feat: add CLARA test (#878)

baltzell · web-flow · commit ccee3f5ca8aa · 2025-10-14T19:44:23.000-04:00
* add new YAML file * use stock schema in yamls * .gitignore clara installation * run clara test * disable validation workflows * add denoising engine * disable validatin * disable validation * reactivate * schema dir in yaml * set memory limit * keep memory contiguous * benchmark * fail fast * let it be null, no waiting * configure threads * fall back to $HOME * add test * single-threaded * reduce verbosity * add gitlab clara test * fix ci test * remove benchmarks * fix file name * asdf * Revert "asdf" This reverts commit f648034. * yes * fix * needs cvmfs for ccdb
diff --git a/.github-ci.yaml b/.github-ci.yaml
@@ -0,0 +1 @@
+.github/workflows/ci.yml
diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
@@ -39,7 +39,9 @@ jobs:
         id: cache
         with:
           key: raw_test_data # fixed key will always hit; clear cache to trigger cache miss
-          path: clas_005038.evio.00000
+          path: |
+            clas_005038.evio.00000
+            clas_021559.evio.00001
           lookup-only: true
       - name: install xrootd-client
         if: ${{ steps.cache.outputs.cache-hit != 'true' }}
@@ -50,6 +52,7 @@ jobs:
         if: ${{ steps.cache.outputs.cache-hit != 'true' }}
         run: |
           xrdcp xroot://sci-xrootd.jlab.org///osgpool/hallb/clas12/validation/clas_005038.evio.00000 ./
+          xrdcp xroot://sci-xrootd.jlab.org///osgpool/hallb/clas12/validation/raw/rg-l/clas_021559.evio.00001 ./
 
   # build
   #############################################################################
@@ -86,14 +89,18 @@ jobs:
         if: ${{ github.ref_type == 'tag' }}
         run: libexec/version-bump.sh ${{ github.ref_name }}
       - name: build
-        run: ./build-coatjava.sh --xrootd --no-progress -T${{ env.nthreads }}
+        run: ./build-coatjava.sh --clara --xrootd --no-progress -T${{ env.nthreads }}
       - name: tar # tarball to preserve permissions
-        run: tar czvf coatjava.tar.gz coatjava
+        run: |
+          tar czvf coatjava.tar.gz coatjava
+          tar czvf clara.tar.gz clara
       - uses: actions/upload-artifact@v4
         with:
           name: build_${{ matrix.runner }}
           retention-days: 1
-          path: coatjava.tar.gz
+          path: |
+              coatjava.tar.gz
+              clara.tar.gz
 
   # tests
   #############################################################################
@@ -147,13 +154,50 @@ jobs:
       - uses: actions/cache/restore@v4
         with:
           key: raw_test_data
-          path: clas_005038.evio.00000
+          path: |
+            clas_005038.evio.00000
+            clas_021559.evio.00001
       - name: untar build
         run: tar xzvf coatjava.tar.gz
       - name: run test
         run: |
+          ls -lhtr
           ./coatjava/bin/decoder -n 10000 -o dog.hipo ./clas_005038.evio.00000
 
+  test_clara:
+    needs: [ build, download_test_data ]
+    strategy:
+      fail-fast: true
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v5
+      - name: Set up JDK
+        uses: actions/setup-java@v5
+        with:
+          java-version: ${{ env.java_version }}
+          distribution: ${{ env.java_distribution }}
+      - name: setup cvmfs
+        uses: cvmfs-contrib/github-action-cvmfs@v5
+        with:
+          cvmfs_repositories: 'oasis.opensciencegrid.org'
+      - uses: actions/download-artifact@v5
+        with:
+          name: build_ubuntu-latest
+      - uses: actions/cache/restore@v4
+        with:
+          key: raw_test_data
+          path: |
+            clas_005038.evio.00000
+            clas_021559.evio.00001
+      - name: untar build
+        run: |
+          tar xzvf clara.tar.gz
+      - name: run test
+        run: |
+          ls -lhtr
+          ./clara/plugins/clas12/bin/run-clara -y ./etc/services/rgl-clarode.yml -t ${{ env.nthreads }} -n 200 -c ./clara -o ./tmp ./clas_021559.evio.00001
+          ls -lhtr
+
   test_coatjava:
     needs: [ build ]
     strategy:
diff --git a/.gitignore b/.gitignore
@@ -1,5 +1,6 @@
 # installed:
 /coatjava
+/clara
 target
 /publish
 /docs/javadoc/*
diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
@@ -23,13 +23,15 @@ stages:
 build:
   stage: build
   script:
-    - ./build-coatjava.sh -T$JL_RUNNER_AVAIL_CPU --quiet --no-progress
+    - ./build-coatjava.sh --clara -T$JL_RUNNER_AVAIL_CPU --quiet --no-progress
     - tar -czf coatjava.tar.gz coatjava
+    - tar -czf clara.tar.gz clara
   artifacts:
     when: always
     expire_in: 1 day
     paths:
       - coatjava.tar.gz
+      - clara.tar.gz
 
