BabelDOC 0.3.43__tar.gz → 0.3.44__tar.gz

{babeldoc-0.3.43 → babeldoc-0.3.44}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: BabelDOC
-Version: 0.3.43
+Version: 0.3.44
 Summary: Yet Another Document Translator
 Project-URL: Homepage, https://github.com/funstory-ai/BabelDOC
 Project-URL: Issues, https://github.com/funstory-ai/BabelDOC/issues

babeldoc-0.3.44/babeldoc/__init__.py

@@ -0,0 +1 @@
+__version__ = "0.3.44"

{babeldoc-0.3.43 → babeldoc-0.3.44}/babeldoc/const.py

@@ -3,7 +3,7 @@ import shutil
 import subprocess
 from pathlib import Path
 
-__version__ = "0.3.43"
+__version__ = "0.3.44"
 
 CACHE_FOLDER = Path.home() / ".cache" / "babeldoc"
 

{babeldoc-0.3.43 → babeldoc-0.3.44}/babeldoc/document_il/backend/pdf_creater.py

@@ -386,20 +386,22 @@ class PDFCreater:
             # Get pages from both PDFs
             orig_page = original_pdf[page_id]
             trans_page = translated_pdf[page_id]
-
-            # Calculate total width and use max height
+            rotate_angle = orig_page.rotation
             total_width = orig_page.rect.width + trans_page.rect.width
             max_height = max(orig_page.rect.height, trans_page.rect.height)
-
-            # Create new page with combined width
-            dual_page = dual.new_page(width=total_width, height=max_height)
-
-            # Define rectangles for left and right sides
             left_width = (
                 orig_page.rect.width
                 if not translation_config.dual_translate_first
                 else trans_page.rect.width
             )
+
+            orig_page.set_rotation(0)
+            trans_page.set_rotation(0)
+
+            # Create new page with combined width
+            dual_page = dual.new_page(width=total_width, height=max_height)
+
+            # Define rectangles for left and right sides
             rect_left = pymupdf.Rect(0, 0, left_width, max_height)
             rect_right = pymupdf.Rect(left_width, 0, total_width, max_height)
 
@@ -414,6 +416,7 @@ class PDFCreater:
                     original_pdf,
                     page_id,
                     keep_proportion=True,
+                    rotate=-rotate_angle,
                 )
             except Exception as e:
                 logger.warning(
@@ -429,6 +432,7 @@ class PDFCreater:
                     translated_pdf,
                     page_id,
                     keep_proportion=True,
+                    rotate=-rotate_angle,
                 )
             except Exception as e:
                 logger.warning(

{babeldoc-0.3.43 → babeldoc-0.3.44}/babeldoc/document_il/midend/layout_parser.py

@@ -6,6 +6,7 @@ import numpy as np
 from pymupdf import Document
 
 from babeldoc.document_il import il_version_1
+from babeldoc.document_il.utils.mupdf_helper import get_no_rotation_img
 from babeldoc.document_il.utils.style_helper import GREEN
 from babeldoc.translation_config import TranslationConfig
 
@@ -124,7 +125,8 @@ class LayoutParser:
                     # Convert coordinate system from picture to il
                     # system to the il coordinate system
                     x0, y0, x1, y1 = layout.xyxy
-                    pix = mupdf_doc[page.page_number].get_pixmap()
+                    pix = get_no_rotation_img(mupdf_doc[page.page_number])
+                    # pix = mupdf_doc[page.page_number].get_pixmap()
                     h, w = pix.height, pix.width
                     x0, y0, x1, y1 = (
                         np.clip(int(x0 - 1), 0, w - 1),

{babeldoc-0.3.43 → babeldoc-0.3.44}/babeldoc/document_il/midend/table_parser.py

@@ -6,6 +6,7 @@ import numpy as np
 from pymupdf import Document
 
 from babeldoc.document_il import il_version_1
+from babeldoc.document_il.utils.mupdf_helper import get_no_rotation_img
 from babeldoc.document_il.utils.style_helper import GREEN
 from babeldoc.translation_config import TranslationConfig
 
