spark-class

#!/usr/bin/env bash

# Debug: Print environment variables
echo "JAVA_HOME: $JAVA_HOME"
echo "PATH: $PATH"

# Hard-code the Java executable path for maximum reliability
JAVA_EXECUTABLE="/usr/bin/java"

# Fallback options if the hard-coded path doesn't work
if [ ! -f "$JAVA_EXECUTABLE" ]; then
  echo "Primary Java path not found, trying alternatives..."
  
  # Try using JAVA_HOME
  if [ -n "$JAVA_HOME" ] && [ -f "$JAVA_HOME/bin/java" ]; then
    JAVA_EXECUTABLE="$JAVA_HOME/bin/java"
    echo "Using Java from JAVA_HOME: $JAVA_EXECUTABLE"
  # Try Java 11 Amazon Corretto path
  elif [ -f "/usr/lib/jvm/java-11-amazon-corretto/bin/java" ]; then
    JAVA_EXECUTABLE="/usr/lib/jvm/java-11-amazon-corretto/bin/java"
    echo "Using Java from Amazon Corretto path: $JAVA_EXECUTABLE"
  # Try common JRE path (for compatibility)
  elif [ -f "/usr/lib/jvm/jre/bin/java" ]; then
    JAVA_EXECUTABLE="/usr/lib/jvm/jre/bin/java"
    echo "Using Java from JRE path: $JAVA_EXECUTABLE"
  else
    # Last resort: search for any java executable
    echo "Searching for Java executable..."
    java_path=$(find /usr -name java -type f -executable 2>/dev/null | head -1)
    if [ -n "$java_path" ]; then
      JAVA_EXECUTABLE="$java_path"
      echo "Found Java at: $java_path"
    else
      echo "Error: Java executable not found" >&2
      exit 1
    fi
  fi
fi

# Verify Java executable exists and is executable
if [ ! -f "$JAVA_EXECUTABLE" ] || [ ! -x "$JAVA_EXECUTABLE" ]; then
  echo "Error: Java executable not found or not executable at $JAVA_EXECUTABLE" >&2
  exit 1
fi

# Print Java version for debugging
echo "Using Java executable: $JAVA_EXECUTABLE"
$JAVA_EXECUTABLE -version

# Use Python 3.12 paths for PySpark
exec "$JAVA_EXECUTABLE" -cp /var/lang/lib/python3.12/site-packages/pyspark/conf/:/var/lang/lib/python3.12/site-packages/pyspark/jars/* -Xmx1g "$@"