Size: 191
Comment:
|
Size: 3217
Comment:
|
Deletions are marked like this. | Additions are marked like this. |
Line 3: | Line 3: |
The plan is: create a MarlinTPC processor chain that completely reflects the MultiFit reconstruction chain. | The plan is: create a MarlinTPC processor chain that completely reflects the !MultiFit reconstruction chain. |
Line 8: | Line 8: |
== The chain == | == The processor chain == The first thing you need is the raw data from your detector. This LCIO file might have a single collection inside of the type '''Tracker'Raw'Data''' with the default name ''''TPCRawData''''. The list is not yet complete, possibly several processors need to be added. In a table like fashion. First the steps specific for ADC based electronics: || step || Processor name || Data output type || Default collection name || Conditions objects || || 1. Data conversion || ''!TrackerRawDataToDataConverterProcessor'' || '''!TrackerData''' || '''''TPCConvertedData''''' || Electronics || || 2. Pedestal Subtraction || ''!PedestalSubtractorProcessor'' || '''!TrackerData''' || '''''TPCData''''' || Pedestal || Then the more generic reconstruction: || step || Processor name || Data output type || Default collection name || Conditions objects || || 3. Pulse finding || ''!PulseFinderProcessor'' || '''!TrackerPulse''' || '''''TPCPulses''''' || !ConversionFactor(?)|| || 4. Channel mapping || ''!ChannelMapperProcessor'' || '''!TrackerPulse''' || '''''TPCPulsesMapped''''' || Channel Mapping || || 5. Hit finding || ''!HitTrackFinderTopoProcessor'' || '''!TrackerHit''' || '''''TPCHits''''' || ? || The next step is actually dependent on which kind of track model is used. The track model (straight line, circle, helix) determines which kind of track is searched and reconstructed. The given chain is for straight lines. || step || Processor name || Data output type || Default collection name || Conditions objects || || 6. Track finding || ''!TrackFinderHoughProcessor'' || '''Track''' || '''''TPCTrackCandidates'''''|| || || 7. Track fitting || ''!LinearRegressionProcessor'' || '''Track''' || '''''TPCSeedTracks''''''|| || == Conditions processing == For the chain to work several conditions objects are needed, although for now they may be implemented as processor parameters. The objects are commonly '''LCGenericObjects''', defined in /tpcconddata Predefined/static input is: || Description || Data type || Default name || || Electronics || '''LCGenericObject''' || ? || || Channel Mapping || '''LCGenericObject''' || '''''TPCADCChannelMapping''''' || Actual processing determines further input to the reconstruction: || step || Processor name || Data output type || Default collection name || || Pedestal Calculation || ''!PedestalCalculator'' || '''LCGenericObject''' || '''''TPCPedestal''''' || || adc2primary_electrons || ??? || ??? || '''''!ConversionFactor''''' || |
The objective
The plan is: create a MarlinTPC processor chain that completely reflects the MultiFit reconstruction chain.
Use data taken with the MediTPC.
The processor chain
The first thing you need is the raw data from your detector. This LCIO file might have a single collection inside of the type Tracker'Raw'Data with the default name 'TPCRawData'. The list is not yet complete, possibly several processors need to be added.
In a table like fashion. First the steps specific for ADC based electronics:
step |
Processor name |
Data output type |
Default collection name |
Conditions objects |
1. Data conversion |
TrackerRawDataToDataConverterProcessor |
TrackerData |
TPCConvertedData |
Electronics |
2. Pedestal Subtraction |
PedestalSubtractorProcessor |
TrackerData |
TPCData |
Pedestal |
Then the more generic reconstruction:
step |
Processor name |
Data output type |
Default collection name |
Conditions objects |
3. Pulse finding |
PulseFinderProcessor |
TrackerPulse |
TPCPulses |
ConversionFactor(?) |
4. Channel mapping |
ChannelMapperProcessor |
TrackerPulse |
TPCPulsesMapped |
Channel Mapping |
5. Hit finding |
HitTrackFinderTopoProcessor |
TrackerHit |
TPCHits |
? |
The next step is actually dependent on which kind of track model is used. The track model (straight line, circle, helix) determines which kind of track is searched and reconstructed. The given chain is for straight lines.
step |
Processor name |
Data output type |
Default collection name |
Conditions objects |
6. Track finding |
TrackFinderHoughProcessor |
Track |
TPCTrackCandidates |
|
7. Track fitting |
LinearRegressionProcessor |
Track |
TPCSeedTracks |
|
Conditions processing
For the chain to work several conditions objects are needed, although for now they may be implemented as processor parameters. The objects are commonly LCGenericObjects, defined in /tpcconddata Predefined/static input is: Description Data type Default name Electronics LCGenericObject ? Channel Mapping LCGenericObject TPCADCChannelMapping Actual processing determines further input to the reconstruction: step Processor name Data output type Default collection name Pedestal Calculation PedestalCalculator LCGenericObject TPCPedestal adc2primary_electrons ??? ??? ConversionFactor