Support a single anonymous stream (AKA 'no streaming').

Phys/Tesla/python/Tesla/Configuration.py

@@ -714,9 +714,9 @@ class Tesla(LHCbConfigurableUser):
         online = mode == 'Online'
         raw_format = self.getProp('RawFormatVersion')
 
-        if not streams and not simulation:
-            log.warning('Not streaming and not running on MC, nothing to do!')
-            return
+        # Treat the no-streaming case as a single anonymous stream
+        if not streams:
+            streams = {'': {'lines': self.getProp('TriggerLines')}}
 
         # Setting up online
         decoders_seq = GaudiSequencer('TeslaDecoders')
@@ -755,12 +755,12 @@ class Tesla(LHCbConfigurableUser):
         )
         unpackers_seq.Members.append(unpack_psandvs)
 
-        # TODO cover the no-streams case, when `streams == {}`
         stream_sequences = []
         for stream_name in streams:
             stream_seq = self._configureOutputTurboSPStream(
                 stream_name,
-                streams[stream_name]['lines']
+                streams[stream_name]['lines'],
+                prpacking
             )
             stream_sequences.append(stream_seq)
 
@@ -768,7 +768,7 @@ class Tesla(LHCbConfigurableUser):
         address_killer = AddressKillerAlg()
 
         streaming_seq = GaudiSequencer(
-            'TeslaStreamSequence',
+            'TeslaStreamsSequence',
             Members=(
                 [decoders_seq, unpackers_seq] +
                 stream_sequences +
@@ -778,13 +778,20 @@ class Tesla(LHCbConfigurableUser):
         )
         return streaming_seq
 
-    def _configureOutputTurboSPStream(name, lines):
-        """Copy line outputs for this stream to a stream-specific location.
+    def _configureOutputTurboSPStream(self, name, lines, packing):
+        """Return a sequence for streaming the lines into the named location.
+
+        Copy line outputs for this stream to a stream-specific location.
+
+        Keyword arguments:
+        name -- Name of the stream
+        lines -- List of HLT2 line names that belong in the stream
+        packing -- Instance of PersistRecoPacking
         """
         output_prefix = name.title()
         decisions = [l + 'Decision' for l in lines]
         # Naming convention for stream-specific algorithms
-        namer = lambda x: '{0}ForStream{1}'.format(x, name)
+        namer = lambda x: '{0}ForStream{1}'.format(x, output_prefix)
         pack = self.getProp('Pack')
         write_fsr = self.getProp('WriteFSR')
 
@@ -798,8 +805,12 @@ class Tesla(LHCbConfigurableUser):
             namer('Hlt2_HLTFilter'),
             Code=filter_code
         )
-        stream_seq.Members.append(filter)
+        # Don't filter an anonymous stream by default
+        if output_prefix or self.getProp('HDRFilter'):
+            stream_seq.Members.append(filter)
 
+        # No need to clone if the output prefix is empty (everything stays
+        # under /Event/Turbo)
         # XXX change this to the TCK svc once finished debugging
         persistence_svc = HltLinePersistenceSvc()
         container_cloner = CopyLinePersistenceLocations(
@@ -822,41 +833,47 @@ class Tesla(LHCbConfigurableUser):
             ],
             IgnoreFilterPassed=True
         )
-        stream_seq.Members.append(copy_line_outputs_seq)
+        if output_prefix:
+            stream_seq.Members.append(copy_line_outputs_seq)
 
+        tes_root = '/Event'
+        # /Event/<stream name>
+        stream_base = os.path.join(tes_root, output_prefix).rstrip('/')
         required_output_locations = [
-            '/Event/DAQ/ODIN#1',
-            '/Event/Rec/Summary#1'
+            os.path.join(tes_root, 'DAQ/ODIN#1'),
+            os.path.join(tes_root, 'Rec/Summary#1')
         ]
         optional_output_locations = []
-        stream_base = os.path.join('/Event', output_prefix)
         if pack:
             packers = []
 
-            # Pack the PersistReco locations that we've cloned
+            # Pack the PersistReco locations for this stream
             prpacking_inputs = {
-                k: v.replace('/Event', stream_base)
-                for k, v in prpacking.outputs.items()
+                k: v.replace(tes_root, stream_base)
+                for k, v in packing.outputs.items()
             }
-            stream_base += '/Turbo'
+            # /Event/<stream name>/Turbo
+            stream_base = os.path.join(stream_base, 'Turbo')
             prpacking_stream = PersistRecoPacking(
                 self.getProp('DataType'),
                 inputs=prpacking_inputs
             )
-            packers += prpacking_stream.packers(identifier=name)
+            packers += prpacking_stream.packers(identifier=namer(''))
             for packer in packers:
                 # Put the packed containers under the stream
                 packer.OutputName = packer.OutputName.replace(
-                    '/Event', stream_base
+                    tes_root, stream_base
                 )
 
             # Pack everything else that we've cloned
             psandvs_packer = PackParticlesAndVertices(
-                name=namer('TurboPacker'),
+                namer('TurboPacker'),
                 InputStream=stream_base,
                 VetoedContainers=prpacking_stream.inputs.values()
             )
-            packers.append(psandvs_packer)
+            # Packed locations already exist at the anonymous stream location
+            if output_prefix:
+                packers.append(psandvs_packer)
 
             for packer in packers:
                 packer.OutputLevel = min(self.getProp('OutputLevel'),
@@ -877,16 +894,21 @@ class Tesla(LHCbConfigurableUser):
                 # os.path.join(stream_base, 'pRec#*')
             ]
         else:
+            # /Event/<stream name>/Turbo
+            stream_base = os.path.join(stream_base, 'Turbo')
             # Save everything under the stream prefix
             optional_output_locations += [
                 os.path.join(stream_base, '#*')
             ]
 
-        fname = '{0}{1}{2}'.format(
-            self.getProp('outputPrefix'),
-            name,
-            self.getProp('outputSuffix')
-        )
+        if output_prefix:
+            fname = '{0}{1}{2}'.format(
+                self.getProp('outputPrefix'),
+                output_prefix,
+                self.getProp('outputSuffix')
+            )
+        else:
+            fname = self.getProp('outputFile')
 
         writer = OutputStream(namer(self.writerName))
         writer.ItemList = required_output_locations