Add support for realistic RICH PMT encoding and decoding

• Adds various utility classes that use the readout mapping conditions added in lhcb-conddb/SIMCOND!115 (merged).
• Updates the RICH decoding algorithm to support the realistic PMT data format.
• Support for the data format used by current MC samples (RichSmartID streaming) is retained of course, and processing existing MC samples does not require the above DB update. That update is only required to when encoding or decoding the new format.

Requires Moore!586 (merged) and Boole!289 (merged)

Rich/RichFutureDAQ/src/component/RichRawBankDecoder.cpp

@@ -336,9 +336,8 @@ OutData RawBankDecoder::operator()( const LHCb::RawEvent& rawEvent,      //
           }
           // Finally, fill the output data from the list of SmartIDs
           fillDecodedData( decodedIDs, decodedData );
-          // clear ID vector for next bank
-          decodedIDs.clear();
         }
+
         // catch and report any exceptions during decoding
         catch ( const GaudiException& expt ) {
           // count errors
@@ -419,10 +418,8 @@ void RawBankDecoder::decodeToSmartIDs_MaPMT1( const LHCb::RawBank& bank,
                            " != DecodedSize:" + std::to_string( nPackedSizeW + nPayloadWords ) );
   }
 
-  // Guess at reserve size for ID vector based on payload size
-  // Note this is only correct for all ZS frames.
-  // For frames with NZS can be an underestimate, so add a scale factor to compensate
-  // decodedIDs.reserve( nPayloadWords * 2 );
+  // Clear ID vector for this bank
+  decodedIDs.clear();
 
   // finally loop over payload words and decode hits
   // note iterator starts from where the above header loop ended...
@@ -560,6 +557,9 @@ void RawBankDecoder::decodeToSmartIDs_StreamIDs( const LHCb::RawBank& bank,
   // Data bank size in 32 bit words
   const auto bankSize = bank.size() / 4;
 
+  // Clear ID vector for this bank
+  decodedIDs.clear();
+
   // reserve size in ID vector for number of hits (bank size here)
   decodedIDs.reserve( bankSize );