VERITAS-Observatory · GernotMaier · Jun 26, 2026 · Jun 26, 2026 · Jun 26, 2026
diff --git a/README.md b/README.md
@@ -53,7 +53,7 @@ Set the following environment variables before running the scripts:
 | `$VERITAS_IRFPRODUCTION_DIR` | Path to the IRF production output directory |
 | `$VERITAS_USER_DATA_DIR` | Path to user-specific data output |
 | `$VERITAS_USER_LOG_DIR` | Path to log files |
-| `$EVNDISP_MLSYS` | Path to the [Eventdisplay-ML](https://github.com/Eventdisplay/Eventdisplay-ML) installation (optional) |
+| `$EVNDISP_ML_SYS` | Path to the [Eventdisplay-ML](https://github.com/Eventdisplay/Eventdisplay-ML) installation (optional) |
 | `$EVNDISP_ML_ENV` | Name of the conda environment for Eventdisplay-ML (default: `eventdisplay_ml`) |
 
 See [`./scripts/set_environment.sh`](./scripts/set_environment.sh) for a template with all variables configured for the DESY computing environment. Submission commands for various batch systems are available in [submissionCommands.dat](./scripts/submissionCommands.dat).

diff --git a/docs/changes/153.maintenance.md b/docs/changes/153.maintenance.md
@@ -0,0 +1 @@
+Centralize XGB helper-script Conda activation and environment logging.
diff --git a/scripts/ANALYSIS.dispXGB.sh b/scripts/ANALYSIS.dispXGB.sh
@@ -60,6 +60,7 @@ rm -f "${LOGDIR}"/x* 2>/dev/null
 
 # Job submission script
 SUBSCRIPT="$(dirname "$0")/helper_scripts/ANALYSIS.dispXGB_sub"
+HELPER_SCRIPTS_DIR="$(cd "$(dirname "$0")/helper_scripts" && pwd)"
 TIMETAG=$(date +"%s")
 CREATED_LOGSUBDIRS=" "
 
@@ -82,6 +83,8 @@ do
         -e "s|XXGB|$XGB|" \
         -e "s|XGB_TTYPE|$XGB_TYPE|" \
         -e "s|ANALYSISTYPE|$ANALYSIS_TYPE|" \
+        -e "s|HHELPER_SCRIPTS_DIR|$HELPER_SCRIPTS_DIR|" \
+        -e "s|EENV_SNAPSHOT_DIR|$LOGDIR|" \
         -e "s|OODIR|$ODIR|" "$SUBSCRIPT".sh > "$FSCRIPT".sh
 
     chmod u+x "$FSCRIPT.sh"

diff --git a/scripts/IRF.dispXGB.sh b/scripts/IRF.dispXGB.sh
@@ -58,6 +58,7 @@ rm -f "${LOGDIR}"/x* 2>/dev/null
 
 # Job submission script
 SUBSCRIPT="$(dirname "$0")/helper_scripts/IRF.dispXGB_sub"
+HELPER_SCRIPTS_DIR="$(cd "$(dirname "$0")/helper_scripts" && pwd)"
 
 for FILE in $FILES
 do
@@ -69,6 +70,8 @@ do
         -e "s|XXGB|$XGB|" \
         -e "s|XGB_TTYPE|$XGB_TYPE|" \
         -e "s|ANALYSISTYPE|$ANALYSIS_TYPE|" \
+        -e "s|HHELPER_SCRIPTS_DIR|$HELPER_SCRIPTS_DIR|" \
+        -e "s|EENV_SNAPSHOT_DIR|$LOGDIR|" \
         -e "s|OODIR|$ODIR|" "$SUBSCRIPT".sh > "$FSCRIPT".sh
 
     chmod u+x "$FSCRIPT.sh"

diff --git a/scripts/IRF.trainXGBforAngularReconstruction.sh b/scripts/IRF.trainXGBforAngularReconstruction.sh
@@ -102,12 +102,15 @@ done
 echo "FILE LIST: ${MSCWLIST}"
 
