BabelDOC 0.6.3__tar.gz → 0.7.0__tar.gz

{babeldoc-0.6.3 → babeldoc-0.7.0}/PKG-INFO

@@ -1,58 +1,57 @@
 Metadata-Version: 2.4
 Name: BabelDOC
-Version: 0.6.3
-Summary: Yet Another Document Translator
-Project-URL: Homepage, https://github.com/funstory-ai/BabelDOC
-Project-URL: Issues, https://github.com/funstory-ai/BabelDOC/issues
-Author-email: awwaawwa <aw@funstory.ai>
-Maintainer-email: awwaawwa <aw@funstory.ai>
-License-Expression: AGPL-3.0
-License-File: LICENSE
-Keywords: PDF
-Classifier: Operating System :: OS Independent
+Version: 0.7.0
 Classifier: Programming Language :: Python :: 3
-Requires-Python: <3.14,>=3.10
-Requires-Dist: bitstring>=4.3.0
-Requires-Dist: chardet>=5.2.0
-Requires-Dist: charset-normalizer>=2.0.0
+Classifier: Operating System :: OS Independent
 Requires-Dist: configargparse>=1.7
-Requires-Dist: cryptography>=46.0.7
-Requires-Dist: freetype-py>=2.5.1
 Requires-Dist: httpx[socks]>=0.27.0
 Requires-Dist: huggingface-hub>=0.27.0
-Requires-Dist: hyperscan>=0.7.13
-Requires-Dist: levenshtein>=0.27.1
-Requires-Dist: msgpack>=1.1.0
 Requires-Dist: numpy>=2.0.2
 Requires-Dist: onnx>=1.21.0
-Requires-Dist: onnxruntime<1.24,>=1.16.1; python_version < '3.11'
-Requires-Dist: onnxruntime>=1.16.1; python_version >= '3.11'
+Requires-Dist: onnxruntime>=1.16.1,<1.24 ; python_full_version < '3.11'
+Requires-Dist: onnxruntime>=1.16.1 ; python_full_version >= '3.11'
 Requires-Dist: openai>=1.59.3
-Requires-Dist: opencv-python-headless>=4.10.0.84
 Requires-Dist: orjson>=3.10.14
+Requires-Dist: charset-normalizer>=2.0.0
+Requires-Dist: cryptography>=46.0.7
 Requires-Dist: peewee>=3.17.8
 Requires-Dist: psutil>=7.0.0
+Requires-Dist: pymupdf>=1.26.7,<2
+Requires-Dist: rich>=13.9.4
+Requires-Dist: toml>=0.10.2
+Requires-Dist: tqdm>=4.67.1
+Requires-Dist: xsdata[cli,lxml,soap]>=24.12
+Requires-Dist: msgpack>=1.1.0
 Requires-Dist: pydantic>=2.10.6
-Requires-Dist: pymupdf>=1.26.7
+Requires-Dist: tenacity>=9.0.0
+Requires-Dist: scikit-image>=0.25.2
+Requires-Dist: freetype-py>=2.5.1
+Requires-Dist: tiktoken>=0.9.0
+Requires-Dist: levenshtein>=0.27.1
+Requires-Dist: opencv-python-headless>=4.10.0.84
 Requires-Dist: pyzstd>=0.17.0
-Requires-Dist: rich>=13.9.4
+Requires-Dist: hyperscan>=0.7.13
 Requires-Dist: rtree>=1.4.0
-Requires-Dist: scikit-image>=0.25.2
-Requires-Dist: scikit-learn>=1.7.1
+Requires-Dist: chardet>=5.2.0
 Requires-Dist: scipy>=1.15.3
-Requires-Dist: tenacity>=9.0.0
-Requires-Dist: tiktoken>=0.9.0
-Requires-Dist: toml>=0.10.2
-Requires-Dist: tqdm>=4.67.1
 Requires-Dist: uharfbuzz>=0.50.2
-Requires-Dist: xsdata[cli,lxml,soap]>=24.12
+Requires-Dist: scikit-learn>=1.7.1
+Requires-Dist: onnxruntime-gpu>=1.16.1 ; extra == 'cuda'
+Requires-Dist: onnxruntime-directml>=1.16.1 ; extra == 'directml'
+Requires-Dist: memray>=1.17.1 ; extra == 'memray'
 Provides-Extra: cuda
-Requires-Dist: onnxruntime-gpu>=1.16.1; extra == 'cuda'
 Provides-Extra: directml
-Requires-Dist: onnxruntime-directml>=1.16.1; extra == 'directml'
 Provides-Extra: memray
-Requires-Dist: memray>=1.17.1; extra == 'memray'
-Description-Content-Type: text/markdown
+License-File: LICENSE
+Summary: Yet Another Document Translator
+Keywords: PDF
+Author-email: awwaawwa <aw@funstory.ai>
+Maintainer-email: awwaawwa <aw@funstory.ai>
+License-Expression: AGPL-3.0
+Requires-Python: >=3.10, <3.14
+Description-Content-Type: text/markdown; charset=UTF-8; variant=GFM
+Project-URL: Homepage, https://github.com/funstory-ai/BabelDOC
+Project-URL: Issues, https://github.com/funstory-ai/BabelDOC/issues
 
