kernel.h@qq.com před 1 rokem
rodič
revize
4de8810c18

+ 1 - 1
magic_pdf/cli/magicpdf.py

@@ -132,7 +132,7 @@ def pdf_command(pdf, model, method):
     local_image_dir, _ = prepare_env()
     local_image_rw = DiskReaderWriter(local_image_dir)
     parse = get_pdf_parse_method(method)
-    parse(pdf_data, jso["doc_layout_result"], local_image_rw, is_debug=True)
+    parse(pdf_data, jso, local_image_rw, is_debug=True)
 
 
 if __name__ == "__main__":

+ 1 - 1
magic_pdf/libs/config_reader.py

@@ -17,7 +17,7 @@ def read_config():
     config_file = os.path.join(home_dir, "magic-pdf.json")
 
     if not os.path.exists(config_file):
-        raise Exception("magic-pdf.json not found")
+        raise Exception(f"{config_file} not found")
 
     with open(config_file, "r") as f:
         config = json.load(f)

+ 4 - 2
magic_pdf/libs/pdf_image_tools.py

@@ -1,11 +1,12 @@
 
+from magic_pdf.io import AbsReaderWriter
 from magic_pdf.libs.commons import fitz
 from loguru import logger
 from magic_pdf.libs.commons import join_path
 from magic_pdf.libs.hash_utils import compute_sha256
 
 
-def cut_image(bbox: tuple, page_num: int, page: fitz.Page, return_path, imageWriter):
+def cut_image(bbox: tuple, page_num: int, page: fitz.Page, return_path, imageWriter:AbsReaderWriter):
     """
     从第page_num页的page中,根据bbox进行裁剪出一张jpg图片,返回图片路径
     save_path:需要同时支持s3和本地, 图片存放在save_path下,文件名是: {page_num}_{bbox[0]}_{bbox[1]}_{bbox[2]}_{bbox[3]}.jpg , bbox内数字取整。
@@ -28,6 +29,7 @@ def cut_image(bbox: tuple, page_num: int, page: fitz.Page, return_path, imageWri
 
     byte_data = pix.tobytes(output='jpeg', jpg_quality=95)
 
+    imageWriter.write(byte_data, path=img_hash256_path, mode="binary")
     imageWriter.write(content=byte_data, path=img_hash256_path, mode="binary")
 
     return img_hash256_path
@@ -73,4 +75,4 @@ def save_images_by_bboxes(page_num: int, page: fitz.Page, pdf_bytes_md5: str,
         image_path = cut_image(bbox, page_num, page, return_path("tables"), imageWriter)
         table_info.append({"bbox": bbox, "image_path": image_path})
 
-    return image_info, image_backup_info, table_info, inline_eq_info, interline_eq_info
+    return image_info, image_backup_info, table_info, inline_eq_info, interline_eq_info