BabelDOC 0.5.23__tar.gz → 0.6.0__tar.gz

{babeldoc-0.5.23 → babeldoc-0.6.0}/.github/workflows/codeql.yml

@@ -55,7 +55,7 @@ jobs:
         # your codebase is analyzed, see https://docs.github.com/en/code-security/code-scanning/creating-an-advanced-setup-for-code-scanning/codeql-code-scanning-for-compiled-languages
     steps:
     - name: Checkout repository
-      uses: actions/checkout@v5
+      uses: actions/checkout@v6
 
     # Initializes the CodeQL tools for scanning.
     - name: Initialize CodeQL

{babeldoc-0.5.23 → babeldoc-0.6.0}/.github/workflows/docs.yml

@@ -9,7 +9,7 @@ jobs:
   deploy:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v5
+      - uses: actions/checkout@v6
         with:
           fetch-depth: 0
       - name: Configure Git Credentials
@@ -24,7 +24,7 @@ jobs:
           cache-dependency-glob: "uv.lock"
           activate-environment: true
       - run: echo "cache_id=$(date --utc '+%V')" >> $GITHUB_ENV 
-      - uses: actions/cache@v4
+      - uses: actions/cache@v5
         with:
           key: mkdocs-material-${{ env.cache_id }}
           path: .cache

{babeldoc-0.5.23 → babeldoc-0.6.0}/.github/workflows/labeler.yml

@@ -22,10 +22,10 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Check out the repository
-        uses: actions/checkout@v5
+        uses: actions/checkout@v6
 
       - name: Run Labeler
-        uses: crazy-max/ghaction-github-labeler@24d110aa46a59976b8a7f35518cb7f14f434c916 # v5.3.0
+        uses: crazy-max/ghaction-github-labeler@548a7c3603594ec17c819e1239f281a3b801ab4d # v6.0.0
         with:
           skip-delete: true
           dry-run: ${{ github.event_name == 'pull_request' }}

{babeldoc-0.5.23 → babeldoc-0.6.0}/.github/workflows/lint.yml

@@ -10,7 +10,7 @@ jobs:
       fail-fast: false
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v5
+      - uses: actions/checkout@v6
       - name: Ruff
         uses: astral-sh/ruff-action@v3
       - name: AutoCorrect

{babeldoc-0.5.23 → babeldoc-0.6.0}/.github/workflows/pr-lint.yml

@@ -9,7 +9,7 @@ jobs:
     name: runner / ruff
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v5
+      - uses: actions/checkout@v6
       
       - name: Install Python
         uses: actions/setup-python@v6
@@ -20,7 +20,7 @@ jobs:
         run: pip install ruff
         
       - name: Install reviewdog
-        uses: reviewdog/action-setup@d8edfce3dd5e1ec6978745e801f9c50b5ef80252 # v1.4.0
+        uses: reviewdog/action-setup@d8a7baabd7f3e8544ee4dbde3ee41d0011c3a93f # v1.5.0
         with:
           reviewdog_version: latest
           
@@ -34,7 +34,7 @@ jobs:
     name: runner / autocorrect
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v5
+      - uses: actions/checkout@v6
       - name: AutoCorrect
         uses: huacnlee/autocorrect-action@bf91ab3904c2908dd8e71312a8a83ed1eb632997 # v2.13.3
       - name: Report ReviewDog

{babeldoc-0.5.23 → babeldoc-0.6.0}/.github/workflows/publish-to-pypi.yml

@@ -28,7 +28,7 @@ jobs:
     outputs:
       is_release: ${{ steps.check-version.outputs.tag }}
     steps:
-      - uses: actions/checkout@v5
+      - uses: actions/checkout@v6
         with:
           persist-credentials: true
           fetch-depth: 2
@@ -55,6 +55,11 @@ jobs:
           version-command: |
             cat pyproject.toml | grep "version = " | head -n 1 | awk -F'"' '{print $2}'
 
+      - name: Check manual release notes
+        if: steps.check-version.outputs.tag
+        run: |
+          test -f "docs/release-notes/${{ steps.check-version.outputs.tag }}.md"
+
       - name: Install Dependencies
         run: |
           uv sync
@@ -69,7 +74,7 @@ jobs:
         run: "uv build"
 
