IndexErrors with ORFlashCamListenerStatusDecoder when running build_raw

[slwatkins]

I've updated pygama et al. to the latest versions and have been testing the build_raw functionality, as I'm writing up some code to validate the packets in FC and ORCA files.

With the latest updates and recent files, the build_raw function fails when attempting to decode the FlashCam Listener Status. Here are two MWEs using files from NERSC (where I've truncated the paths for privacy reasons):

from daq2lh5 import build_raw
build_raw(f'{full_path_to_raw_data_on_nersc}/l200-p04-r006-tst-20230510T101033Z.orca', out_spec="raw-data.lh5")

This errors with:

no implementation of OR1DHistoDecoder, corresponding packets will be skipped
no implementation of ORScriptDecoderForRecord, corresponding packets will be skipped
no implementation of ORScriptDecoderForState, corresponding packets will be skipped
got 2 ORFlashCam Waveform Decoders in memory!

---------------------------------------------------------------------------
IndexError                                Traceback (most recent call last)
Cell In[6], line 1
----> 1 build_raw(f'{full_path_to_raw_data_on_nersc}/l200-p04-r006-tst-20230510T101033Z.orca', out_spec="raw-data.lh5")

File /opt/anaconda3/envs/py310/lib/python3.10/site-packages/daq2lh5/build_raw.py:231, in build_raw(in_stream, in_stream_type, out_spec, buffer_size, n_max, overwrite, compass_config_file, hdf5_compression, **kwargs)
    229 n_bytes_last = streamer.n_bytes_read
    230 while True:
--> 231     chunk_list = streamer.read_chunk()
    233     if log.getEffectiveLevel() <= logging.INFO and n_max == np.inf:
    234         progress_bar.update(streamer.n_bytes_read - n_bytes_last)

File /opt/anaconda3/envs/py310/lib/python3.10/site-packages/daq2lh5/data_streamer.py:283, in DataStreamer.read_chunk(self, chunk_mode_override, rp_max, clear_full_buffers)
    281 still_has_data = True
    282 while True:
--> 283     still_has_data = self.read_packet()
    284     if not still_has_data:
    285         break

File /opt/anaconda3/envs/py310/lib/python3.10/site-packages/daq2lh5/orca/orca_streamer.py:312, in OrcaStreamer.read_packet(self)
    310 decoder = self.decoder_id_dict[data_id]
    311 rbl = self.rbl_id_dict[data_id]
--> 312 self.any_full |= decoder.decode_packet(packet, self.packet_id, rbl)
    313 return True

File /opt/anaconda3/envs/py310/lib/python3.10/site-packages/daq2lh5/orca/orca_flashcam.py:391, in ORFlashCamListenerStatusDecoder.decode_packet(self, packet, packet_id, rbl)
    389 for i_card in range(n_cards):
    390     i_cd = 15 + i_card * cd_size  # start index of this card's data
--> 391     tbl["card_id"].flattened_data.nda[fd0 + i_card] = packet[i_cd + 0]
    392     tbl["card_status"].flattened_data.nda[fd0 + i_card] = packet[i_cd + 1]
    393     tbl["card_eventnumber"].flattened_data.nda[fd0 + i_card] = packet[i_cd + 2]

IndexError: index 1195 is out of bounds for axis 0 with size 187

Trying a different file from a more recent period:

from daq2lh5 import build_raw
build_raw(
    f'{full_path_to_raw_data_on_nersc}/l200-p06-r000-phy-20230611T222756Z.orca',
    out_spec='./test.lh5',
    overwrite=True,
)

Which errors with:

no implementation of OR1DHistoDecoder, corresponding packets will be skipped
no implementation of ORScriptDecoderForRecord, corresponding packets will be skipped
no implementation of ORScriptDecoderForState, corresponding packets will be skipped
got 4 ORFlashCam Waveform Decoders in memory!

---------------------------------------------------------------------------
IndexError                                Traceback (most recent call last)
Cell In[6], line 1
----> 1 build_raw(
      2          f'{full_path_to_raw_data_on_nersc}/l200-p06-r000-phy-20230611T222756Z.orca',
      3          out_spec='./test.lh5', overwrite=True)

File ~/.conda/envs/py39/lib/python3.9/site-packages/daq2lh5/build_raw.py:231, in build_raw(in_stream, in_stream_type, out_spec, buffer_size, n_max, overwrite, compass_config_file, hdf5_compression, **kwargs)
    229 n_bytes_last = streamer.n_bytes_read
    230 while True:
--> 231     chunk_list = streamer.read_chunk()
    233     if log.getEffectiveLevel() <= logging.INFO and n_max == np.inf:
    234         progress_bar.update(streamer.n_bytes_read - n_bytes_last)

File ~/.conda/envs/py39/lib/python3.9/site-packages/daq2lh5/data_streamer.py:283, in DataStreamer.read_chunk(self, chunk_mode_override, rp_max, clear_full_buffers)
    281 still_has_data = True
    282 while True:
--> 283     still_has_data = self.read_packet()
    284     if not still_has_data:
    285         break

File ~/.conda/envs/py39/lib/python3.9/site-packages/daq2lh5/orca/orca_streamer.py:312, in OrcaStreamer.read_packet(self)
    310 decoder = self.decoder_id_dict[data_id]
    311 rbl = self.rbl_id_dict[data_id]
--> 312 self.any_full |= decoder.decode_packet(packet, self.packet_id, rbl)
    313 return True

File ~/.conda/envs/py39/lib/python3.9/site-packages/daq2lh5/orca/orca_flashcam.py:421, in ORFlashCamListenerStatusDecoder.decode_packet(self, packet, packet_id, rbl)
    415 tbl["card_n_link_errors"].flattened_data.nda[fd0 + i_card] = packet[
    416     i_cd + 13
    417 ]
    418 for j in range(5):
    419     tbl["card_n_other_errors"].flattened_data.nda[
    420         (fd0 + i_card) * 5 + j
--> 421     ] = packet[i_cd + 14 + j]
    422 for j in range(5):
    423     tbl["card_temps"].flattened_data.nda[
    424         (fd0 + i_card) * 5 + j
    425     ] = int_packet[i_cd + 19 + j]

IndexError: index 2391 is out of bounds for axis 0 with size 2391

It's odd that the index errors are happening at different lines, which could be a difference between tst and phy files.

Note that the test for build_raw uses an old file that does not have any packets for the FlashCam Status Listener, so that test will need to be updated to include more packets.

[gipert]

I have uploaded to legend-testdata a non-truncated file that triggers this error: legend-exp/legend-testdata@ed54f0c. Has been produced by Brady a while ago and should include all the important packets. We should add unit tests that use it.

[jasondet]

This is my fault -- my implementation of this decoder has bugs. Let me try to find time to fix it.

[jasondet]

Fixing the status decoder has to wait for Simon to add some methods to pyfcutils. In the meantime I have updated the status decoder to just skip decoding. I'll turn status decoding back on as part of resolving the other issue.