 SUBSCRIPT="$(dirname "$0")/helper_scripts/IRF.trainXGBforAngularReconstruction_sub.sh"
+HELPER_SCRIPTS_DIR="$(cd "$(dirname "$0")/helper_scripts" && pwd)"
 
 echo "Processing Zenith = $ZA, Noise = $NOISE, Wobble = $WOBBLE"
 
 FSCRIPT="$LOGDIR/trainXGBStereoAnalysis.ID${RECID}.${EPOCH}.ATM${ATM}.${ZA}.sh"
 sed -e "s|OUTPUTDIR|$ODIR|" \
     -e "s|MSCWLIST|$MSCWLIST|" \
+    -e "s|HHELPER_SCRIPTS_DIR|$HELPER_SCRIPTS_DIR|" \
+    -e "s|EENV_SNAPSHOT_DIR|$LOGDIR|" \
     -e "s|NCORES|$ncore|" "$SUBSCRIPT" > "$FSCRIPT"
 
 chmod u+x "$FSCRIPT"

diff --git a/scripts/IRF.trainXGBforGammaHadronSeparationTraining.sh b/scripts/IRF.trainXGBforGammaHadronSeparationTraining.sh
@@ -116,6 +116,7 @@ get_run_prefix()
 
 # Job submission script
 SUBSCRIPT="$(dirname "$0")/helper_scripts/IRF.trainXGBforGammaHadronSeparation_sub.sh"
+HELPER_SCRIPTS_DIR="$(cd "$(dirname "$0")/helper_scripts" && pwd)"
 
 SIGNALLIST="${ODIR}/signal_files.list"
 rm -f "${SIGNALLIST}"
@@ -167,6 +168,8 @@ for (( i=0; i < NENE; i++ )); do
         -e "s|MODELPARA|$RUNPAR|" \
         -e "s|ENERGYBIN|$i|" \
 	-e "s|NCORES|$ncore|" \
+        -e "s|HHELPER_SCRIPTS_DIR|$HELPER_SCRIPTS_DIR|" \
+        -e "s|EENV_SNAPSHOT_DIR|$LOGDIR|" \
         -e "s|OUTPUTDIR|${ODIR}|" "$SUBSCRIPT" > "$FSCRIPT"
 
     chmod u+x "$FSCRIPT"

diff --git a/scripts/helper_scripts/ANALYSIS.dispXGB_sub.sh b/scripts/helper_scripts/ANALYSIS.dispXGB_sub.sh
@@ -10,6 +10,8 @@
 RUN=RRUN
 ODIR=OODIR
 env_name="${EVNDISP_ML_ENV:-eventdisplay_ml}"
+HELPER_SCRIPTS_DIR="HHELPER_SCRIPTS_DIR"
+ENV_SNAPSHOT_DIR="EENV_SNAPSHOT_DIR"
 XGB="XXGB"
 XGB_TYPE=XGB_TTYPE
 ANATYPE=ANALYSISTYPE
@@ -26,29 +28,10 @@ mkdir -p "$TEMPDIR"
 mkdir -p "${ODIR}"
 echo -e "Output files will be written to:\n ${ODIR}"
 
-check_conda_installation()
-{
-    if command -v conda &> /dev/null; then
-        echo "Found conda installation."
-    else
-        echo "Error: found no conda installation."
-        echo "exiting..."
-        exit
-    fi
-    env_info=$(conda info --envs)
-    if [[ "$env_info" == *"$env_name"* ]]; then
-        echo "Found conda environment '$env_name'"
-    else
-        echo "Error: the conda environment '$env_name' does not exist."
-        echo "exiting..."
-        exit
-    fi
-}
-
-check_conda_installation
-
-eval "$(conda shell.bash hook)"
-conda activate "$env_name"
+# shellcheck source=scripts/helper_scripts/UTILITY.conda_env.sh
+source "${HELPER_SCRIPTS_DIR}/UTILITY.conda_env.sh"
+evndisp_ml_setup_python_cache "$TEMPDIR" "$RUN"
+evndisp_ml_activate_conda "$env_name"
 
 # directory schema for preprocessed files
 getNumberedDirectory()
@@ -113,7 +96,6 @@ $ML_EXEC --input_file "$MSCW_FILE" \
     --model_prefix "$DISPDIR" \
     --output_file "$OFIL.root" > "${LOGFILE}" 2>&1
 
