Make it installable on Windows

.gitignore

@@ -1,3 +1,55 @@
+<<<<<<< HEAD
+# These are some examples of commonly ignored file patterns.
+# You should customize this list as applicable to your project.
+# Learn more about .gitignore:
+#     https://www.atlassian.com/git/tutorials/saving-changes/gitignore
+
+# Node artifact files
+node_modules/
+dist/
+
+# Compiled Java class files
+*.class
+
+# Compiled Python bytecode
+*.py[cod]
+
+# Log files
+*.log
+
+# Package files
+*.jar
+
+# Maven
+target/
+dist/
+
+# JetBrains IDE
+.idea/
+
+# Unit test reports
+TEST*.xml
+
+# Generated by MacOS
+.DS_Store
+
+# Generated by Windows
+Thumbs.db
+
+# Applications
+*.app
+*.exe
+*.war
+
+# Large media files
+*.mp4
+*.tiff
+*.avi
+*.flv
+*.mov
+*.wmv
+
+=======
 # Temporary and binary files
 *~
 *.py[cod]
@@ -55,3 +107,4 @@ MANIFEST
 
 extern/rds2cpp*
 src/rds2py/lib/parser.cpp
+>>>>>>> bd-fix-windows

CMakeFiles/CMakeSystem.cmake

@@ -0,0 +1,15 @@
+set(CMAKE_HOST_SYSTEM "Windows")
+set(CMAKE_HOST_SYSTEM_NAME "Windows")
+set(CMAKE_HOST_SYSTEM_VERSION "")
+set(CMAKE_HOST_SYSTEM_PROCESSOR "AMD64")
+
+
+
+set(CMAKE_SYSTEM "Windows")
+set(CMAKE_SYSTEM_NAME "Windows")
+set(CMAKE_SYSTEM_VERSION "")
+set(CMAKE_SYSTEM_PROCESSOR "AMD64")
+
+set(CMAKE_CROSSCOMPILING "FALSE")
+
+set(CMAKE_SYSTEM_LOADED 1)

lib/CMakeLists.txt

@@ -3,11 +3,36 @@ cmake_minimum_required(VERSION 3.24)
 project(rds2py
     VERSION 1.0.0
     DESCRIPTION "Building the rds shared library"
-    LANGUAGES CXX)
+    LANGUAGES C CXX)
 
 # Importing all of the dependencies with pinned versions (even for transitive dependencies).
 include(FetchContent)
 
+# On Windows (or when zlib is not found), build zlib from source so that
+# byteme can use GzipFileReader / GzipFileWriter.
+find_package(ZLIB QUIET)
+if(NOT ZLIB_FOUND)
+    message(STATUS "System zlib not found -- building zlib from source via FetchContent")
+    FetchContent_Declare(
+      zlib
+      GIT_REPOSITORY https://github.com/madler/zlib
+      GIT_TAG v1.3.1
+    )
+    set(ZLIB_BUILD_EXAMPLES OFF CACHE BOOL "" FORCE)
+    FetchContent_MakeAvailable(zlib)
+
+    # Make the zlib headers visible to downstream targets and set the
+    # ZLIB_FOUND / ZLIB::ZLIB variables that byteme's CMakeLists.txt looks for.
+    add_library(ZLIB::ZLIB ALIAS zlibstatic)
+    target_include_directories(zlibstatic PUBLIC
+        "${zlib_SOURCE_DIR}"
+        "${zlib_BINARY_DIR}"
+    )
+    set(ZLIB_FOUND TRUE CACHE BOOL "" FORCE)
+    set(ZLIB_LIBRARIES zlibstatic CACHE STRING "" FORCE)
+    set(ZLIB_INCLUDE_DIRS "${zlib_SOURCE_DIR};${zlib_BINARY_DIR}" CACHE STRING "" FORCE)
+endif()
+
 FetchContent_Declare(
   rds2cpp
   GIT_REPOSITORY https://github.com/LTLA/rds2cpp
@@ -44,6 +69,11 @@ set_property(TARGET ${TARGET} PROPERTY CXX_STANDARD 17)
 
 target_link_libraries(${TARGET} PRIVATE rds2cpp pybind11::pybind11)
 