 <div align="center">
 
@@ -204,9 +203,9 @@ uv run babeldoc --files example.pdf --files example2.pdf --openai --openai-model
 - `--disable-rich-text-translate`: Disable rich text translation (may help improve compatibility with some PDFs)
 - `--enhance-compatibility`: Enable all compatibility enhancement options (equivalent to --skip-clean --dual-translate-first --disable-rich-text-translate)
 - `--use-alternating-pages-dual`: Use alternating pages mode for dual PDF. When enabled, original and translated pages are arranged in alternate order. When disabled (default), original and translated pages are shown side by side on the same page.
-- `--watermark-output-mode`: Control watermark output mode: 'watermarked' (default) adds watermark to translated PDF, 'no_watermark' doesn't add watermark, 'both' outputs both versions.
+- `--watermark-output-mode`: [DEPRECATED] Watermarks are no longer added; this option still controls output file names and count.
 - `--max-pages-per-part`: Maximum number of pages per part for split translation. If not set, no splitting will be performed.
-- `--no-watermark`: [DEPRECATED] Use --watermark-output-mode=no_watermark instead.
+- `--no-watermark`: [DEPRECATED] Watermarks are no longer added; this option still affects output file names and count.
 - `--translate-table-text`: Translate table text (experimental, default: False)
 - `--formular-font-pattern`: Font pattern to identify formula text (default: None)
 - `--formular-char-pattern`: Character pattern to identify formula text (default: None)
@@ -333,10 +332,10 @@ skip-clean = false
 dual-translate-first = false
 disable-rich-text-translate = false
 use-alternating-pages-dual = false
-watermark-output-mode = "watermarked"  # Choices: "watermarked", "no_watermark", "both"
+watermark-output-mode = "watermarked"  # DEPRECATED: No watermark is added; this still controls output file names and count.
 max-pages-per-part = 50  # Automatically split the document for translation and merge it back.
 only_include_translated_page = false # Only include translated pages in the output PDF. Effective only when `pages` is used.
-# no-watermark = false  # DEPRECATED: Use watermark-output-mode instead
+# no-watermark = false  # DEPRECATED: No watermark is added; this still affects output file names and count.
 skip-scanned-detection = false  # Skip scanned document detection for faster processing
 auto_extract_glossary = true # Set to false to disable automatic term extraction
 formular_font_pattern = "" # Font pattern for formula text
@@ -379,7 +378,7 @@ report-interval = 0.5
 
 ## Python API
 
-The current recommended way to call BabelDOC in Python is to call the `high_level.do_translate_async_stream` function of [pdf2zh next](https://github.com/PDFMathTranslate/PDFMathTranslate-next).
+The current recommended way to call BabelDOC in Python is to call the `high_level.do_translate_async_stream` function of [pdf2zh next](https://pdf2zh-next.com/).
 