-python --version >> "${LOGFILE}"
-conda list -n "$env_name" >> "${LOGFILE}"
+evndisp_ml_log_environment "${LOGFILE}" "$env_name" "$ENV_SNAPSHOT_DIR"
 
 conda deactivate
diff --git a/scripts/helper_scripts/IRF.dispXGB_sub.sh b/scripts/helper_scripts/IRF.dispXGB_sub.sh
@@ -9,6 +9,8 @@
 MSCW_FILE=FFILE
 ODIR=OODIR
 env_name="${EVNDISP_ML_ENV:-eventdisplay_ml}"
+HELPER_SCRIPTS_DIR="HHELPER_SCRIPTS_DIR"
+ENV_SNAPSHOT_DIR="EENV_SNAPSHOT_DIR"
 XGB="XXGB"
 XGB_TYPE=XGB_TTYPE
 ANATYPE=ANALYSISTYPE
@@ -26,34 +28,10 @@ mkdir -p "$TEMPDIR"
 mkdir -p "${ODIR}"
 echo -e "Output files will be written to:\n ${ODIR}"
 
-check_conda_installation()
-{
-    if command -v conda &> /dev/null; then
-        echo "Found conda installation."
-    else
-        echo "Error: found no conda installation."
-        echo "exiting..."
-        exit
-    fi
-    env_info=$(conda info --envs)
-    if [[ "$env_info" == *"$env_name"* ]]; then
-        echo "Found conda environment '$env_name'"
-    else
-        echo "Error: the conda environment '$env_name' does not exist."
-        echo "exiting..."
-        exit
-    fi
-}
-
-check_conda_installation
-
-# Avoid plugin/cache race conditions
-export CONDA_NO_PLUGINS=true
-PYTHONPYCACHEPREFIX="${TEMPDIR}/pycache_$(basename "$MSCW_FILE" .root)"
-export PYTHONPYCACHEPREFIX
-mkdir -p "$PYTHONPYCACHEPREFIX"
-eval "$(conda shell.bash hook)"
-conda activate "$env_name"
+# shellcheck source=scripts/helper_scripts/UTILITY.conda_env.sh
+source "${HELPER_SCRIPTS_DIR}/UTILITY.conda_env.sh"
+evndisp_ml_setup_python_cache "$TEMPDIR" "$(basename "$MSCW_FILE" .root)"
+evndisp_ml_activate_conda "$env_name"
 
 if [[ ! -e ${MSCW_FILE} ]]; then
     echo "File ${MSCW_FILE} not found. Exiting."
@@ -104,5 +82,6 @@ $ML_EXEC --input_file "$MSCW_FILE" \
     --max_cores $MAXCORES \
     --output_file "$OFIL.root" > "${LOGFILE}" 2>&1
 
-python --version >> "${LOGFILE}"
-python -m pip freeze >> "${LOGFILE}" || echo "pip freeze failed, continuing" >> "${LOGFILE}"
+evndisp_ml_log_environment "${LOGFILE}" "$env_name" "$ENV_SNAPSHOT_DIR"
+
+conda deactivate
diff --git a/scripts/helper_scripts/IRF.trainXGBforAngularReconstruction_sub.sh b/scripts/helper_scripts/IRF.trainXGBforAngularReconstruction_sub.sh
@@ -8,6 +8,8 @@
 LLIST=MSCWLIST
 ODIR=OUTPUTDIR
 env_name="${EVNDISP_ML_ENV:-eventdisplay_ml}"
+HELPER_SCRIPTS_DIR="HHELPER_SCRIPTS_DIR"
+ENV_SNAPSHOT_DIR="EENV_SNAPSHOT_DIR"
 P="0.5"
 N="5000000"
 MAXCORES=NCORES
@@ -24,29 +26,10 @@ mkdir -p "$TEMPDIR"
 mkdir -p "${ODIR}"
 echo -e "Output files will be written to:\n ${ODIR}"
 