       - name: Store the distribution packages
-        uses: actions/upload-artifact@v4.6.2
+        uses: actions/upload-artifact@v7.0.0
         with:
           name: python-package-distributions
           path: dist/
@@ -139,7 +144,7 @@ jobs:
       contents: write
       pull-requests: write
     steps:
-      - uses: actions/checkout@v5
+      - uses: actions/checkout@v6
         with:
           persist-credentials: true
           fetch-depth: 2
@@ -151,4 +156,22 @@ jobs:
           publish: ${{ needs.build.outputs.is_release != '' }}
           tag: ${{ needs.build.outputs.is_release }}
         env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Prepend manual release notes
+        if: needs.build.outputs.is_release != ''
+        env:
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          TAG: ${{ needs.build.outputs.is_release }}
+        run: |
+          set -euo pipefail
+          notes_path="docs/release-notes/${TAG}.md"
+          current_body="$(mktemp)"
+          combined_body="$(mktemp)"
+          gh release view "${TAG}" --json body --jq .body > "${current_body}"
+          {
+            cat "${notes_path}"
+            printf '\n\n---\n\n'
+            cat "${current_body}"
+          } > "${combined_body}"
+          gh release edit "${TAG}" --notes-file "${combined_body}"

{babeldoc-0.5.23 → babeldoc-0.6.0}/.github/workflows/test.yml

@@ -18,12 +18,12 @@ jobs:
         python-version: ["3.10", "3.11", "3.12", "3.13"]
 
     steps:
-      - uses: actions/checkout@v5
+      - uses: actions/checkout@v6
         with:
           persist-credentials: false
       - name: Cached Assets
         id: cache-assets
-        uses: actions/cache@v4.2.0
+        uses: actions/cache@v5
         with:
           path: ~/.cache/babeldoc
           key: babeldoc-assets-${{ hashFiles('babeldoc/assets/embedding_assets_metadata.py') }}

{babeldoc-0.5.23 → babeldoc-0.6.0}/.gitignore

@@ -25,7 +25,9 @@ memray*
 .vscode
 __pycache__
 .ruff_cache
+.tmp
 yadt.toml
+yadt.toml*
 examples/
 /make_gif.py
 /dist
@@ -41,9 +43,9 @@ babeldoc/format/txt
 # uv
 uv.lock
 
-# Claude Code memory file
-CLAUDE.md
 /.claude
 babeldoc/format/playground
 temp.jpg
-AGENTS.md
+.serena
+.codex
+

{babeldoc-0.5.23 → babeldoc-0.6.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: BabelDOC
-Version: 0.5.23
+Version: 0.6.0
 Summary: Yet Another Document Translator
 Project-URL: Homepage, https://github.com/funstory-ai/BabelDOC
 Project-URL: Issues, https://github.com/funstory-ai/BabelDOC/issues
@@ -34,7 +34,6 @@ Requires-Dist: psutil>=7.0.0
 Requires-Dist: pydantic>=2.10.6
 Requires-Dist: pymupdf>=1.25.1
 Requires-Dist: pyzstd>=0.17.0
-Requires-Dist: rapidocr-onnxruntime>=1.4.4
 Requires-Dist: rich>=13.9.4
 Requires-Dist: rtree>=1.4.0
 Requires-Dist: scikit-image>=0.25.2

babeldoc-0.6.0/babeldoc/__init__.py

@@ -0,0 +1 @@
+__version__ = "0.6.0"

{babeldoc-0.5.23 → babeldoc-0.6.0}/babeldoc/assets/assets.py

@@ -17,10 +17,6 @@ from babeldoc.assets.embedding_assets_metadata import (
 from babeldoc.assets.embedding_assets_metadata import EMBEDDING_FONT_METADATA
 from babeldoc.assets.embedding_assets_metadata import FONT_METADATA_URL
 from babeldoc.assets.embedding_assets_metadata import FONT_URL_BY_UPSTREAM
-from babeldoc.assets.embedding_assets_metadata import (
-    TABLE_DETECTION_RAPIDOCR_MODEL_SHA3_256,
-)
-from babeldoc.assets.embedding_assets_metadata import TABLE_DETECTION_RAPIDOCR_MODEL_URL
 from babeldoc.assets.embedding_assets_metadata import TIKTOKEN_CACHES
 from babeldoc.const import get_cache_file_path
 from tenacity import retry
@@ -38,20 +34,33 @@ _FASTEST_FONT_METADATA: dict | None = None
 class ResultContainer:
     def __init__(self):
         self.result = None
+        self.exception: BaseException | None = None
 
     def set_result(self, result):
         self.result = result
 
+    def set_exception(self, exc: BaseException):
+        self.exception = exc
+
 
 def run_in_another_thread(coro):
     result_container = ResultContainer()
 
     def _wrapper():
-        result_container.set_result(asyncio.run(coro))
+        try:
+            result_container.set_result(asyncio.run(coro))
+        except BaseException as exc:  # noqa: BLE001
+            result_container.set_exception(exc)
 
     thread = threading.Thread(target=_wrapper)
     thread.start()
     thread.join()
+    if result_container.exception is not None:
+        msg = (
+            "asset coroutine failed: "
+            f"{type(result_container.exception).__name__}: {result_container.exception}"
+        )
+        raise RuntimeError(msg) from result_container.exception
     return result_container.result
 