 > [!WARNING]
 > **All APIs of BabelDOC should be considered as internal APIs, and any direct use of BabelDOC is not supported.**
@@ -439,7 +438,7 @@ This project uses a combination of [Semantic Versioning](https://semver.org/) an
 
 > [!NOTE]
 >
-> The API compatibility here mainly refers to the compatibility with [pdf2zh_next](https://github.com/PDFMathTranslate/PDFMathTranslate-next).
+> The API compatibility here mainly refers to the compatibility with [pdf2zh_next](https://pdf2zh-next.com/).
 
 
 - MAJOR: Incremented by 1 when API incompatible changes are made or when proud improvements are implemented.
@@ -500,3 +499,4 @@ step by step.
 >     *   If the document is identified as heavily scanned (e.g., >80% scanned pages) AND `auto_enable_ocr_workaround` is `true` (i.e., `translation_config.auto_enable_ocr_workaround` is true), the system will then attempt to set both `ocr_workaround` to `true` and `skip_scanned_detection` to `true`.
 >
 > This means that `--auto-enable-ocr-workaround` effectively gives the system control to enable OCR processing for scanned documents, potentially overriding manual settings for `--ocr-workaround` and `--skip_scanned_detection` based on its detection results. If the document is *not* detected as heavily scanned, then the initial `false` values for `ocr_workaround` and `skip_scanned_detection` (forced by `--auto-enable-ocr-workaround` at the `TranslationConfig` initialization stage) will remain in effect unless changed by other logic.
+

{babeldoc-0.6.3 → babeldoc-0.7.0}/README.md

@@ -148,9 +148,9 @@ uv run babeldoc --files example.pdf --files example2.pdf --openai --openai-model
 - `--disable-rich-text-translate`: Disable rich text translation (may help improve compatibility with some PDFs)
 - `--enhance-compatibility`: Enable all compatibility enhancement options (equivalent to --skip-clean --dual-translate-first --disable-rich-text-translate)
 - `--use-alternating-pages-dual`: Use alternating pages mode for dual PDF. When enabled, original and translated pages are arranged in alternate order. When disabled (default), original and translated pages are shown side by side on the same page.
-- `--watermark-output-mode`: Control watermark output mode: 'watermarked' (default) adds watermark to translated PDF, 'no_watermark' doesn't add watermark, 'both' outputs both versions.
+- `--watermark-output-mode`: [DEPRECATED] Watermarks are no longer added; this option still controls output file names and count.
 - `--max-pages-per-part`: Maximum number of pages per part for split translation. If not set, no splitting will be performed.
-- `--no-watermark`: [DEPRECATED] Use --watermark-output-mode=no_watermark instead.
+- `--no-watermark`: [DEPRECATED] Watermarks are no longer added; this option still affects output file names and count.
 - `--translate-table-text`: Translate table text (experimental, default: False)
 - `--formular-font-pattern`: Font pattern to identify formula text (default: None)
 - `--formular-char-pattern`: Character pattern to identify formula text (default: None)
@@ -277,10 +277,10 @@ skip-clean = false
 dual-translate-first = false
 disable-rich-text-translate = false
 use-alternating-pages-dual = false
-watermark-output-mode = "watermarked"  # Choices: "watermarked", "no_watermark", "both"
+watermark-output-mode = "watermarked"  # DEPRECATED: No watermark is added; this still controls output file names and count.
 max-pages-per-part = 50  # Automatically split the document for translation and merge it back.
 only_include_translated_page = false # Only include translated pages in the output PDF. Effective only when `pages` is used.
-# no-watermark = false  # DEPRECATED: Use watermark-output-mode instead
+# no-watermark = false  # DEPRECATED: No watermark is added; this still affects output file names and count.
 skip-scanned-detection = false  # Skip scanned document detection for faster processing
 auto_extract_glossary = true # Set to false to disable automatic term extraction
 formular_font_pattern = "" # Font pattern for formula text
@@ -323,7 +323,7 @@ report-interval = 0.5
 