@@ -136,7 +137,8 @@ class TableParser:
                     # Convert coordinate system from picture to il
                     # system to the il coordinate system
                     x0, y0, x1, y1 = layout.xyxy
-                    pix = mupdf_doc[page.page_number].get_pixmap()
+                    # pix = mupdf_doc[page.page_number].get_pixmap()
+                    pix = get_no_rotation_img(mupdf_doc[page.page_number])
                     h, w = pix.height, pix.width
                     x0, y0, x1, y1 = (
                         np.clip(int(x0 - 1), 0, w - 1),

{babeldoc-0.3.43 → babeldoc-0.3.44}/babeldoc/document_il/translator/translator.py

@@ -255,6 +255,7 @@ class OpenAITranslator(BaseTranslator):
         response = self.client.chat.completions.create(
             model=self.model,
             **self.options,
+            max_tokens=2048,
             messages=[
                 {
                     "role": "user",

babeldoc-0.3.44/babeldoc/document_il/utils/mupdf_helper.py

@@ -0,0 +1,10 @@
+import pymupdf
+
+
+def get_no_rotation_img(page: pymupdf.Page):
+    # return page.get_pixmap(dpi=72)
+    original_rotation = page.rotation
+    page.set_rotation(0)
+    pix = page.get_pixmap(dpi=72)
+    page.set_rotation(original_rotation)
+    return pix

{babeldoc-0.3.43 → babeldoc-0.3.44}/babeldoc/docvision/doclayout.py

@@ -9,6 +9,8 @@ from collections.abc import Generator
 import cv2
 import numpy as np
 
+from babeldoc.document_il.utils.mupdf_helper import get_no_rotation_img
+
 try:
     import onnx
     import onnxruntime
@@ -282,7 +284,8 @@ 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 = mupdf_doc[page.page_number].get_pixmap(dpi=72)
+                pix = get_no_rotation_img(mupdf_doc[page.page_number])
             image = np.fromstring(pix.samples, np.uint8).reshape(
                 pix.height,
                 pix.width,

{babeldoc-0.3.43 → babeldoc-0.3.44}/babeldoc/docvision/rpc_doclayout.py

@@ -14,6 +14,7 @@ from tenacity import stop_after_attempt
 from tenacity import wait_exponential
 
 import babeldoc
+from babeldoc.document_il.utils.mupdf_helper import get_no_rotation_img
 from babeldoc.docvision.doclayout import DocLayoutModel
 from babeldoc.docvision.doclayout import YoloBox
 from babeldoc.docvision.doclayout import YoloResult
@@ -256,7 +257,8 @@ class RpcDocLayoutModel(DocLayoutModel):
     ):
         translate_config.raise_if_cancelled()
         with self.lock:
-            pix = mupdf_doc[page.page_number].get_pixmap(dpi=72)
+            # pix = mupdf_doc[page.page_number].get_pixmap(dpi=72)
+            pix = get_no_rotation_img(mupdf_doc[page.page_number])
         image = np.fromstring(pix.samples, np.uint8).reshape(
             pix.height,
             pix.width,

{babeldoc-0.3.43 → babeldoc-0.3.44}/babeldoc/docvision/table_detection/rapidocr.py

@@ -6,6 +6,7 @@ from collections.abc import Generator
 import cv2
 import numpy as np
 from babeldoc.assets.assets import get_table_detection_rapidocr_model_path
+from babeldoc.document_il.utils.mupdf_helper import get_no_rotation_img
 from babeldoc.docvision.doclayout import YoloBox
 from babeldoc.docvision.doclayout import YoloResult
 from rapidocr_onnxruntime import RapidOCR
@@ -236,7 +237,8 @@ class RapidOCRModel:
         for page in pages:
             translate_config.raise_if_cancelled()
             with self.lock:
-                pix = mupdf_doc[page.page_number].get_pixmap(dpi=72)
+                # pix = mupdf_doc[page.page_number].get_pixmap(dpi=72)
+                pix = get_no_rotation_img(mupdf_doc[page.page_number])
             image = np.fromstring(pix.samples, np.uint8).reshape(
                 pix.height,
                 pix.width,

{babeldoc-0.3.43 → babeldoc-0.3.44}/babeldoc/main.py

@@ -23,7 +23,7 @@ from babeldoc.translation_config import TranslationConfig
 from babeldoc.translation_config import WatermarkOutputMode
 
 logger = logging.getLogger(__name__)
-__version__ = "0.3.43"
+__version__ = "0.3.44"
 
 
 def create_parser():

