pyodide · stefdoerr · Mar 18, 2026 · Mar 18, 2026 · Mar 18, 2026 · Mar 18, 2026
diff --git a/packages/libboost-python/meta.yaml b/packages/libboost-python/meta.yaml
@@ -0,0 +1,59 @@
+package:
+  name: libboost-python
+  version: 1.84.0
+  tag:
+    - library
+    - static_library
+
+source:
+  url: https://github.com/boostorg/boost/releases/download/boost-1.84.0/boost-1.84.0.tar.gz
+  sha256: 4d27e9efed0f6f152dc28db6430b9d3dfb40c0345da7342eaa5a987dde57bd95
+
+requirements:
+  host:
+    - libzlib
+
+build:
+  type: static_library
+  script: |
+    # Patch boost.numpy for NumPy 2.x (PyArray_Descr->elsize removed)
+    sed -i 's/reinterpret_cast<PyArray_Descr\*>(ptr())->elsize/PyDataType_ELSIZE(reinterpret_cast<PyArray_Descr*>(ptr()))/' \
+        libs/python/src/numpy/dtype.cpp
+
+    ./bootstrap.sh --prefix=${WASM_LIBRARY_DIR} \
+        --with-libraries=system,python,serialization,iostreams \
+        --with-python=python${PYMAJOR}.${PYMINOR}
+
+    # https://github.com/emscripten-core/emscripten/issues/17052
+    # Without this, boost outputs WASM modules not static library archives.
+    printf "using clang : emscripten : emcc : <archiver>emar <ranlib>emranlib <linker>emlink ;\n" \
+        | tee -a ./project-config.jam
+
+    # Bypass standard python detection, specify cross-compilation paths
+    sed -i 's/using python/#using python/' ./project-config.jam
+    NUMPY_INC=$(python3 -c "import numpy; print(numpy.get_include())")
+    printf "using python : ${PYMAJOR}.${PYMINOR} : python${PYMAJOR}.${PYMINOR} : ${PYTHONINCLUDE} ${NUMPY_INC} ;\n" \
+        | tee -a ./project-config.jam
+
+    ./b2 variant=release toolset=clang-emscripten link=static threading=single \
+        address-model=32 --disable-icu \
+        cxxflags="$SIDE_MODULE_CXXFLAGS -fwasm-exceptions -std=c++20 -DBOOST_SP_DISABLE_THREADS=1" \
+        cflags="$SIDE_MODULE_CFLAGS -fwasm-exceptions -DBOOST_SP_DISABLE_THREADS=1" \
+        linkflags="-fpic $SIDE_MODULE_LDFLAGS" \
+        python=${PYMAJOR}.${PYMINOR} \
+        -sZLIB_INCLUDE=${WASM_LIBRARY_DIR}/include \
+        -sZLIB_LIBPATH=${WASM_LIBRARY_DIR}/lib \
+        --layout=system -j"${PYODIDE_JOBS:-3}" --prefix=${WASM_LIBRARY_DIR} \
+        install || true
+
+    # Verify key libraries were built
+    echo "=== Boost libraries ==="
+    ls -la ${WASM_LIBRARY_DIR}/lib/libboost_python*.a
+    ls -la ${WASM_LIBRARY_DIR}/lib/libboost_numpy*.a
+    ls -la ${WASM_LIBRARY_DIR}/lib/libboost_serialization*.a
+    ls -la ${WASM_LIBRARY_DIR}/lib/libboost_iostreams*.a
+
+about:
+  home: https://www.boost.org/
+  summary: Boost C++ libraries with Python and NumPy support for Emscripten
+  license: Boost
diff --git a/packages/librdkit/meta.yaml b/packages/librdkit/meta.yaml
@@ -0,0 +1,158 @@
+package:
+  name: librdkit
+  version: 2025.9.6
+  tag:
+    - library
+    - shared_library
+
+source:
+  url: https://github.com/rdkit/rdkit/archive/refs/tags/Release_2025_09_6.tar.gz
+  sha256: 57b92e8f47d9dbd559bd808d5cf6c48a628bc36118bc35b832a35e2ca8a0c7a1
+  extract_dir: rdkit-Release_2025_09_6
+
+requirements:
+  host:
+    - libboost-python
+    - libzlib
+
+build:
+  type: shared_library
