Fix channel name in test with 'modern' corrected AGIPD data

extra_data/tests/conftest.py

@@ -161,8 +161,7 @@ def mock_spb_raw_and_proc_run():
 @pytest.fixture(scope='session')
 def mock_modern_spb_proc_run(format_version):
  with TemporaryDirectory() as td:
- make_examples.make_modern_spb_proc_run(
- td, format_version=format_version)
+ make_examples.make_modern_spb_proc_run(td)
  yield td
 
 

extra_data/tests/make_examples.py

@@ -346,13 +346,13 @@ def make_reduced_spb_run(dir_path, raw=True, rng=None, format_version='0.5'):
  format_version=format_version)
 
 
-def make_modern_spb_proc_run(dir_path, format_version='0.5'):
+def make_modern_spb_proc_run(dir_path, format_version='1.2'):
  for modno in range(16):
  path = osp.join(dir_path, f'CORR-R0142-AGIPD{modno:0>2}-S00000.h5')
  write_file(path, [
- AGIPDModule(f'SPB_DET_AGIPD1M-1/CORR/{modno}CH0', raw=False,
- frames_per_train=32,
-  legacy_name=f'SPB_DET_AGIPD1M-1/DET/{modno}CH0')
+ AGIPDModule(f'SPB_DET_AGIPD1M-1/CORR/{modno}CH0', channel_name='output',
+ raw=False, frames_per_train=32,
+ legacy_name=f'SPB_DET_AGIPD1M-1/DET/{modno}CH0')
  ], ntrains=64, chunksize=32, format_version=format_version)
 
 

extra_data/tests/mockdata/detectors.py

@@ -19,14 +19,15 @@ class DetectorModule:
  ]
 
  def __init__(self, device_id, frames_per_train=64, raw=True,
- legacy_name=None):
+ channel_name='xtdf', legacy_name=None):
  self.device_id = device_id
  self._frames_per_train = frames_per_train
  if not raw:
  # Raw data has an extra dimension, used in AGIPD to separate data
  # and gain. This dimension is removed by the calibration process.
  self.image_dims = self.image_dims[1:]
  self.raw = raw
+ self.channel_name = channel_name
  self.legacy_name = legacy_name
 
  def write_control(self, f):
@@ -81,13 +82,13 @@ def write_instrument(self, f):
  if ntrains_pad % self.chunksize:
  ntrains_pad += + self.chunksize - (ntrains_pad % self.chunksize)
 
+ inst_source = f'{self.device_id}:{self.channel_name}'
+
  # INDEX
  for part in self.output_parts:
- dev_chan = '%s:xtdf/%s' % (self.device_id, part)
-
- i_first = f.create_dataset('INDEX/%s/first' % dev_chan,
+ i_first = f.create_dataset(f'INDEX/{inst_source}/{part}/first',
  (self.ntrains,), 'u8', maxshape=(None,))
- i_count = f.create_dataset('INDEX/%s/count' % dev_chan,
+ i_count = f.create_dataset(f'INDEX/{inst_source}/{part}/count',
  (self.ntrains,), 'u8', maxshape=(None,))
  if part == 'image':
  # First first is always 0
@@ -108,63 +109,65 @@ def write_instrument(self, f):
  if self.raw:
  # Raw data have an extra dimension (length 1) and an unlimited max
  # for the first dimension.
- ds = f.create_dataset('INSTRUMENT/%s:xtdf/image/trainId' % self.device_id,
+ ds = f.create_dataset(f'INSTRUMENT/{inst_source}/image/trainId',
  (nframes, 1), 'u8', maxshape=(None, 1))
  ds[:, 0] = tid_index
 
- pid = f.create_dataset('INSTRUMENT/%s:xtdf/image/pulseId' % self.device_id,
+ pid = f.create_dataset(f'INSTRUMENT/{inst_source}/image/pulseId',
  (nframes, 1), 'u8', maxshape=(None, 1))
  pid[:, 0] = pid_index
 
