py-chunks 0.2.1__tar.gz → 0.2.3__tar.gz

{py_chunks-0.2.1 → py_chunks-0.2.3}/.github/workflows/release.yml

@@ -131,7 +131,7 @@ jobs:
         with:
           python-version: "3.13"
 
-      - name: Download and bundle PDFium
+      - name: Download and bundle PDFium + MSVC runtime
         shell: pwsh
         run: |
           Invoke-WebRequest `
@@ -144,6 +144,17 @@ jobs:
           Copy-Item -Path $dll.FullName -Destination "py_chunks\pdfium.dll"
           Write-Host "Bundled $($dll.Name) ($('{0:N0}' -f (Get-Item py_chunks\pdfium.dll).Length) bytes)"
 
+          # Bundle MSVC runtime DLLs alongside pdfium.dll so the Rust
+          # LOAD_WITH_ALTERED_SEARCH_PATH preload finds them in py_chunks/.
+          $runtimeDlls = @("vcruntime140.dll", "msvcp140.dll", "vcruntime140_1.dll")
+          foreach ($rt in $runtimeDlls) {
+            $src = "C:\Windows\System32\$rt"
+            if (Test-Path $src) {
+              Copy-Item -Path $src -Destination "py_chunks\$rt"
+              Write-Host "Bundled runtime $rt"
+            }
+          }
+
       - name: Build wheel
         uses: PyO3/maturin-action@v1
         with:

{py_chunks-0.2.1 → py_chunks-0.2.3}/.gitignore

@@ -2,6 +2,9 @@
 py_chunks/libpdfium.dylib
 py_chunks/libpdfium.so
 py_chunks/pdfium.dll
+py_chunks/vcruntime140.dll
+py_chunks/msvcp140.dll
+py_chunks/vcruntime140_1.dll
 
 # Python artifacts
 __pycache__/

{py_chunks-0.2.1 → py_chunks-0.2.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: py-chunks
-Version: 0.2.1
+Version: 0.2.3
 Classifier: Programming Language :: Python :: 3
 Classifier: License :: OSI Approved :: MIT License
 Classifier: Operating System :: OS Independent

{py_chunks-0.2.1 → py_chunks-0.2.3}/py_chunks/__init__.py

@@ -6,6 +6,7 @@ URLs.
 """
 
 import os
+import sys
 import tempfile
 from os import PathLike, fspath
 from pathlib import Path
@@ -13,11 +14,30 @@ from typing import Any
 from urllib.parse import urlparse
 from urllib.request import urlopen
 
-# Register the package directory so the Rust layer can find a bundled
-# libpdfium binary (libpdfium.dylib / pdfium.dll / libpdfium.so) placed
-# alongside _rust.so inside this directory.  Must be set before any Rust
-# symbol that calls get_pdfium() is first invoked.
-os.environ.setdefault("PY_CHUNKS_PACKAGE_DIR", str(Path(__file__).parent))
+_pkg_dir = Path(__file__).parent
+
+# Tell the Rust layer where to find the bundled PDFium binary.
+os.environ.setdefault("PY_CHUNKS_PACKAGE_DIR", str(_pkg_dir))
+
+# Directly resolve the bundled binary and set PDFIUM_LIBRARY_PATH to its
+# absolute path.  This hits the highest-priority branch in the Rust resolver
+# so no directory scanning is needed — the path is always exact.
+_PDFIUM_NAMES = {
+    "win32":  "pdfium.dll",
+    "darwin": "libpdfium.dylib",
+    "linux":  "libpdfium.so",
+}
+_pdfium_bin = _pkg_dir / _PDFIUM_NAMES.get(sys.platform, "")
+if _pdfium_bin.exists():
+    os.environ.setdefault("PDFIUM_LIBRARY_PATH", str(_pdfium_bin))
+
+# On Windows, register the package directory as a DLL search directory so
+# pdfium.dll's own dependencies (vcruntime140.dll, msvcp140.dll …) are found
+# in py_chunks/ rather than failing with LoadLibrary error 126.
+# os.add_dll_directory() wraps AddDllDirectory() — available on Python 3.8+,
+# which we always satisfy (requires-python = ">=3.9").
+if sys.platform == "win32" and hasattr(os, "add_dll_directory"):
+    os.add_dll_directory(str(_pkg_dir))
 
 from .chunkers.docx import chunk_docx, stream_chunk_docx
 from .chunkers.html import chunk_html, stream_chunk_html

