Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

implemented a new service to be used in DCALIGN cooking, #380

Open
wants to merge 5 commits into
base: development
Choose a base branch
from
Open
Show file tree
Hide file tree
Changes from 3 commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
19 changes: 19 additions & 0 deletions etc/bankdefs/hipo4/neuralnetwork.json
Original file line number Diff line number Diff line change
@@ -1,4 +1,23 @@
[
{
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can the new bank be added at the bottom of the file, so that they are ordered by item?

"name": "MLDC::dc",
"group": 4200,
"item" : 5,
"info": "bank with DC information to trina neural networks",
"entries": [
{"name":"id", "type":"S", "info":"the ID"},
{"name":"value", "type":"S", "info":"value"}
]
},
{
"name": "MLDC::tracks",
"group": 4200,
"item" : 6,
"info": "track information",
"entries": [
{"name":"bytes", "type":"B", "info":"byte array of composite format"}
]
},
{
"name": "nn::tracks",
"group": 4200,
Expand Down
2 changes: 2 additions & 0 deletions etc/services/dcalign.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -51,6 +51,8 @@ services:
name: RICH
- class: org.jlab.service.rtpc.RTPCEngine
name: RTPC
- class: org.jlab.service.mltn.MLDCEngine
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

dcalign.yaml is really only used for straight tracks/alignment runs. The new service should be added to data-ai.yaml or data-cv.yaml or data-aicv.yaml.

name: MLDC
configuration:
global:
variation: rgb_spring2019
Expand Down
Original file line number Diff line number Diff line change
@@ -0,0 +1,165 @@
/*
* Click nbfs://nbhost/SystemFileSystem/Templates/Licenses/license-default.txt to change this license
* Click nbfs://nbhost/SystemFileSystem/Templates/Classes/Class.java to edit this template
*/
package org.jlab.service.mltn;

import java.nio.ByteBuffer;
import java.nio.ByteOrder;
import java.util.ArrayList;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import org.jlab.clas.reco.ReconstructionEngine;
import org.jlab.io.base.DataBank;
import org.jlab.io.base.DataEvent;
import org.jlab.io.hipo.HipoDataBank;
import org.jlab.jnp.hipo4.data.Bank;
import org.jlab.jnp.hipo4.data.CompositeBank;

/**
*
* @author gavalian
*/
public class MLDCEngine extends ReconstructionEngine {


public MLDCEngine(){
super("MLDC","gavalian","1.0");
}

@Override
public boolean init() {
return true;
}

@Override
public boolean processDataEvent(DataEvent de) {

List<ByteBuffer> trackBuffers = new ArrayList<>();

if(de.hasBank("TimeBasedTrkg::TBTracks")&&de.hasBank("TimeBasedTrkg::TBClusters")){
//System.out.println("writing trakcing bank");
DataBank tbt = de.getBank("TimeBasedTrkg::TBTracks");
DataBank tbc = de.getBank("TimeBasedTrkg::TBClusters");
ByteBuffer bb = getTracks(tbt,tbc,10);
trackBuffers.add(bb);
/*DataBank output = de.createBank("MLDC::tracks", bb.array().length);
for(int j = 0; j < bb.array().length; j++)
output.setByte("bytes", j,bb.array()[j]);
de.appendBank(output);*/
}

if(de.hasBank("TimeBasedTrkg::AITracks")&&de.hasBank("TimeBasedTrkg::AIClusters")){
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Note that the banks with the AI prefix are created only when data-aicv.yaml is used. When data-ai.yaml is used, the tracking banks have the TB prefix even if AI tracking is used.

