yomitoku 0.9.3__tar.gz → 0.9.4__tar.gz

{yomitoku-0.9.3 → yomitoku-0.9.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: yomitoku
-Version: 0.9.3
+Version: 0.9.4
 Summary: Yomitoku is an AI-powered document image analysis package designed specifically for the Japanese language.
 Author-email: Kotaro Kinoshita <kotaro.kinoshita@mlism.com>
 License: CC BY-NC-SA 4.0

{yomitoku-0.9.3 → yomitoku-0.9.4}/src/yomitoku/cli/main.py

@@ -158,6 +158,7 @@ def process_single_file(args, analyzer, path, format):
                     args.ignore_line_break,
                     img,
                     args.figure,
+                    args.figure_letter,
                     args.figure_dir,
                 )
             else:
@@ -167,6 +168,7 @@ def process_single_file(args, analyzer, path, format):
                     encoding=args.encoding,
                     img=img,
                     export_figure=args.figure,
+                    export_figure_letter=args.figure_letter,
                     figure_dir=args.figure_dir,
                 )
 

{yomitoku-0.9.3 → yomitoku-0.9.4}/src/yomitoku/export/export_csv.py

@@ -63,6 +63,7 @@ def convert_csv(
     ignore_line_break,
     img=None,
     export_figure: bool = True,
+    export_figure_letter: bool = False,
     figure_dir="figures",
 ):
     elements = []
@@ -89,6 +90,20 @@ def convert_csv(
             }
         )
 
+    if export_figure_letter:
+        for figure in inputs.figures:
+            paragraphs = sorted(figure.paragraphs, key=lambda x: x.order)
+            for paragraph in paragraphs:
+                contents = paragraph_to_csv(paragraph, ignore_line_break)
+                elements.append(
+                    {
+                        "type": "paragraph",
+                        "box": paragraph.box,
+                        "element": contents,
+                        "order": figure.order,
+                    }
+                )
+
     elements = sorted(elements, key=lambda x: x["order"])
 
     if export_figure:
@@ -109,6 +124,7 @@ def export_csv(
     encoding: str = "utf-8",
     img=None,
     export_figure: bool = True,
+    export_figure_letter: bool = False,
     figure_dir="figures",
 ):
     elements = convert_csv(
@@ -117,6 +133,7 @@ def export_csv(
         ignore_line_break,
         img,
         export_figure,
+        export_figure_letter,
         figure_dir,
     )
 

{yomitoku-0.9.3 → yomitoku-0.9.4}/src/yomitoku/utils/searchable_pdf.py

@@ -72,8 +72,6 @@ def create_searchable_pdf(images, ocr_results, output_path, font_path=None):
 
     for i, (image, ocr_result) in enumerate(zip(images, ocr_results)):
         image = Image.fromarray(image[:, :, ::-1])  # Convert BGR to RGB
-        pdfmetrics.registerFont(TTFont("MPLUS1p-Medium", FONT_PATH))
-
         image_path = f"tmp_{i}.png"
         image.save(image_path)
         w, h = image.size