+  script: |
+    set -e
+
+    # ========================
+    # 1. Setup — derive Python paths from pyodide config
+    # ========================
+    PYINC=$(pyodide config get python_include_dir)
+    PYVER=$(pyodide config get python_version | cut -d. -f1-2)
+    PYMAJ=$(echo $PYVER | cut -d. -f1)
+    PYMIN=$(echo $PYVER | cut -d. -f2)
+    NUMPY_INC=$(python3 -c "import numpy; print(numpy.get_include())")
+
+    # Download Eigen3 headers (header-only)
+    if [ ! -d "eigen-3.4.0" ]; then
+        wget -q https://gitlab.com/libeigen/eigen/-/archive/3.4.0/eigen-3.4.0.tar.gz
+        tar xf eigen-3.4.0.tar.gz
+    fi
+
+    # Dummy libpython — side modules don't link against it, but CMake's FindPython3 needs it
+    emar rcs /tmp/libpython${PYVER}.a
+
+    # Patch CMakeLists.txt for Emscripten support
+    sed -i 's/OR RDK_BUILD_MINIMAL_LIB)/OR EMSCRIPTEN OR RDK_BUILD_MINIMAL_LIB)/' CMakeLists.txt
+
+    # ========================
+    # 2. CMake configure
+    # ========================
+    mkdir -p build && cd build
+
+    emcmake cmake \
+        -DCMAKE_BUILD_TYPE=Release \
+        -DRDK_BUILD_PYTHON_WRAPPERS=ON \
+        -DRDK_BUILD_MINIMAL_LIB=OFF \
+        -DRDK_BUILD_CPP_TESTS=OFF \
+        -DRDK_BUILD_INCHI_SUPPORT=ON \
+        -DRDK_USE_BOOST_SERIALIZATION=ON \
+        -DRDK_USE_BOOST_IOSTREAMS=ON \
+        -DRDK_OPTIMIZE_POPCNT=OFF \
+        -DRDK_BUILD_THREADSAFE_SSS=OFF \
+        -DRDK_BUILD_DESCRIPTORS3D=ON \
+        -DRDK_TEST_MULTITHREADED=OFF \
+        -DRDK_BUILD_CHEMDRAW_SUPPORT=ON \
+        -DRDK_BUILD_MAEPARSER_SUPPORT=ON \
+        -DRDK_BUILD_COORDGEN_SUPPORT=ON \
+        -DRDK_BUILD_SLN_SUPPORT=ON \
+        -DRDK_BUILD_CAIRO_SUPPORT=OFF \
+        -DRDK_BUILD_QT_SUPPORT=OFF \
+        -DRDK_BUILD_PGSQL=OFF \
+        -DRDK_BUILD_SWIG_WRAPPERS=OFF \
+        -DRDK_BUILD_FREETYPE_SUPPORT=OFF \
+        -DRDK_BUILD_AVALON_SUPPORT=ON \
+        -DRDK_BUILD_YAEHMOP_SUPPORT=OFF \
+        -DRDK_BUILD_XYZ2MOL_SUPPORT=ON \
+        -DRDK_INSTALL_INTREE=ON \
+        -DRDK_INSTALL_STATIC_LIBS=ON \
+        -DRDK_BUILD_STATIC_LIBS_ONLY=ON \
+        -DBoost_USE_STATIC_LIBS=ON \
+        -DCMAKE_PREFIX_PATH="${WASM_LIBRARY_DIR}" \
+        -DCMAKE_FIND_ROOT_PATH="${WASM_LIBRARY_DIR}" \
+        -DCMAKE_FIND_ROOT_PATH_MODE_PACKAGE=BOTH \
+        -DBoost_INCLUDE_DIR=${WASM_LIBRARY_DIR}/include \
+        -DCMAKE_INCLUDE_PATH="${WASM_LIBRARY_DIR}/include" \
+        -DCMAKE_LIBRARY_PATH="${WASM_LIBRARY_DIR}/lib" \
+        -DZLIB_INCLUDE_DIR=${WASM_LIBRARY_DIR}/include \
+        -DZLIB_LIBRARY=${WASM_LIBRARY_DIR}/lib/libz.a \
+        -DEIGEN3_INCLUDE_DIR=$PWD/../eigen-3.4.0 \
+        -DPython3_INCLUDE_DIR=${PYINC} \
+        -DPython3_LIBRARY=/tmp/libpython${PYVER}.a \
+        -DPython3_NumPy_INCLUDE_DIR=${NUMPY_INC} \
+        -DPython3_EXECUTABLE=$(which python${PYVER}) \
+        -DCMAKE_CXX_FLAGS="-fwasm-exceptions -O2 -DNDEBUG -fPIC -std=c++20" \
+        -DCMAKE_C_FLAGS="-fwasm-exceptions -O2 -DNDEBUG -fPIC" \
+        ..
+
+    # ========================
+    # 3. Build
+    # ========================
+    emmake make -k -j ${PYODIDE_JOBS:-3} || true
+    cd ..