DataBank tbt = de.getBank("TimeBasedTrkg::AITracks");
DataBank tbc = de.getBank("TimeBasedTrkg::AIClusters");
ByteBuffer bb = getTracks(tbt,tbc,20);
trackBuffers.add(bb);
}

if(trackBuffers.size()>0){
int[] lengths = new int[trackBuffers.size()];
int[] offsets = new int[trackBuffers.size()];
int size = 0;
for(int k = 0; k < offsets.length; k++){
lengths[k] = trackBuffers.get(k).array().length;
offsets[k] = size;
size += trackBuffers.get(k).array().length;
}
byte[] buffer = new byte[size];
for(int k = 0; k < trackBuffers.size(); k++){
System.arraycopy(trackBuffers.get(k).array(), 0, buffer, offsets[k], lengths[k]);
}
DataBank output = de.createBank("MLDC::tracks", buffer.length);
for(int j = 0; j < buffer.length; j++)
output.setByte("bytes", j,buffer[j]);
de.appendBank(output);
}

if(de.hasBank("DC::tdc")==true){
DataBank bank = de.getBank("DC::tdc");
DataBank output = de.createBank("MLDC::dc", bank.rows());

HipoDataBank hbank = (HipoDataBank) bank;
byte[] sector = hbank.getByte("sector");
byte[] layer = hbank.getByte("layer");
short[] wire = hbank.getShort("component");
int[] tdc = hbank.getInt("TDC");

for(int j = 0; j < bank.rows(); j++){
int id = (sector[j]-1)*(112*36) + (layer[j]-1)*112 + (wire[j]-1);
output.setShort("id", j, (short) id);
output.setShort("value", j, (short) tdc[j]);
}
de.appendBank(output);
}
return true;
}

public static Map<Integer,Integer> getMap(DataBank bank){
Map<Integer,Integer> map = new HashMap<>();
int[] ids = bank.getInt("id");
for(int j = 0; j < ids.length; j++)
map.put(ids[j], j);
return map;
}


public static ByteBuffer getTracks(DataBank trkg, DataBank clusters, int status){
Map<Integer,Integer> map = getMap(clusters);
int size = trkg.rows();
int bsize = 110;
byte[] bytes = new byte[size*bsize];
ByteBuffer b = ByteBuffer.wrap(bytes);
b.order(ByteOrder.LITTLE_ENDIAN);
HipoDataBank ht = (HipoDataBank) trkg;
HipoDataBank hc = (HipoDataBank) clusters;
int[] cid = new int[6];
for(int j = 0; j < size; j++){
int offset = j*bsize;
int charge = (int) ht.getByte("q",j);
int ps = status + (charge<0?2:1);
//System.out.printf(" row = %4d , status = %5d, charge = %3d, pstat = %5d\n",
// j,status,charge,ps);
b.putShort(offset+0, (short) ps);
b.putFloat(offset+2, 0.0f);
b.putShort(offset+6, (short) ht.getByte("sector",j));
b.putShort(offset+8, (short) ht.getByte("q",j));
b.putFloat(offset+10, ht.getFloat("chi2", j));
b.putFloat(offset+14, ht.getFloat("p0_x", j));
b.putFloat(offset+18, ht.getFloat("p0_y", j));
b.putFloat(offset+22, ht.getFloat("p0_z", j));

b.putFloat(offset+26, ht.getFloat("Vtx0_x", j));
b.putFloat(offset+30, ht.getFloat("Vtx0_y", j));
b.putFloat(offset+34, ht.getFloat("Vtx0_z", j));

cid[0] = ht.getShort("Cluster1_ID", j);
cid[1] = ht.getShort("Cluster2_ID", j);
cid[2] = ht.getShort("Cluster3_ID", j);
cid[3] = ht.getShort("Cluster4_ID", j);
cid[4] = ht.getShort("Cluster5_ID", j);
cid[5] = ht.getShort("Cluster6_ID", j);

b.putInt(offset+38, cid[0]);
b.putInt(offset+42, cid[1]);
b.putInt(offset+46, cid[2]);
b.putInt(offset+50, cid[3]);
b.putInt(offset+54, cid[4]);
b.putInt(offset+58, cid[5]);


for(int i = 0; i < 6; i++){
if(map.containsKey(cid[i])==true){
float avg = hc.getFloat("avgWire", map.get(cid[i]));
b.putFloat(offset+i*4+62, avg);
b.putFloat(offset+i*4+62+4*6, avg);
} else {
b.putFloat(offset+i*4+62, 0.0f);
b.putFloat(offset+i*4+62+4*6, 0.0f);
}
}
}
return b;
}
}
Loading