+# On Windows, also link zlib so the gzip symbols are available at link time.
+if(NOT ZLIB_FOUND OR TARGET zlibstatic)
+    target_link_libraries(${TARGET} PRIVATE ZLIB::ZLIB)
+endif()
+
 set_target_properties(${TARGET} PROPERTIES
     OUTPUT_NAME lib_rds_parser
     PREFIX ""

lib/src/rdswrapper.cpp

@@ -3,22 +3,24 @@
 #include <pybind11/numpy.h>
 #include <rds2cpp/rds2cpp.hpp>
 #include <stdexcept>
+#include <algorithm>
 #include <pybind11/iostream.h>
 
 namespace py = pybind11;
 
 class RdsReader {
 private:
     const rds2cpp::RObject* ptr;
+    const std::vector<rds2cpp::Symbol>* symbols;
 
 public:
-    RdsReader(const rds2cpp::RObject* p) : ptr(p) {
+    RdsReader(const rds2cpp::RObject* p, const std::vector<rds2cpp::Symbol>* syms) : ptr(p), symbols(syms) {
         if (!p) throw std::runtime_error("Null pointer passed to 'RdsReader'.");
+        if (!syms) throw std::runtime_error("Null symbols pointer passed to 'RdsReader'.");
     }
 
     std::string get_rtype() const {
         if (!ptr) throw std::runtime_error("Null pointer in 'get_rtype'.");
-        // py::print("arg::", static_cast<int>(ptr->type()));
         switch (ptr->type()) {
             case rds2cpp::SEXPType::S4: return "S4";
             case rds2cpp::SEXPType::INT: return "integer";
@@ -69,23 +71,36 @@ class RdsReader {
             throw std::runtime_error("Invalid type for 'string_arr'");
         }
         const auto& data = static_cast<const rds2cpp::StringVector*>(ptr)->data;
-        return py::cast(data);
+        py::list result;
+        for (const auto& s : data) {
+            if (s.value.has_value()) {
+                result.append(py::str(s.value.value()));
+            } else {
+                result.append(py::none());
+            }
+        }
+        return result;
     }
 
     py::list get_attribute_names() const {
         if (!ptr) throw std::runtime_error("Null pointer in 'get_attribute_names'");
-        return py::cast(get_attributes().names);
+        const auto& attrs = get_attributes();
+        py::list names;
+        for (const auto& attr : attrs) {
+            names.append((*symbols)[attr.name.index].name);
+        }
+        return names;
     }
 
     py::object load_attribute_by_name(const std::string& name) const {
         if (!ptr) throw std::runtime_error("Null pointer in 'load_attribute_by_name'");
-        const auto& attributes = get_attributes();
-        auto it = std::find(attributes.names.begin(), attributes.names.end(), name);
-        if (it == attributes.names.end()) {
-            throw std::runtime_error("Attribute not found: " + name);
+        const auto& attrs = get_attributes();
+        for (const auto& attr : attrs) {
+            if ((*symbols)[attr.name.index].name == name) {
+                return py::cast(new RdsReader(attr.value.get(), symbols));
+            }
         }
-        size_t index = std::distance(attributes.names.begin(), it);
-        return py::cast(new RdsReader(attributes.values[index].get()));
+        throw std::runtime_error("Attribute not found: " + name);
     }
 
     py::object load_vec_element(int index) const {
@@ -97,7 +112,7 @@ class RdsReader {
         if (index < 0 || static_cast<size_t>(index) >= data.size()) {
             throw std::out_of_range("Vector index out of range");
         }
-        return py::cast(new RdsReader(data[index].get()));
+        return py::cast(new RdsReader(data[index].get(), symbols));
     }
 
     std::string get_package_name() const {
@@ -126,7 +141,7 @@ class RdsReader {
     }
 