 depana:
   allow_failure: true
@@ -38,6 +40,18 @@ depana:
     - libexec/dependency-tree.sh
     - libexec/dependency-analysis.sh
 
+download:
+  stage: build
+  script:
+    - xrdcp xroot://sci-xrootd.jlab.org///osgpool/hallb/clas12/validation/clas_005038.evio.00000 ./
+    - xrdcp xroot://sci-xrootd.jlab.org///osgpool/hallb/clas12/validation/raw/rg-l/clas_021559.evio.00001 ./
+  artifacts:
+    when: always
+    expire_in: 1 day
+    paths:
+      - clas_005038.evio.00000
+      - clas_021559.evio.00000
+
 spotbugs:
   stage: test
   needs: [build]
@@ -100,6 +114,18 @@ eb:
       - ARG: electronneutronC
       - ARG: electronFTpion
 
+clara:
+  allow_failure: true
+  stage: test
+  needs: [build]
+  dependencies: [build]
+  script:
+    - >
+      tar -xzf clara.tar.gz &&
+      ./clara/plugins/clas12/bin/run-clara -c ./clara
+        -t $JL_RUNNER_AVAIL_CPU -y ./etc/services/rgl-clarode.yml
+        -n 100 -o ./out clas_021*.hipo
+
 profile:
   stage: test
   needs: [build]
diff --git a/bin/run-clara b/bin/run-clara
@@ -72,7 +72,7 @@ done
 [ $(cat $CLARA_USER_DATA/filelist.txt | wc -l) -gt 0 ] || error "Found no input files"
 
 # Set some JVM options:
-export JAVA_OPTS="$JAVA_OPTS -XX:+IgnoreUnrecognizedVMOptions"
+export JAVA_OPTS="$JAVA_OPTS -Xmx${threads}g -XX:+IgnoreUnrecognizedVMOptions"
 export JAVA_OPTS="$JAVA_OPTS -Djava.io.tmpdir=$CLARA_USER_DATA -Dorg.sqlite.tmpdir=$CLARA_USER_DATA"
 # Set verbosity:
 [ -z ${quiet+x} ] && stub=fine || stub=info
diff --git a/etc/services/rgl-clarode.yml b/etc/services/rgl-clarode.yml
@@ -0,0 +1,59 @@
+mime-types:
+  - binary/data-hipo
+configuration:
+  global:
+    variation: rge_spring2024
+    timestamp: 10/08/2025-17:00:00
+  io-services:
+    writer:
+      schema_dir: full
+  services:
+    MAGFIELDS:
+      magfieldSolenoidMap: Symm_solenoid_r601_phi1_z1201_13June2018.dat
+      magfieldTorusMap: Full_torus_r251_phi181_z251_25Jan2021.dat
+io-services:
+  reader:
+    class: org.jlab.io.clara.DecoderReader
+    name: DecoderReader
+  writer:
+    class: org.jlab.io.clara.DecoderWriter
+    name: DecoderWriter
+services:
+  - class: org.jlab.service.ai.DCDenoiseEngine
+    name: DCDN
+  - class: org.jlab.clas.swimtools.MagFieldsEngine
+    name: MAGFIELDS
+  - class: org.jlab.rec.ft.cal.FTCALEngine
+    name: FTCAL
+  - class: org.jlab.rec.ft.hodo.FTHODOEngine
+    name: FTHODO
+  - class: org.jlab.rec.ft.FTEBEngine
+    name: FTEB
+  - class: org.jlab.service.dc.DCHBClustering
+    name: DCCR
+  - class: org.jlab.service.dc.DCHBPostClusterConv
+    name: DCHB
+  - class: org.jlab.service.ftof.FTOFHBEngine
+    name: FTOFHB
+  - class: org.jlab.service.ec.ECEngine
+    name: EC
+  - class: org.jlab.service.ctof.CTOFEngine
+    name: CTOF
+  - class: org.jlab.service.cnd.CNDCalibrationEngine
+    name: CND
+  - class: org.jlab.service.band.BANDEngine
+    name: BAND
+  - class: org.jlab.service.htcc.HTCCReconstructionService
+    name: HTCC
+  - class: org.jlab.service.ltcc.LTCCEngine
+    name: LTCC
+  - class: org.jlab.service.eb.EBHBEngine
+    name: EBHB
+  - class: org.jlab.service.dc.DCTBEngine
+    name: DCTB
+  - class: org.jlab.service.ftof.FTOFTBEngine
+    name: FTOFTB
+  - class: org.jlab.service.eb.EBTBEngine
+    name: EBTB
+  - class: org.jlab.rec.rich.RICHEBEngine
+    name: RICH
diff --git a/install-clara b/install-clara
@@ -34,7 +34,7 @@ function litter() {
         then
             echo -e "\nSee temporary build directory for more info:\n\t$tmp_dir"
         else
-            rm -rf $tmp_dir && echo -e "\nRemoved temporary build directory:\n\t$tmp_dir"
+            rm -rf $tmp_dir
         fi
     fi
 }
@@ -117,7 +117,6 @@ fi
 tmp_dir=$(mktemp -d tmp.install-clara.XXXXXX)
 [ $? -ne 0 ] && error "Cannot create temporary directory:  $tmp_dir"
 tmp_dir=$(cd $tmp_dir && pwd)
-echo -e "\nCreating temporary build directory:\n\t$tmp_dir"
 cd $tmp_dir
 
