dist = np.linalg.norm(np.array(loc) - np.array(peak_location))
        distances.append(dist)
    sorted_indices = np.argsort(distances)
    neighborhood_channel_ids = []
    for ii in range(min(channel_neighborhood_size, len(channel_ids))):
        neighborhood_channel_ids.append(int(channel_ids[sorted_indices[ii]]))
    return neighborhood_channel_ids

subsample(x, num)

Subsample an array.

Source code in src/spyglass/spikesorting/v0/figurl_views/prepare_spikesortingview_data.py 
256
257
258
259
260
261
def subsample(x: np.array, num: int):
    """Subsample an array."""
    if num >= len(x):
        return x
    stride = math.floor(len(x) / num)
    return x[0 : stride * num : stride]

extract_spike_snippets(*, traces, times, snippet_len)

Extract spike snippets.

Source code in src/spyglass/spikesorting/v0/figurl_views/prepare_spikesortingview_data.py 
264
265
266
267
268
269
270
271
272
273
274
275
276
277
def extract_spike_snippets(
    *, traces: np.ndarray, times: np.array, snippet_len: Tuple[int]
):
    """Extract spike snippets."""
    a = snippet_len[0]
    b = snippet_len[1]
    T = a + b
    M = traces.shape[1]
    L = len(times)
    ret = np.zeros((L, T, M), dtype=traces.dtype)
    if L > 0:
        for t in range(T):
            ret[:, t, :] = traces[times - a + t, :]
    return ret