- cid = f.create_dataset('INSTRUMENT/%s:xtdf/image/cellId' % self.device_id,
+ cid = f.create_dataset(f'INSTRUMENT/{inst_source}/image/cellId',
  (nframes, 1), 'u2', maxshape=(None, 1))
  cid[:, 0] = pid_index # Cell IDs mirror pulse IDs for now
  else:
  # Corrected data drops the extra dimension, and maxshape==shape.
  f.create_dataset(
- 'INSTRUMENT/%s:xtdf/image/trainId' % self.device_id,
+ f'INSTRUMENT/{inst_source}/image/trainId',
  (nframes,), 'u8', chunks=True, data=tid_index
  )
 
  f.create_dataset(
- 'INSTRUMENT/%s:xtdf/image/pulseId' % self.device_id,
+ f'INSTRUMENT/{inst_source}/image/pulseId',
  (nframes,), 'u8', chunks=True, data=pid_index
  )
 
  f.create_dataset( # Cell IDs mirror pulse IDs for now
- 'INSTRUMENT/%s:xtdf/image/cellId' % self.device_id,
+ f'INSTRUMENT/{inst_source}/image/cellId',
  (nframes,), 'u2', chunks=True, data=pid_index
  )
 
  max_len = None if self.raw else nframes
  for (key, datatype, dims) in self.image_keys:
- f.create_dataset('INSTRUMENT/%s:xtdf/image/%s' % (self.device_id, key),
+ f.create_dataset(f'INSTRUMENT/{inst_source}/image/{key}',
  (nframes,) + dims, datatype, maxshape=((max_len,) + dims))
 
 
  # INSTRUMENT (other parts)
  for part in ['detector', 'header', 'trailer']:
- ds = f.create_dataset('INSTRUMENT/%s:xtdf/%s/trainId' % (self.device_id, part),
+ ds = f.create_dataset(f'INSTRUMENT/{inst_source}/{part}/trainId',
  (ntrains_pad,), 'u8', maxshape=(None,))
  ds[:self.ntrains] = trainids
 
  for (key, datatype, dims) in self.other_keys:
- f.create_dataset('INSTRUMENT/%s:xtdf/%s' % (self.device_id, key),
+ f.create_dataset(f'INSTRUMENT/{inst_source}/{key}',
  (ntrains_pad,) + dims, datatype, maxshape=((None,) + dims))
 
  if self.legacy_name is not None:
+ # The legacy source name for corrected data is the same as for
+ # raw data, which for these detectors always has the xtdf channel.
  f[f'INDEX/{self.legacy_name}:xtdf'] = h5py.SoftLink(
- f'/INDEX/{self.device_id}:xtdf')
+ f'/INDEX/{inst_source}')
  f[f'INSTRUMENT/{self.legacy_name}:xtdf'] = h5py.SoftLink(
- f'/INSTRUMENT/{self.device_id}:xtdf')
+ f'/INSTRUMENT/{inst_source}')
 
  def datasource_ids(self):
  for part in self.output_parts:
- yield 'INSTRUMENT/%s:xtdf/%s' % (self.device_id, part)
+ yield f'INSTRUMENT/{self.device_id}:{self.channel_name}/{part}'
 
  if self.legacy_name is not None:
  for part in self.output_parts:
- yield 'INSTRUMENT/%s:xtdf/%s' % (self.legacy_name, part)
+ yield f'INSTRUMENT/{self.legacy_name}:xtdf/{part}'
 
 
 class AGIPDModule(DetectorModule):

extra_data/tests/test_reader_mockdata.py

@@ -1153,9 +1153,9 @@ def test_run_metadata_no_trains(mock_scs_run):
 def test_proc_legacy_sources(mock_modern_spb_proc_run):
  run = RunDirectory(mock_modern_spb_proc_run)
 
- src_pattern = 'SPB_DET_AGIPD1M-1/{}/{}CH0:xtdf'
- corr_sources = {src_pattern.format('CORR', i) for i in range(16)}
- det_sources = {src_pattern.format('DET', i) for i in range(16)}
+ src_pattern = 'SPB_DET_AGIPD1M-1/{}/{}CH0:{}'
+ corr_sources = {src_pattern.format('CORR', i, 'output') for i in range(16)}
+ det_sources = {src_pattern.format('DET', i, 'xtdf') for i in range(16)}
 
  # Should contain both canonical and legacy names.
  assert run.all_sources == corr_sources | det_sources
@@ -1169,7 +1169,7 @@ def test_proc_legacy_sources(mock_modern_spb_proc_run):
  assert run.legacy_sources == dict(zip(
  sorted(det_sources), sorted(corr_sources)))
 
- det_mod0 = src_pattern.format('DET', 0)
+ det_mod0 = src_pattern.format('DET', 0, 'xtdf')
 
  # Classic APIs continue to work as normal, but raise warnings
  # whenever data is accessed through creation of SourceData object.