 # CLARA:
diff --git a/reconstruction/ai/src/main/java/org/jlab/service/ai/DCDenoiseEngine.java b/reconstruction/ai/src/main/java/org/jlab/service/ai/DCDenoiseEngine.java
@@ -17,7 +17,7 @@
 import ai.djl.translate.TranslateException;
 import java.io.IOException;
 import java.util.concurrent.BlockingQueue;
-import java.util.concurrent.LinkedBlockingQueue;
+import java.util.concurrent.ArrayBlockingQueue;
 
 import org.jlab.clas.reco.ReconstructionEngine;
 import org.jlab.io.base.DataBank;
@@ -28,6 +28,7 @@ public class DCDenoiseEngine extends ReconstructionEngine {
 
     final static String[] BANK_NAMES = {"DC::tot","DC::tdc"};
     final static String CONF_THRESHOLD = "threshold";
+    final static String CONF_THREADS = "threads";
     final static int LAYERS = 36;
     final static int WIRES = 112;
 
@@ -39,11 +40,11 @@ public class DCDenoiseEngine extends ReconstructionEngine {
     public static class PredictorPool {
         final BlockingQueue<Predictor> pool;
         public PredictorPool(int size, ZooModel model) {
-            pool = new LinkedBlockingQueue<>(size);
-            for (int i=0; i<size; i++) pool.offer(model.newPredictor());
+            pool = new ArrayBlockingQueue<>(size);
+            for (int i=0; i<size; i++) pool.add(model.newPredictor());
         }
         public Predictor get() throws InterruptedException {
-            return pool.take();
+            return pool.poll();
         }
         public void put(Predictor p) {
             if (p != null) pool.offer(p);
@@ -56,6 +57,9 @@ public DCDenoiseEngine() {
 
     @Override
     public boolean init() {
+        System.setProperty("ai.djl.pytorch.num_interop_threads", "1");
+        System.setProperty("ai.djl.pytorch.num_threads", "1");
+        System.setProperty("ai.djl.pytorch.graph_optimizer", "false");
         if (getEngineConfigString(CONF_THRESHOLD) != null)
             threshold = Float.parseFloat(getEngineConfigString(CONF_THRESHOLD));
         try {
@@ -67,14 +71,23 @@ public boolean init() {
                 .optProgress(new ProgressBar())
                 .build();
             model = criteria.loadModel();
-            predictors = new PredictorPool(64, model);
+            int threads = Integer.parseInt(getEngineConfigString(CONF_THREADS,"64"));
+            predictors = new PredictorPool(threads, model);
             return true;
         } catch (NullPointerException | MalformedModelException | IOException | ModelNotFoundException ex) {
             System.getLogger(DCDenoiseEngine.class.getName()).log(System.Logger.Level.ERROR, (String) null, ex);
             return false;
         }
     }
 
+    public static void main(String args[]){
+        DCDenoiseEngine dn = new DCDenoiseEngine();
+        dn.init();
+        for (int i=0; i<10000; i++) {
+            dn.processFakeEvent();
+        }
+    }
+    
     @Override
     public boolean processDataEvent(DataEvent event) {
 