{py_chunks-0.2.1 → py_chunks-0.2.3}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "maturin"
 
 [project]
 name = "py-chunks"
-version = "0.2.1"
+version = "0.2.3"
 description = "Rust-backed Python chunking library"
 requires-python = ">=3.9"
 dependencies = []

{py_chunks-0.2.1 → py_chunks-0.2.3}/src/extensions/pdf/structural.rs

@@ -434,6 +434,14 @@ pub(super) fn get_pdfium() -> Result<Pdfium, String> {
             add_library_search_path(parent);
         }
 
+        // On Windows, pre-load with LOAD_WITH_ALTERED_SEARCH_PATH so the
+        // loader searches pdfium.dll's own directory for its dependencies
+        // (e.g. bundled vcruntime140.dll).  If successful the module is
+        // cached; pdfium-render's subsequent LoadLibraryExW returns the
+        // already-loaded handle without a new search.
+        #[cfg(windows)]
+        windows_preload_pdfium(&path);
+
         if let Ok(bindings) = Pdfium::bind_to_library(&path) {
             return Ok(Pdfium::new(bindings));
         }
@@ -443,12 +451,50 @@ pub(super) fn get_pdfium() -> Result<Pdfium, String> {
         .map(|b| Pdfium::new(b))
         .map_err(|e| {
             format!(
-                "PDFium native library not found or could not be loaded. The py-chunks wheel expects pypdfium2 to supply a native PDFium binary automatically. If this still fails, verify that the installed pypdfium2 package is intact and that your environment has access to the bundled pdfium library. On Windows, ensure pdfium.dll is present on PATH. Loader error: {}",
+                "PDFium native library not found or could not be loaded. \
+                 The py-chunks wheel bundles pdfium.dll — if you see this on \
+                 Windows, install the Visual C++ Redistributable 2022 (x64): \
+                 https://aka.ms/vs/17/release/vc_redist.x64.exe \
+                 Or set PDFIUM_LIBRARY_PATH to point at a working pdfium.dll. \
+                 Loader error: {}",
                 e
             )
         })
 }
 
+#[cfg(windows)]
+fn windows_preload_pdfium(path: &str) {
+    use std::ffi::OsStr;
+    use std::os::windows::ffi::OsStrExt;
+
+    extern "system" {
+        fn LoadLibraryExW(
+            lp_file_name: *const u16,
+            h_file: *mut std::ffi::c_void,
+            dw_flags: u32,
+        ) -> *mut std::ffi::c_void;
+    }
+
+    // LOAD_WITH_ALTERED_SEARCH_PATH: use the DLL's own directory as the
+    // base when resolving its import dependencies.
+    const LOAD_WITH_ALTERED_SEARCH_PATH: u32 = 0x0000_0008;
+
+    let wide: Vec<u16> = OsStr::new(path)
+        .encode_wide()
+        .chain(std::iter::once(0u16))
+        .collect();
+
+    // Ignore the return value — on failure we fall through to pdfium-render's
+    // own load attempt which will surface the OS error code.
+    unsafe {
+        LoadLibraryExW(
+            wide.as_ptr(),
+            std::ptr::null_mut(),
+            LOAD_WITH_ALTERED_SEARCH_PATH,
+        );
+    }
+}
+
 fn collect_paragraph_records(
     grouped_paragraphs: Vec<(String, bool, usize, f32)>,
 ) -> Vec<ParagraphRecord> {