 ## Python API
 
-The current recommended way to call BabelDOC in Python is to call the `high_level.do_translate_async_stream` function of [pdf2zh next](https://github.com/PDFMathTranslate/PDFMathTranslate-next).
+The current recommended way to call BabelDOC in Python is to call the `high_level.do_translate_async_stream` function of [pdf2zh next](https://pdf2zh-next.com/).
 
 > [!WARNING]
 > **All APIs of BabelDOC should be considered as internal APIs, and any direct use of BabelDOC is not supported.**
@@ -383,7 +383,7 @@ This project uses a combination of [Semantic Versioning](https://semver.org/) an
 
 > [!NOTE]
 >
-> The API compatibility here mainly refers to the compatibility with [pdf2zh_next](https://github.com/PDFMathTranslate/PDFMathTranslate-next).
+> The API compatibility here mainly refers to the compatibility with [pdf2zh_next](https://pdf2zh-next.com/).
 
 
 - MAJOR: Incremented by 1 when API incompatible changes are made or when proud improvements are implemented.

babeldoc-0.7.0/babeldoc/__init__.py

@@ -0,0 +1 @@
+__version__ = "0.7.0"

{babeldoc-0.6.3 → babeldoc-0.7.0}/babeldoc/const.py

@@ -6,7 +6,7 @@ import subprocess
 import threading
 from pathlib import Path
 
-__version__ = "0.6.3"
+__version__ = "0.7.0"
 
 CACHE_FOLDER = Path.home() / ".cache" / "babeldoc"
 

{babeldoc-0.6.3 → babeldoc-0.7.0}/babeldoc/docvision/doclayout.py

@@ -10,7 +10,8 @@ import numpy as np
 
 from babeldoc.docvision.base_doclayout import DocLayoutModel
 from babeldoc.docvision.base_doclayout import YoloResult
-from babeldoc.format.pdf.document_il.utils.mupdf_helper import get_no_rotation_img
+from babeldoc.format.pdf.document_il.utils.raster_geometry import RasterGeometry
+from babeldoc.format.pdf.document_il.utils.raster_geometry import with_target_long_edge
 
 try:
     import onnx
@@ -178,7 +179,14 @@ class OnnxModel(DocLayoutModel):
         boxes[..., :4] = (boxes[..., :4] - [pad_x, pad_y, pad_x, pad_y]) / gain
         return boxes
 
-    def predict(self, image, imgsz=800, batch_size=16, **kwargs):
+    def predict(
+        self,
+        image,
+        imgsz=800,
+        batch_size=16,
+        geometry: RasterGeometry | None = None,
+        **kwargs,
+    ):
         """
         Predict the layout of document pages.
 
@@ -237,6 +245,13 @@ class OnnxModel(DocLayoutModel):
                         preds[..., :4],
                         orig_shapes[j],
                     )
+                    if geometry is not None:
+                        preds[..., [0, 2]] = geometry.px_len_to_pt(
+                            preds[..., [0, 2]], "x"
+                        )
+                        preds[..., [1, 3]] = geometry.px_len_to_pt(
+                            preds[..., [1, 3]], "y"
+                        )
                 results.append(YoloResult(boxes_data=preds, names=self._names))
 
         return results
@@ -253,14 +268,14 @@ class OnnxModel(DocLayoutModel):
         for page in pages:
             translate_config.raise_if_cancelled()
             with self.lock:
-                # pix = mupdf_doc[page.page_number].get_pixmap(dpi=72)
-                pix = get_no_rotation_img(mupdf_doc[page.page_number])
-            image = np.frombuffer(pix.samples, np.uint8).reshape(
-                pix.height,
-                pix.width,
-                3,
-            )[:, :, ::-1]
-            predict_result = self.predict(image)[0]
+                geometry = with_target_long_edge(
+                    mupdf_doc[page.page_number],
+                    72,
+                    self._FIXED_IMGSZ,
+                    normalize_rotation=True,
+                )
+            image = geometry.image[:, :, ::-1]
+            predict_result = self.predict(image, geometry=geometry)[0]
             save_debug_image(
                 image,
                 predict_result,