+
+    # ========================
+    # 4. Relink core into a single WASM side module
+    # ========================
+    find build/lib build/External -name '*.a' 2>/dev/null | sort > /tmp/all_libs.txt
+    for lib in ${WASM_LIBRARY_DIR}/lib/libboost_python*.a \
+               ${WASM_LIBRARY_DIR}/lib/libboost_numpy*.a \
+               ${WASM_LIBRARY_DIR}/lib/libboost_system*.a \
+               ${WASM_LIBRARY_DIR}/lib/libboost_serialization*.a \
+               ${WASM_LIBRARY_DIR}/lib/libboost_iostreams*.a; do
+        [ -f "$lib" ] && echo "$lib" >> /tmp/all_libs.txt
+    done
+    echo "${WASM_LIBRARY_DIR}/lib/libz.a" >> /tmp/all_libs.txt
+    echo "Core libraries: $(wc -l < /tmp/all_libs.txt)"
+
+    WHOLE_ARGS=""
+    while read -r lib; do
+        WHOLE_ARGS="$WHOLE_ARGS -Wl,--whole-archive $lib -Wl,--no-whole-archive"
+    done < /tmp/all_libs.txt
+
+    em++ -fwasm-exceptions -sSIDE_MODULE=1 -O2 -shared \
+        -Wl,--no-gc-sections -Wl,--export-all \
+        $WHOLE_ARGS \
+        -o build/librdkit_core.so
+    ls -lh build/librdkit_core.so
+
+    # Copy core .so to $DISTDIR (Pyodide loads shared_library packages
+    # asynchronously, avoiding Chrome's 8MB sync WebAssembly.Compile limit)
+    cp build/librdkit_core.so ${DISTDIR}
+
+    # Also install to $WASM_LIBRARY_DIR/lib/ so the rdkit recipe can link
+    # wrapper .so modules against it (proper dynamic linking, no RTLD_GLOBAL)
+    cp build/librdkit_core.so ${WASM_LIBRARY_DIR}/lib/
+
+    # ========================
+    # 5. Stage build artifacts for the rdkit Python package recipe
+    # ========================
+    STAGING=${WASM_LIBRARY_DIR}/share/rdkit
+
+    # Stage wrapper .o files and cmake build.make (for output path derivation)
+    mkdir -p ${STAGING}/wrappers
+    for wrap_dir in $(find build -path '*/Wrap/CMakeFiles/*.dir' -type d 2>/dev/null); do
+        modname=$(basename "$wrap_dir" .dir)
+        obj_files=$(find "$wrap_dir" -name '*.o' 2>/dev/null)
+        if [ -z "$obj_files" ]; then continue; fi
+        mkdir -p "${STAGING}/wrappers/${modname}"
+        cp $obj_files "${STAGING}/wrappers/${modname}/"
+        cp "$wrap_dir/build.make" "${STAGING}/wrappers/${modname}/"
+    done
+
+    # Stage Python source files and Data directory
+    cp -r rdkit ${STAGING}/python
+    cp -r Data ${STAGING}/Data
+
+    echo "Staged $(ls ${STAGING}/wrappers | wc -l) wrapper modules"
+
+about:
+  home: https://github.com/rdkit/rdkit
+  summary: RDKit core C++ libraries as a WASM side module
+  license: BSD-3-Clause
diff --git a/packages/rdkit/extras/patch_init.py b/packages/rdkit/extras/patch_init.py
@@ -0,0 +1,50 @@
+"""Patch rdkit/__init__.py for emscripten/Pyodide support.
+
+The librdkit_core.so shared library is loaded by Pyodide via the librdkit
+shared_library package (asynchronously during loadPackage). Each wrapper
+.so.wasm module links against librdkit_core.so directly (like scipy links
+against libopenblas), so no RTLD_GLOBAL hack is needed.
+
+This patch:
+1. Sets RDBASE for RDConfig.py path resolution
+2. Registers a custom MetaPathFinder to load .so.wasm wrapper modules
+"""
+
+init_path = "rdkit/__init__.py"
+init = open(init_path).read()
+
+loader = '''import sys as _sys
+
+if _sys.platform == 'emscripten':
+    import os as _os
+    import importlib.abc
+    import importlib.machinery
+
+    # Set RDBASE so RDConfig.py finds Data/, Docs/, etc. relative to this package
+    _os.environ['RDBASE'] = _os.path.dirname(__file__)
+
+    class _RDKitExtensionFinder(importlib.abc.MetaPathFinder):
+        def find_spec(self, fullname, path, target=None):
+            parts = fullname.split('.')