{babeldoc-0.3.43 → babeldoc-0.3.44}/babeldoc/pdfinterp.py

@@ -377,6 +377,11 @@ class PDFPageInterpreterEx(PDFPageInterpreter):
             ctm = (0, 1, -1, 0, y1, -x0)
         else:
             ctm = (1, 0, 0, 1, -x0, -y0)
+        # ctm_for_ops = copy.copy(ctm)
+        ctm_for_ops = (1, 0, 0, 1, -x0, -y0)
+        ctm = (1, 0, 0, 1, -x0, -y0)
+        if page.rotate == 90 or page.rotate == 270:
+            (x0, y0, x1, y1) = (y0, x1, y1, x0)
         self.il_creater.on_page_start()
         self.il_creater.on_page_crop_box(x0, y0, x1, y1)
         self.device.begin_page(page, ctm)
@@ -391,7 +396,8 @@ class PDFPageInterpreterEx(PDFPageInterpreter):
         # )
         # for obj in page.contents:
         #     self.obj_patch[obj.objid] = ""
-        return f"q {ops_base} Q 1 0 0 1 {x0} {y0} cm"
+        return f"q {ops_base} Q {' '.join(f'{x:f}' for x in ctm_for_ops)} cm"
+        # return f"q {ops_base} Q 1 0 0 1 {x0} {y0} cm"
 
     def render_contents(
         self,

{babeldoc-0.3.43 → babeldoc-0.3.44}/docs/CONTRIBUTING.md

@@ -28,6 +28,12 @@ Please pay special attention to:
 - This project is not intended for direct end-user use, and the supported translators are mainly for debugging purposes. Unless it clearly helps with development and debugging, PRs for directly adding translators will not be accepted.
 - You can directly use [PDFMathTranslate](https://github.com/Byaidu/PDFMathTranslate) to get support for more translators.
 
+### **If you want to add new accelerator support for the layout model**
+
+- This project only plans to support various accelerators through onnxruntime. Please submit your accelerator support directly to onnxruntime.
+
+- Additionally, [translation_config.py](https://github.com/funstory-ai/BabelDOC/blob/9e5be3a05c15ecae98024ba695e4a2db1412c062/babeldoc/translation_config.py#L41) shows that the layout model implementation actually used in this project is passed in from outside. You can implement a layout model class according to the relevant interface, and then pass it through this parameter at runtime.
+
 ### **If you wish to contribute to BabelDOC**
 
 > [!TIP]
@@ -40,7 +46,7 @@ Please pay special attention to:
 > 
 > Due to the current high rate of code changes, this project only accepts small PRs. If you would like to suggest a change and you include a patch as a proof-of-concept, that would be great. However, please do not be offended if we rewrite your patch from scratch.
 >
-> Please discuss with the maintainers in the [Issue](https://github.com/funstory-ai/BabelDOC/issues) before sending a PR.
+> **This project cannot accept all PRs. We recommend that you discuss with the maintainers via [Issue](https://github.com/funstory-ai/BabelDOC/issues) before submitting a PR.**
 
 [//]: # (> We welcome pull requests and will review your contributions.)
 

{babeldoc-0.3.43 → babeldoc-0.3.44}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "BabelDOC"
-version = "0.3.43"
+version = "0.3.44"
 description = "Yet Another Document Translator"
 license = "AGPL-3.0"
 readme = "README.md"
@@ -150,7 +150,7 @@ pythonpath = [".", "src"]
 testpaths = ["tests"]
 
 [bumpver]
-current_version = "0.3.43"
+current_version = "0.3.44"
 version_pattern = "MAJOR.MINOR.PATCH[.PYTAGNUM]"
 
 [bumpver.file_patterns]

{babeldoc-0.3.43 → babeldoc-0.3.44}/uv.lock

@@ -48,7 +48,7 @@ wheels = [
 
 [[package]]
 name = "babeldoc"
-version = "0.3.43"
+version = "0.3.44"
 source = { editable = "." }
 dependencies = [
     { name = "bitstring" },

babeldoc-0.3.43/babeldoc/__init__.py

@@ -1 +0,0 @@
-__version__ = "0.3.43"