 private:
-    const rds2cpp::Attributes& get_attributes() const {
+    const std::vector<rds2cpp::Attribute>& get_attributes() const {
         if (!ptr) throw std::runtime_error("Null pointer in get_attributes");
         switch (ptr->type()) {
             case rds2cpp::SEXPType::INT: return static_cast<const rds2cpp::IntegerVector*>(ptr)->attributes;
@@ -142,18 +157,18 @@ class RdsReader {
 
 class RdsObject {
 private:
-    std::unique_ptr<rds2cpp::Parsed> parsed;
+    std::unique_ptr<rds2cpp::RdsFile> parsed;
     std::unique_ptr<RdsReader> reader;
 
 public:
     RdsObject(const std::string& file) {
         try {
             rds2cpp::ParseRdsOptions options;
-            parsed = std::make_unique<rds2cpp::Parsed>(rds2cpp::parse_rds(file, options));
+            parsed = std::make_unique<rds2cpp::RdsFile>(rds2cpp::parse_rds(file, options));
             if (!parsed || !parsed->object) {
                 throw std::runtime_error("Failed to parse RDS file");
             }
-            reader = std::make_unique<RdsReader>(parsed->object.get());
+            reader = std::make_unique<RdsReader>(parsed->object.get(), &parsed->symbols);
         } catch (const std::exception& e) {
             throw std::runtime_error(std::string("Error in 'RdsObject' constructor: ") + e.what());
         }
@@ -181,38 +196,31 @@ class RdaObject {
 
     py::list get_object_names() const {
         if (!parsed) throw std::runtime_error("Null parsed in 'get_object_names'");
-        const auto& pairlist = parsed->contents;
         py::list names;
-        for (size_t i = 0; i < pairlist.tag_names.size(); ++i) {
-            if (pairlist.has_tag[i]) {
-                names.append(pairlist.tag_names[i]);
-            } else {
-                names.append(py::none());
-            }
+        for (const auto& obj : parsed->objects) {
+            names.append(parsed->symbols[obj.name.index].name);
         }
         return names;
     }
 
     int get_object_count() const {
         if (!parsed) throw std::runtime_error("Null parsed in 'get_object_count'");
-        return static_cast<int>(parsed->contents.data.size());
+        return static_cast<int>(parsed->objects.size());
     }
 
     RdsReader* get_object_by_index(int index) const {
         if (!parsed) throw std::runtime_error("Null parsed in 'get_object_by_index'");
-        const auto& data = parsed->contents.data;
-        if (index < 0 || static_cast<size_t>(index) >= data.size()) {
+        if (index < 0 || static_cast<size_t>(index) >= parsed->objects.size()) {
             throw std::out_of_range("Object index out of range");
         }
-        return new RdsReader(data[index].get());
+        return new RdsReader(parsed->objects[index].value.get(), &parsed->symbols);
     }
 
     RdsReader* get_object_by_name(const std::string& name) const {
         if (!parsed) throw std::runtime_error("Null parsed in 'get_object_by_name'");
-        const auto& pairlist = parsed->contents;
-        for (size_t i = 0; i < pairlist.tag_names.size(); ++i) {
-            if (pairlist.has_tag[i] && pairlist.tag_names[i] == name) {
-                return new RdsReader(pairlist.data[i].get());
+        for (const auto& obj : parsed->objects) {
+            if (parsed->symbols[obj.name.index].name == name) {
+                return new RdsReader(obj.value.get(), &parsed->symbols);
             }
         }
         throw std::runtime_error("Object not found: " + name);
@@ -234,7 +242,6 @@ PYBIND11_MODULE(lib_rds_parser, m) {
         .def("get_object_by_name", &RdaObject::get_object_by_name, py::return_value_policy::take_ownership, py::keep_alive<0, 1>());
 
     py::class_<RdsReader>(m, "RdsReader")
-        .def(py::init<const rds2cpp::RObject*>())
         .def("get_rtype", &RdsReader::get_rtype)
         .def("get_rsize", &RdsReader::get_rsize)
         .def("get_numeric_data", &RdsReader::get_numeric_data)

setup.cfg

@@ -24,7 +24,7 @@ project_urls =
 #    Twitter = https://twitter.com/PyScaffold
 
 # Change if running only on Windows, Mac or Linux (comma-separated)
-platforms = Mac, Linux
+platforms = Mac, Linux, Windows
 
