feat(build): clean up build artifacts and optimize wheel installation process in Linux script

Author: singjc

scripts/build/build_linux.sh

@@ -32,6 +32,18 @@ echo "Installing build tools..."
 sudo apt-get update -qq
 sudo apt-get install -y binutils
 
+# Save the original directory
+ORIGINAL_DIR="$(pwd)"
+
+# Clean ALL build artifacts that might confuse PyInstaller
+echo "Cleaning build artifacts..."
+rm -rf build dist pyprophet.spec *.egg-info
+rm -rf .eggs
+# Clean any Python cache that might have source references
+find . -type d -name __pycache__ -exec rm -rf {} + 2>/dev/null || true
+find . -type f -name "*.pyc" -delete 2>/dev/null || true
+find . -type f -name "*.pyo" -delete 2>/dev/null || true
+
 # Install/upgrade build dependencies
 $PYTHON -m pip install --upgrade pip setuptools wheel cython numpy pyinstaller
 
@@ -59,9 +71,20 @@ for dep in deps:
             continue
 PYEOF
 
-# Build and install pyprophet as a wheel (cleanest approach)
-echo "Building and installing pyprophet package..."
+# Build pyprophet wheel in a completely isolated temp directory
+echo "Building pyprophet wheel in isolated directory..."
+WHEEL_BUILD_DIR=$(mktemp -d)
+cp -r "${ORIGINAL_DIR}"/* "${WHEEL_BUILD_DIR}/" 2>/dev/null || true
+cd "${WHEEL_BUILD_DIR}"
+
 $PYTHON -m pip wheel --no-deps --wheel-dir /tmp/pyprophet_wheels .
+
+# Return to original directory and clean up wheel build directory
+cd "${ORIGINAL_DIR}"
+rm -rf "${WHEEL_BUILD_DIR}"
+
+# Install pyprophet from wheel
+echo "Installing pyprophet from wheel..."
 $PYTHON -m pip install --force-reinstall --no-deps /tmp/pyprophet_wheels/pyprophet-*.whl
 
 # Verify installation
@@ -72,6 +95,10 @@ $PYTHON -c "import pyprophet; print(f'PyProphet installed at: {pyprophet.__file_
 SITE_PACKAGES=$($PYTHON -c "import pyprophet, os; print(os.path.dirname(pyprophet.__file__))")
 echo "PyProphet package location: ${SITE_PACKAGES}"
 
+# Verify numpy is installed correctly (not from source)
+echo "Verifying numpy installation..."
+$PYTHON -c "import numpy; print(f'NumPy installed at: {numpy.__file__}'); import pandas; print('Pandas imports successfully')"
+
 # Collect compiled extension binaries from the installed package
 ADD_BINARY_ARGS=()
 for so in "${SITE_PACKAGES}"/pyprophet/scoring/_optimized*.so; do
@@ -81,27 +108,29 @@ for so in "${SITE_PACKAGES}"/pyprophet/scoring/_optimized*.so; do
   fi
 done
 
-# Save the original directory
-ORIGINAL_DIR="$(pwd)"
-
-# Clean previous PyInstaller builds in original directory
-rm -rf build dist pyprophet.spec
-
 # Create dist directory in original location
 mkdir -p "${ORIGINAL_DIR}/dist"
 
-# Change to a temporary directory to avoid picking up source files
-# This ensures PyInstaller only sees installed packages
+# Change to a temporary directory to avoid picking up ANY source files
 BUILD_DIR=$(mktemp -d)
 echo "Using temporary build directory: ${BUILD_DIR}"
 cd "${BUILD_DIR}"
 
-# Copy only the necessary files
+# Copy only the necessary files (NOT the entire source tree)
 cp "${ORIGINAL_DIR}/packaging/pyinstaller/run_pyprophet.py" .
-cp -r "${ORIGINAL_DIR}/packaging/pyinstaller/hooks" .
+mkdir -p hooks
+cp "${ORIGINAL_DIR}/packaging/pyinstaller/hooks"/*.py hooks/ 2>/dev/null || true
 
-# Run PyInstaller in onefile mode (single executable)
+# Verify we're in a clean directory with no source pollution
+echo "Build directory contents:"
+ls -la
+
+# Run PyInstaller in onefile mode
 echo "Running PyInstaller (onefile mode)..."
+echo "Current directory: $(pwd)"
+echo "Python sys.path will be:"
+$PYTHON -c "import sys; import pprint; pprint.pprint(sys.path)"
+
 $PYTHON -m PyInstaller \
   --clean \
   --noconfirm \
@@ -125,15 +154,15 @@ $PYTHON -m PyInstaller \
   --exclude-module black \
   --exclude-module ruff \
   --collect-submodules pyprophet \
-  --collect-all numpy \
-  --collect-all pandas \
-  --collect-all scipy \
-  --collect-all sklearn \
-  --collect-all pyopenms \
+  --copy-metadata numpy \
+  --copy-metadata pandas \
+  --copy-metadata scipy \
+  --copy-metadata sklearn \
+  --copy-metadata scikit-learn \
+  --copy-metadata pyopenms \
   --copy-metadata duckdb \
   --copy-metadata duckdb-extensions \
   --copy-metadata duckdb-extension-sqlite-scanner \
-  --copy-metadata pyopenms \
   "${ADD_BINARY_ARGS[@]}" \
   run_pyprophet.py
 
@@ -144,7 +173,7 @@ mv dist/pyprophet "${ORIGINAL_DIR}/dist/pyprophet"
 # Return to original directory
 cd "${ORIGINAL_DIR}"
 
-# Clean up temporary build directory and wheel directory
+# Clean up temporary directories
 rm -rf "${BUILD_DIR}" /tmp/pyprophet_wheels
 
 # NOTE: UPX compression is NOT applied on Linux because it breaks PyInstaller executables