-check_conda_installation()
-{
-    if command -v conda &> /dev/null; then
-        echo "Found conda installation."
-    else
-        echo "Error: found no conda installation."
-        echo "exiting..."
-        exit
-    fi
-    env_info=$(conda info --envs)
-    if [[ "$env_info" == *"$env_name"* ]]; then
-        echo "Found conda environment '$env_name'"
-    else
-        echo "Error: the conda environment '$env_name' does not exist."
-        echo "exiting..."
-        exit
-    fi
-}
-
-check_conda_installation
-
-eval "$(conda shell.bash hook)"
-conda activate "$env_name"
+# shellcheck source=scripts/helper_scripts/UTILITY.conda_env.sh
+source "${HELPER_SCRIPTS_DIR}/UTILITY.conda_env.sh"
+evndisp_ml_setup_python_cache "$TEMPDIR" "train_stereo_$(basename "$LLIST" .list)"
+evndisp_ml_activate_conda "$env_name"
 
 PREFIX="${ODIR}/dispdir_bdt"
 LOGFILE="${PREFIX}.log"
@@ -60,7 +43,6 @@ eventdisplay-ml-train-xgb-stereo \
     --min_images 2 --memory_profile \
     --train_test_fraction $P --max_events $N >| "${LOGFILE}" 2>&1
 
-python --version >> "${LOGFILE}"
-conda list -n "$env_name" >> "${LOGFILE}"
+evndisp_ml_log_environment "${LOGFILE}" "$env_name" "$ENV_SNAPSHOT_DIR"
 
 conda deactivate
diff --git a/scripts/helper_scripts/IRF.trainXGBforGammaHadronSeparation_sub.sh b/scripts/helper_scripts/IRF.trainXGBforGammaHadronSeparation_sub.sh
@@ -11,6 +11,8 @@ PARA=MODELPARA
 EBIN=ENERGYBIN
 ODIR=OUTPUTDIR
 env_name="${EVNDISP_ML_ENV:-eventdisplay_ml}"
+HELPER_SCRIPTS_DIR="HHELPER_SCRIPTS_DIR"
+ENV_SNAPSHOT_DIR="EENV_SNAPSHOT_DIR"
 P="0.5"
 N="5000000"
 MAXCORES=NCORES
@@ -27,29 +29,10 @@ mkdir -p "$TEMPDIR"
 mkdir -p "${ODIR}"
 echo -e "Output files will be written to:\n ${ODIR}"
 
-check_conda_installation()
-{
-    if command -v conda &> /dev/null; then
-        echo "Found conda installation."
-    else
-        echo "Error: found no conda installation."
-        echo "exiting..."
-        exit
-    fi
-    env_info=$(conda info --envs)
-    if [[ "$env_info" == *"$env_name"* ]]; then
-        echo "Found conda environment '$env_name'"
-    else
-        echo "Error: the conda environment '$env_name' does not exist."
-        echo "exiting..."
-        exit
-    fi
-}
-
-check_conda_installation
-
-eval "$(conda shell.bash hook)"
-conda activate "$env_name"
+# shellcheck source=scripts/helper_scripts/UTILITY.conda_env.sh
+source "${HELPER_SCRIPTS_DIR}/UTILITY.conda_env.sh"
+evndisp_ml_setup_python_cache "$TEMPDIR" "train_gh_ebin${EBIN}"
+evndisp_ml_activate_conda "$env_name"
 
 PREFIX="${ODIR}/gammahadron_bdt"
 LOGFILE="${PREFIX}_ebin${EBIN}.log"
@@ -66,7 +49,6 @@ eventdisplay-ml-train-xgb-classify \
     --balance_class_zenith_weights \
     --train_test_fraction $P --max_events $N  >| "${LOGFILE}" 2>&1
 