 # Add here all kinds of additional classifiers as defined under
 # https://pypi.org/classifiers/

setup.py

@@ -7,6 +7,7 @@
 
 from setuptools import setup, Extension
 from setuptools.command.build_ext import build_ext as build_ext_orig
+import glob
 import pathlib
 import os
 import shutil
@@ -33,23 +34,26 @@ def build_cmake(self, ext):
         outpath = os.path.join(build_lib.absolute(), ext.name)
 
         build_temp = os.path.join(build_temp, "build")
-        if not os.path.exists(build_temp):
-            cmd = [
-                "cmake",
-                "-S",
-                "lib",
-                "-B",
-                build_temp,
-                "-Dpybind11_DIR=" + os.path.join(os.path.dirname(pybind11.__file__), "share", "cmake", "pybind11"),
-                "-DPYTHON_EXECUTABLE=" + sys.executable,
-            ]
-            if os.name != "nt":
-                cmd.append("-DCMAKE_BUILD_TYPE=Release")
-                cmd.append("-DCMAKE_LIBRARY_OUTPUT_DIRECTORY=" + outpath)
+        # Remove stale build dir to avoid FetchContent conflicts with
+        # leftover _deps from a previous pip build environment.
+        if os.path.exists(build_temp):
+            shutil.rmtree(build_temp, ignore_errors=True)
+        cmd = [
+            "cmake",
+            "-S",
+            "lib",
+            "-B",
+            build_temp,
+            "-Dpybind11_DIR=" + os.path.join(os.path.dirname(pybind11.__file__), "share", "cmake", "pybind11"),
+            "-DPYTHON_EXECUTABLE=" + sys.executable,
+        ]
+        if os.name != "nt":
+            cmd.append("-DCMAKE_BUILD_TYPE=Release")
+            cmd.append("-DCMAKE_LIBRARY_OUTPUT_DIRECTORY=" + outpath)
 
-            if "MORE_CMAKE_OPTIONS" in os.environ:
-                cmd += os.environ["MORE_CMAKE_OPTIONS"].split()
-            self.spawn(cmd)
+        if "MORE_CMAKE_OPTIONS" in os.environ:
+            cmd += os.environ["MORE_CMAKE_OPTIONS"].split()
+        self.spawn(cmd)
 
         if not self.dry_run:
             cmd = ["cmake", "--build", build_temp]
@@ -59,10 +63,20 @@ def build_cmake(self, ext):
             if os.name == "nt":
                 # Gave up trying to get MSVC to respect the output directory.
                 # Delvewheel also needs it to have a 'pyd' suffix... whatever.
-                shutil.copyfile(
-                    os.path.join(build_temp, "Release", "_core.dll"),
-                    os.path.join(outpath, "_core.pyd"),
-                )
+                # The CMake target name is lib_rds_parser; MSVC puts it under Release/.
+                # pybind11 may add an ABI tag (e.g. lib_rds_parser.cp312-win_amd64.pyd).
+                release_dir = os.path.join(build_temp, "Release")
+                candidates = glob.glob(os.path.join(release_dir, "lib_rds_parser*"))
+                if not candidates:
+                    raise RuntimeError(
+                        f"Cannot find compiled library in {release_dir}. "
+                        f"Contents: {os.listdir(release_dir) if os.path.isdir(release_dir) else 'dir not found'}"
+                    )
+                # Prefer .pyd files over .lib/.exp
+                pyd_files = [c for c in candidates if c.endswith(".pyd")]
+                src_name = pyd_files[0] if pyd_files else candidates[0]
+                os.makedirs(outpath, exist_ok=True)
+                shutil.copyfile(src_name, os.path.join(outpath, "lib_rds_parser.pyd"))
 
 
 if __name__ == "__main__":