yomitoku 0.9.3__py3-none-any.whl → 0.9.4__py3-none-any.whl

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/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/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

{yomitoku-0.9.3.dist-info → yomitoku-0.9.4.dist-info}/METADATA

@@ -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.dist-info → yomitoku-0.9.4.dist-info}/RECORD

@@ -10,7 +10,7 @@ yomitoku/table_structure_recognizer.py,sha256=tHjex6deT_FjRK5ePz9bUXA_QIhgv_vYtK
 yomitoku/text_detector.py,sha256=6IwEJJKp_F8YH0Oki0QV-Mqi--P2LGbNKo-_kxBB_eo,4383
 yomitoku/text_recognizer.py,sha256=eaxozNu-Ms6iv8efbKZzn8pJNW1Wo4f86bGhzSMtv3s,5992
 yomitoku/cli/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-yomitoku/cli/main.py,sha256=5An9usBfBYqNiBA6QqZTCaYI4b3W1j-efAsggK_HCss,13522
+yomitoku/cli/main.py,sha256=v1UYsnQdnylhLvDURuxLODU3IU-ssVGqOJT9r-TCVns,13623
 yomitoku/cli/mcp_server.py,sha256=WnWzxd13HaemC3b-5i9B9NVBGc3WGfum2nYhoBolEnk,5641
 yomitoku/configs/__init__.py,sha256=x5-ccjGiP6xxRtDPT7f1Enl7SsE0hSk0G8f7eF9V85I,886
 yomitoku/configs/cfg_layout_parser_rtdtrv2.py,sha256=8PRxB2Ar9UF7-DLtbgSokhrzdXb0veWI6Wc-X8qigRw,2329
@@ -25,7 +25,7 @@ yomitoku/data/__init__.py,sha256=KAofFc9rk9ZdTKBjemu9RM8Vj9XnKbWC2MPZ2RWtOdE,82
 yomitoku/data/dataset.py,sha256=lpBcpkMuQzRIyLJ4_mqtuhR9s2ZmzgBgc-XYuE_b2Sc,1326
 yomitoku/data/functions.py,sha256=RExCUxI3-gccIMw-H0ribX2jeGKkrJWhS4fNn_12c3Y,7878
 yomitoku/export/__init__.py,sha256=gmlikMHRXfzfJ_8q4fyDlnpGms-x1oggQOwJEWHMgBU,508
-yomitoku/export/export_csv.py,sha256=VY8mntUCPDbDco_dyvq5O0_Q4wga9_GTyjHCS-y4UiQ,3399
+yomitoku/export/export_csv.py,sha256=4U4KQ2RcBQmyUZ9O7a4uLoB6RUw80HPL1EEJUDwQlcI,4044
 yomitoku/export/export_html.py,sha256=LQDyZgbzmI0qJ0-FEK-54r9816H3L9hD10ChMcw0KyA,5620
 yomitoku/export/export_json.py,sha256=iNG37tdIuYG2x3NiiZemKaB6-X45WrhVPZhbX7RUzRI,2410
 yomitoku/export/export_markdown.py,sha256=KrdxDmKzVP_LbTKuDNGGsT31QOPKVsNNlb6wtLEW-1Q,4705
@@ -51,9 +51,9 @@ yomitoku/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 yomitoku/utils/graph.py,sha256=LKNB8ZhSQwOZMfeAimPMF5UCVVr2ZaUWoGDkz8z-uGU,456
 yomitoku/utils/logger.py,sha256=uOmtQDr0A0JD7wyFshedL08BiNrQorHnpktRXba8bjU,424
 yomitoku/utils/misc.py,sha256=r92x45kQR8lC5jO1MZaHBDtcCWBkQXg_WS9H4RXJzSY,4127
-yomitoku/utils/searchable_pdf.py,sha256=7JQCFhwpBJVV1Fx9q4p6fFGlEsJ-SmR0arddI3NzEeo,3567
+yomitoku/utils/searchable_pdf.py,sha256=taZ-XtXN4RItePMDv4q0fRVlryusdkexA3TCXzwlXRo,3497
 yomitoku/utils/visualizer.py,sha256=DjDwHiAu1iFRKh96H3Egq4vuI2s_-9dLCDeykhKi8jo,5251
-yomitoku-0.9.3.dist-info/METADATA,sha256=0r3tOl0ohoegcYQXWM3ROCSOr5px3IK-0zwqyADc9Mc,8872
-yomitoku-0.9.3.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-yomitoku-0.9.3.dist-info/entry_points.txt,sha256=n3c8bQSj5Be5GHAOv_NZ8cldJFmWeigQxSmteFTmu_k,96
-yomitoku-0.9.3.dist-info/RECORD,,
+yomitoku-0.9.4.dist-info/METADATA,sha256=oDIp-lxMIQjIfVtrzQXBcY2PJFHlRwktVGFXndQRJZo,8872
+yomitoku-0.9.4.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+yomitoku-0.9.4.dist-info/entry_points.txt,sha256=n3c8bQSj5Be5GHAOv_NZ8cldJFmWeigQxSmteFTmu_k,96
+yomitoku-0.9.4.dist-info/RECORD,,