{babeldoc-0.6.3 → babeldoc-0.7.0}/babeldoc/docvision/rpc_doclayout.py

@@ -17,7 +17,8 @@ import babeldoc
 from babeldoc.docvision.base_doclayout import DocLayoutModel
 from babeldoc.docvision.base_doclayout import YoloBox
 from babeldoc.docvision.base_doclayout import YoloResult
-from babeldoc.format.pdf.document_il.utils.mupdf_helper import get_no_rotation_img
+from babeldoc.format.pdf.document_il.utils.raster_geometry import RasterGeometry
+from babeldoc.format.pdf.document_il.utils.raster_geometry import with_target_long_edge
 
 logger = logging.getLogger(__name__)
 
@@ -209,6 +210,7 @@ class RpcDocLayoutModel(DocLayoutModel):
         host: str = None,
         result_container: ResultContainer | None = None,
         imgsz: int = 1024,
+        geometry: RasterGeometry | None = None,
     ) -> ResultContainer:
         """Predict the layout of document pages using RPC service."""
         if result_container is None:
@@ -232,6 +234,10 @@ class RpcDocLayoutModel(DocLayoutModel):
                     )
                     for x in pred["boxes"]
                 ]
+                if geometry is not None:
+                    for box in boxes:
+                        box.xyxy[[0, 2]] = geometry.px_len_to_pt(box.xyxy[[0, 2]], "x")
+                        box.xyxy[[1, 3]] = geometry.px_len_to_pt(box.xyxy[[1, 3]], "y")
                 result_container.result = YoloResult(
                     boxes=boxes,
                     names={int(k): v for k, v in pred["names"].items()},
@@ -244,11 +250,17 @@ class RpcDocLayoutModel(DocLayoutModel):
         if isinstance(image, np.ndarray) and len(image.shape) == 3:
             image = [image]
 
+        geometry = kwargs.get("geometry")
         result_containers = [ResultContainer() for _ in image]
         predict_thread = ThreadPoolExecutor(max_workers=len(image))
         for img, result_container in zip(image, result_containers, strict=True):
             predict_thread.submit(
-                self.predict_image, img, self.host, result_container, 800
+                self.predict_image,
+                img,
+                self.host,
+                result_container,
+                800,
+                geometry,
             )
         predict_thread.shutdown(wait=True)
         result = [result_container.result for result_container in result_containers]
@@ -259,14 +271,20 @@ class RpcDocLayoutModel(DocLayoutModel):
     ):
         translate_config.raise_if_cancelled()
         with self.lock:
-            # pix = mupdf_doc[page.page_number].get_pixmap(dpi=72)
-            pix = get_no_rotation_img(mupdf_doc[page.page_number])
-        image = np.frombuffer(pix.samples, np.uint8).reshape(
-            pix.height,
-            pix.width,
-            3,
-        )[:, :, ::-1]
-        predict_result = self.predict_image(image, self.host, None, 800)
+            geometry = with_target_long_edge(
+                mupdf_doc[page.page_number],
+                72,
+                800,
+                normalize_rotation=True,
+            )
+        image = geometry.image[:, :, ::-1]
+        predict_result = self.predict_image(
+            image,
+            self.host,
+            None,
+            800,
+            geometry,
+        )
         save_debug_image(image, predict_result, page.page_number + 1)
         return page, predict_result