 
@@ -248,23 +257,10 @@ async def get_doclayout_onnx_model_path_async(client: httpx.AsyncClient | None =
 async def get_table_detection_rapidocr_model_path_async(
     client: httpx.AsyncClient | None = None,
 ):
-    onnx_path = get_cache_file_path("ch_PP-OCRv4_det_infer.onnx", "models")
-    if verify_file(onnx_path, TABLE_DETECTION_RAPIDOCR_MODEL_SHA3_256):
-        return onnx_path
-
-    logger.info("table detection rapidocr model not found or corrupted, downloading...")
-    fastest_upstream, _ = await get_fastest_upstream_for_model(client)
-    if fastest_upstream is None:
-        logger.error("Failed to get fastest upstream")
-        exit(1)
-
-    url = TABLE_DETECTION_RAPIDOCR_MODEL_URL[fastest_upstream]
-
-    await download_file(client, url, onnx_path, TABLE_DETECTION_RAPIDOCR_MODEL_SHA3_256)
-    logger.info(
-        f"Download table detection rapidocr model from {fastest_upstream} success"
-    )
-    return onnx_path
+    _ = client
+    dummy_path = get_cache_file_path("rapidocr-retired-dummy.txt", "models")
+    dummy_path.write_text("RapidOCR table detection is retired.\n", encoding="utf-8")
+    return dummy_path
 
 
 def get_doclayout_onnx_model_path():
@@ -457,12 +453,9 @@ async def async_warmup():
     _ = encoding_for_model("gpt-4o")
     async with httpx.AsyncClient() as client:
         onnx_task = asyncio.create_task(get_doclayout_onnx_model_path_async(client))
-        onnx_task2 = asyncio.create_task(
-            get_table_detection_rapidocr_model_path_async(client)
-        )
         font_tasks = asyncio.create_task(download_all_fonts_async(client))
         cmap_tasks = asyncio.create_task(download_all_cmaps_async(client))
-        await asyncio.gather(onnx_task, onnx_task2, font_tasks, cmap_tasks)
+        await asyncio.gather(onnx_task, font_tasks, cmap_tasks)
 
 
 def warmup():
@@ -502,12 +495,6 @@ def generate_all_assets_file_list():
             "sha3_256": DOCLAYOUT_YOLO_DOCSTRUCTBENCH_IMGSZ1024ONNX_SHA3_256,
         },
     )
-    result["models"].append(
-        {
-            "name": "ch_PP-OCRv4_det_infer.onnx",
-            "sha3_256": TABLE_DETECTION_RAPIDOCR_MODEL_SHA3_256,
-        },
-    )
     return result
 
 
@@ -620,6 +607,4 @@ if __name__ == "__main__":
     logging.getLogger("httpcore").setLevel(logging.WARNING)
     # warmup()
     # generate_offline_assets_package()
-    # restore_offline_assets_package(Path(
-    #     '/Users/aw/.cache/babeldoc/assets/offline_assets_33971e4940e90ba0c35baacda44bbe83b214f4703a7bdb8b837de97d0383508c.zip'))
     # warmup()

{babeldoc-0.5.23 → babeldoc-0.6.0}/babeldoc/assets/embedding_assets_metadata.py

@@ -4,10 +4,6 @@ DOCLAYOUT_YOLO_DOCSTRUCTBENCH_IMGSZ1024ONNX_SHA3_256 = (
     "60be061226930524958b5465c8c04af3d7c03bcb0beb66454f5da9f792e3cf2a"
 )
 
-TABLE_DETECTION_RAPIDOCR_MODEL_SHA3_256 = (
-    "062f4619afe91b33147c033acadecbb53f2a7b99ac703d157b96d5b10948da5e"
-)
-
 TIKTOKEN_CACHES = {
     "fb374d419588a4632f3f557e76b4b70aebbca790": "cb04bcda5782cfbbe77f2f991d92c0ea785d9496ef1137c91dfc3c8c324528d6"
 }
@@ -39,12 +35,6 @@ DOC_LAYOUT_ONNX_MODEL_URL = {
     "modelscope": "https://www.modelscope.cn/models/AI-ModelScope/DocLayout-YOLO-DocStructBench-onnx/resolve/master/doclayout_yolo_docstructbench_imgsz1024.onnx",
 }
 