+            if parts[0] != 'rdkit':
+                return None
+            modname = parts[-1]
+            if path:
+                for d in path:
+                    candidate = _os.path.join(d, modname + '.so.wasm')
+                    if _os.path.exists(candidate):
+                        loader = importlib.machinery.ExtensionFileLoader(
+                            fullname, candidate
+                        )
+                        return importlib.util.spec_from_file_location(
+                            fullname, candidate, loader=loader,
+                        )
+            return None
+
+    import importlib.util
+    _sys.meta_path.insert(0, _RDKitExtensionFinder())
+
+'''
+
+open(init_path, "w").write(loader + init)
+print("Patched rdkit/__init__.py")
diff --git a/packages/rdkit/extras/pyproject.toml b/packages/rdkit/extras/pyproject.toml
@@ -0,0 +1,15 @@
+[build-system]
+requires = ["setuptools>=68.0"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "rdkit"
+version = "0.0.0"
+description = "RDKit cheminformatics library for Pyodide"
+
+[tool.setuptools.packages.find]
+include = ["rdkit*"]
+
+[tool.setuptools.package-data]
+"*" = ["*.so.wasm"]
+rdkit = ["Data/**/*"]
diff --git a/packages/rdkit/meta.yaml b/packages/rdkit/meta.yaml
@@ -0,0 +1,91 @@
+package:
+  name: rdkit
+  version: 2025.9.6
+  top-level:
+    - rdkit
+
+source:
+  url: https://github.com/rdkit/rdkit/archive/refs/tags/Release_2025_09_6.tar.gz
+  sha256: 57b92e8f47d9dbd559bd808d5cf6c48a628bc36118bc35b832a35e2ca8a0c7a1
+  extract_dir: rdkit-Release_2025_09_6
+
+  extras:
+    - [extras/pyproject.toml, pyproject.toml]
+    - [extras/patch_init.py, patch_init.py]
+
+requirements:
+  host:
+    - librdkit
+    - numpy
+  run:
+    - librdkit
+    - numpy
+
+build:
+  script: |
+    set -e
+
+    STAGING=${WASM_LIBRARY_DIR}/share/rdkit
+
+    # Set version in pyproject.toml
+    sed -i "s/^version = .*/version = \"${PKG_VERSION}\"/" pyproject.toml
+
+    # ========================
+    # 1. Relink wrappers from staged .o files
+    # ========================
+    WRAPPER_COUNT=0
+    for wrapper_dir in ${STAGING}/wrappers/*/; do
+        modname=$(basename "$wrapper_dir")
+        obj_files=$(find "$wrapper_dir" -name '*.o' 2>/dev/null)
+        if [ -z "$obj_files" ]; then
+            echo "SKIP: $modname (no .o files)"
+            continue
+        fi
+        # Derive the correct output path from cmake's build.make
+        dest_path=$(grep -oP "rdkit/[^ ]*${modname}\.so" "$wrapper_dir/build.make" 2>/dev/null | head -1)
+        if [ -z "$dest_path" ]; then
+            echo "SKIP: $modname (no output path found)"
+            continue
+        fi
+        mkdir -p "$(dirname "$dest_path")"
+        em++ -fwasm-exceptions -sSIDE_MODULE=1 -O2 -shared \
+            $obj_files \
+            -L${WASM_LIBRARY_DIR}/lib -lrdkit_core \
+            -o "$dest_path"
+        echo "LINKED: $modname -> $dest_path"
+        WRAPPER_COUNT=$((WRAPPER_COUNT + 1))
+    done
+    echo "Total wrapper .so files: $WRAPPER_COUNT"
+
+    # ========================
+    # 2. Assemble Python package
+    # ========================
+    # Overlay cmake-generated Python files (e.g. inchi.py) from staging
+    cp -rn ${STAGING}/python/* rdkit/ 2>/dev/null || true
+
+    # Copy Data directory into package
+    cp -r ${STAGING}/Data rdkit/Data
+
+    # Rename .so -> .so.wasm (prevents micropip from auto-loading before
+    # core is ready — the core is loaded by Pyodide via the librdkit
+    # shared_library package with {global: true})
+    find rdkit -name "*.so" -exec sh -c 'mv "$1" "$1.wasm"' _ {} \;
+
+    # Patch __init__.py to set RDBASE and register custom import finder
+    python3 patch_init.py
+
+    echo "=== Package contents ==="
+    find rdkit -name "*.so.wasm" | sort
+    echo "---"
+    du -sh rdkit
+
+test:
+  imports:
+    - rdkit
+    - rdkit.Chem
+
+about:
+  home: https://github.com/rdkit/rdkit
+  PyPI: https://pypi.org/project/rdkit
+  summary: RDKit cheminformatics library for Pyodide
+  license: BSD-3-Clause