Merge pull request #195 from funstory-ai/pdf2zh-v2-rc

fix: Fix form ctm matrix

Author: awwaawwa

babeldoc/__init__.py

@@ -1 +1 @@
-__version__ = "0.2.19"
+__version__ = "0.2.20"

babeldoc/const.py

@@ -3,7 +3,7 @@
 import subprocess
 from pathlib import Path
 
-__version__ = "0.2.19"
+__version__ = "0.2.20"
 
 CACHE_FOLDER = Path.home() / ".cache" / "babeldoc"
 

babeldoc/document_il/backend/pdf_creater.py

@@ -673,12 +673,13 @@ def write(self, translation_config: TranslationConfig) -> TranslateResult:
                 }
                 page_op = BitStream()
                 # q {ops_base}Q 1 0 0 1 {x0} {y0} cm {ops_new}
-                page_op.append(b"q ")
+                # page_op.append(b"q ")
                 page_op.append(page.base_operations.value.encode())
-                page_op.append(b" Q ")
-                page_op.append(
-                    f"q Q 1 0 0 1 {page.cropbox.box.x} {page.cropbox.box.y} cm \n".encode(),
-                )
+                page_op.append(b" \n")
+                # page_op.append(b" Q ")
+                # page_op.append(
+                #     f"q Q 1 0 0 1 {page.cropbox.box.x} {page.cropbox.box.y} cm \n".encode(),
+                # )
                 # 收集所有字符
                 chars = []
                 # 首先添加页面级别的字符

babeldoc/main.py

@@ -23,7 +23,7 @@
 from babeldoc.translation_config import WatermarkOutputMode
 
 logger = logging.getLogger(__name__)
-__version__ = "0.2.19"
+__version__ = "0.2.20"
 
 
 def create_parser():

babeldoc/pdfinterp.py

@@ -290,7 +290,7 @@ def do_Do(self, xobjid_arg: PDFStackT) -> None:
             self.scs = interpreter.scs
             self.il_creater.on_xobj_end(
                 x_id,
-                f"q {ops_base}Q {a} {b} {c} {d} {e} {f} cm ",
+                f"q {ops_base} Q {a} {b} {c} {d} {e} {f} cm ",
             )
             try:  # 有的时候 form 字体加不上这里会烂掉
                 self.device.fontid = interpreter.fontid
@@ -345,7 +345,7 @@ def process_page(self, page: PDFPage) -> None:
         # )
         # for obj in page.contents:
         #     self.obj_patch[obj.objid] = ""
-        return ops_base
+        return f"q {ops_base} Q 1 0 0 1 {x0} {y0} cm"
 
     def render_contents(
         self,

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "BabelDOC"
-version = "0.2.19"
+version = "0.2.20"
 description = "Yet Another Document Translator"
 license = "AGPL-3.0"
 readme = "README.md"
@@ -148,7 +148,7 @@ pythonpath = [".", "src"]
 testpaths = ["tests"]
 
 [bumpver]
-current_version = "0.2.19"
+current_version = "0.2.20"
 version_pattern = "MAJOR.MINOR.PATCH[.PYTAGNUM]"
 
 [bumpver.file_patterns]