-python --version >> "${LOGFILE}"
-conda list -n "$env_name" >> "${LOGFILE}"
+evndisp_ml_log_environment "${LOGFILE}" "$env_name" "$ENV_SNAPSHOT_DIR"
 
 conda deactivate
diff --git a/scripts/helper_scripts/UTILITY.conda_env.sh b/scripts/helper_scripts/UTILITY.conda_env.sh
@@ -0,0 +1,89 @@
+#!/bin/bash
+# Shared Conda setup and environment logging helpers for job scripts.
+
+evndisp_ml_setup_python_cache()
+{
+    local tempdir="$1"
+    local cache_tag="$2"
+
+    [[ -z "$tempdir" ]] && return
+
+    cache_tag="${cache_tag//\//_}"
+    export PYTHONPYCACHEPREFIX="${tempdir}/pycache_${cache_tag}"
+    mkdir -p "$PYTHONPYCACHEPREFIX"
+    echo "Python cache dir: $PYTHONPYCACHEPREFIX"
+}
+
+evndisp_ml_activate_conda()
+{
+    local env_name="${1:-${EVNDISP_ML_ENV:-eventdisplay_ml}}"
+
+    export CONDA_NO_PLUGINS="${CONDA_NO_PLUGINS:-true}"
+
+    if ! command -v conda > /dev/null 2>&1; then
+        echo "Error: found no conda installation."
+        echo "exiting..."
+        exit 1
+    fi
+
+    eval "$(conda shell.bash hook)"
+    if ! conda activate "$env_name"; then
+        echo "Error: failed to activate conda environment '$env_name'."
+        echo "exiting..."
+        exit 1
+    fi
+    echo "Activated conda environment '$env_name'"
+}
+
+evndisp_ml_log_environment()
+{
+    local logfile="$1"
+    local env_name="${2:-${EVNDISP_ML_ENV:-eventdisplay_ml}}"
+    local snapshot_dir="$3"
+    local snapshot_tag
+    local snapshot_file
+    local lock_dir
+    local tmp_file
+
+    {
+        echo "Python executable: $(command -v python)"
+        python --version
+    } >> "$logfile" 2>&1
+
+    [[ "${EVNDISP_ML_LOG_ENVIRONMENT:-snapshot}" == "none" ]] && return
+
+    if [[ -z "$snapshot_dir" ]]; then
+        python -m pip freeze >> "$logfile" 2>&1 \
+            || echo "pip freeze failed, continuing" >> "$logfile"
+        return
+    fi
+
+    mkdir -p "$snapshot_dir"
+    snapshot_tag="${env_name//\//_}"
+    snapshot_file="${snapshot_dir}/conda_environment_${snapshot_tag}.txt"
+    lock_dir="${snapshot_file}.lock"
+
+    if [[ -f "$snapshot_file" ]]; then
+        echo "Conda environment snapshot: $snapshot_file" >> "$logfile"
+        return
+    fi
+
+    if mkdir "$lock_dir" 2> /dev/null; then
+        tmp_file="${snapshot_file}.$$"
+        {
+            echo "Conda environment: $env_name"
+            echo "Python executable: $(command -v python)"
+            python --version
+            echo
+            echo "pip freeze:"
+            python -m pip freeze || echo "pip freeze failed"
+            echo
+            echo "conda list:"
+            conda list
+        } > "$tmp_file" 2>&1
+        mv "$tmp_file" "$snapshot_file"
+        rmdir "$lock_dir"
+    fi
+
+    echo "Conda environment snapshot: $snapshot_file" >> "$logfile"
+}
diff --git a/scripts/set_environment.sh b/scripts/set_environment.sh
@@ -59,7 +59,7 @@ export EVNDISPSCRIPTS
 ########################################################################
 # software settings
 export V2DL3SYS=${USERAFSDIR}/EVNDISP/EVNDISP-400/GITHUB_Eventdisplay/PreProcessing/V2DL3-v0.8.1
-export EVNDISP_MLSYS=${USERAFSDIR}/EVNDISP/EVNDISP-400/GITHUB_Eventdisplay/Eventdisplay-ML
+export EVNDISP_ML_SYS=${USERAFSDIR}/EVNDISP/EVNDISP-400/GITHUB_Eventdisplay/Eventdisplay-ML
 # EVENTDISPLAY using apptainers
 if [[ $PROCESS == "apptainer"* ]]; then
     export EVNDISP_APPTAINER="$VERITAS_DATA_DIR/shared/APPTAINERS/eventdisplay_v4_${EVNDISPVERSION}-rc5.sif"
Original file line number	Diff line number	Diff line change
		@@ -0,0 +1 @@
		Centralize XGB helper-script Conda activation and environment logging.