-TABLE_DETECTION_RAPIDOCR_MODEL_URL = {
-    "huggingface": "https://huggingface.co/spaces/RapidAI/RapidOCR/resolve/main/models/text_det/ch_PP-OCRv4_det_infer.onnx",
-    "hf-mirror": "https://hf-mirror.com/spaces/RapidAI/RapidOCR/resolve/main/models/text_det/ch_PP-OCRv4_det_infer.onnx",
-    "modelscope": "https://www.modelscope.cn/models/RapidAI/RapidOCR/resolve/master/onnx/PP-OCRv4/det/ch_PP-OCRv4_det_infer.onnx",
-}
-
 # from https://github.com/funstory-ai/BabelDOC-Assets/blob/main/font_metadata.json
 EMBEDDING_FONT_METADATA = {
     "GoNotoKurrent-Bold.ttf": {

{babeldoc-0.5.23 → babeldoc-0.6.0}/babeldoc/const.py

@@ -6,7 +6,7 @@ import subprocess
 import threading
 from pathlib import Path
 
-__version__ = "0.5.23"
+__version__ = "0.6.0"
 
 CACHE_FOLDER = Path.home() / ".cache" / "babeldoc"
 

{babeldoc-0.5.23 → babeldoc-0.6.0}/babeldoc/docvision/doclayout.py

@@ -37,6 +37,8 @@ os_name = platform.system()
 
 
 class OnnxModel(DocLayoutModel):
+    _FIXED_IMGSZ = 1024  # fixed input size for static-shape CoreML
+
     def __init__(self, model_path: str):
         self.model_path = model_path
 
@@ -47,18 +49,52 @@ class OnnxModel(DocLayoutModel):
         providers = []
 
         available_providers = onnxruntime.get_available_providers()
-        for provider in available_providers:
-            # disable dml|cuda|
-            # directml/cuda may encounter problems under special circumstances
-            if re.match(r"cpu", provider, re.IGNORECASE):
-                logger.info(f"Available Provider: {provider}")
-                providers.append(provider)
+        use_coreml = os_name == "Darwin" and any(
+            re.match(r"coreml", p, re.IGNORECASE) for p in available_providers
+        )
+        if use_coreml:
+            # Fix input shapes to [1, 3, 1024, 1024] so CoreML can take over
+            # 96%+ of the graph nodes (658/681) instead of just 3/823.
+            model = self._make_static_model(model, self._FIXED_IMGSZ)
+            providers = [
+                "CoreMLExecutionProvider",
+                "CPUExecutionProvider",
+            ]
+            logger.info(
+                "Using CoreMLExecutionProvider with static input "
+                f"[1, 3, {self._FIXED_IMGSZ}, {self._FIXED_IMGSZ}]"
+            )
+        else:
+            for provider in available_providers:
+                # disable dml|cuda|
+                # directml/cuda may encounter problems under special circumstances
+                if re.match(r"cpu", provider, re.IGNORECASE):
+                    logger.info(f"Available Provider: {provider}")
+                    providers.append(provider)
         self.model = onnxruntime.InferenceSession(
             model.SerializeToString(),
             providers=providers,
         )
         self.lock = threading.Lock()
 
+    @staticmethod
+    def _make_static_model(model, imgsz):
+        """Rewrite dynamic input dims to fixed [1, 3, imgsz, imgsz] and run
+        shape inference so all intermediate shapes become static."""
+        from onnx import shape_inference
+
+        for inp in model.graph.input:
+            if inp.name == "images":
+                dim = inp.type.tensor_type.shape.dim
+                for i, val in enumerate([1, 3, imgsz, imgsz]):
+                    dim[i].ClearField("dim_param")
+                    dim[i].dim_value = val
+        try:
+            model = shape_inference.infer_shapes(model)
+        except Exception:
+            pass  # best-effort; CoreML still benefits from fixed input
+        return model
+
     @staticmethod
     def from_pretrained():
         pth = get_doclayout_onnx_model_path()
@@ -97,9 +133,9 @@ class OnnxModel(DocLayoutModel):
             interpolation=cv2.INTER_LINEAR,
         )
 
-        # Calculate padding size and align to stride multiple
-        pad_w = (new_w - resized_w) % self.stride
-        pad_h = (new_h - resized_h) % self.stride
+        # Pad to full target size (enables static-shape CoreML acceleration)
+        pad_w = new_w - resized_w
+        pad_h = new_h - resized_h
         top, bottom = pad_h // 2, pad_h - pad_h // 2
         left, right = pad_w // 2, pad_w - pad_w // 2