Merge branch 'dev' into dev-table-model-update

magic_pdf/pdf_parse_by_ocr.py

+from magic_pdf.config.enums import SupportedPdfParseMethod
+from magic_pdf.data.dataset import PymuDocDataset
 from magic_pdf.pdf_parse_union_core_v2 import pdf_parse_union
 
 
@@ -8,10 +10,11 @@ def parse_pdf_by_ocr(pdf_bytes,
                      end_page_id=None,
                      debug_mode=False,
                      ):
-    return pdf_parse_union(pdf_bytes,
+    dataset = PymuDocDataset(pdf_bytes)
+    return pdf_parse_union(dataset,
                            model_list,
                            imageWriter,
-                           "ocr",
+                           SupportedPdfParseMethod.OCR,
                            start_page_id=start_page_id,
                            end_page_id=end_page_id,
                            debug_mode=debug_mode,

magic_pdf/pdf_parse_by_txt.py

+from magic_pdf.config.enums import SupportedPdfParseMethod
+from magic_pdf.data.dataset import PymuDocDataset
 from magic_pdf.pdf_parse_union_core_v2 import pdf_parse_union
 
 
@@ -9,10 +11,11 @@ def parse_pdf_by_txt(
     end_page_id=None,
     debug_mode=False,
 ):
-    return pdf_parse_union(pdf_bytes,
+    dataset = PymuDocDataset(pdf_bytes)
+    return pdf_parse_union(dataset,
                            model_list,
                            imageWriter,
-                           "txt",
+                           SupportedPdfParseMethod.TXT,
                            start_page_id=start_page_id,
                            end_page_id=end_page_id,
                            debug_mode=debug_mode,

magic_pdf/pipe/AbsPipe.py

@@ -17,7 +17,7 @@ class AbsPipe(ABC):
     PIP_TXT = "txt"
 
     def __init__(self, pdf_bytes: bytes, model_list: list, image_writer: AbsReaderWriter, is_debug: bool = False,
-                 start_page_id=0, end_page_id=None, lang=None):
+                 start_page_id=0, end_page_id=None, lang=None, layout_model=None, formula_enable=None, table_enable=None):
         self.pdf_bytes = pdf_bytes
         self.model_list = model_list
         self.image_writer = image_writer
@@ -26,6 +26,9 @@ class AbsPipe(ABC):
         self.start_page_id = start_page_id
         self.end_page_id = end_page_id
         self.lang = lang
+        self.layout_model = layout_model
+        self.formula_enable = formula_enable
+        self.table_enable = table_enable
     
     def get_compress_pdf_mid_data(self):
         return JsonCompressor.compress_json(self.pdf_mid_data)
@@ -95,9 +98,7 @@ class AbsPipe(ABC):
         """
         pdf_mid_data = JsonCompressor.decompress_json(compressed_pdf_mid_data)
         pdf_info_list = pdf_mid_data["pdf_info"]
-        parse_type = pdf_mid_data["_parse_type"]
-        lang = pdf_mid_data.get("_lang", None)
-        content_list = union_make(pdf_info_list, MakeMode.STANDARD_FORMAT, drop_mode, img_buket_path, parse_type, lang)
+        content_list = union_make(pdf_info_list, MakeMode.STANDARD_FORMAT, drop_mode, img_buket_path)
         return content_list
 
     @staticmethod
@@ -107,9 +108,7 @@ class AbsPipe(ABC):
         """
         pdf_mid_data = JsonCompressor.decompress_json(compressed_pdf_mid_data)
         pdf_info_list = pdf_mid_data["pdf_info"]
-        parse_type = pdf_mid_data["_parse_type"]
-        lang = pdf_mid_data.get("_lang", None)
-        md_content = union_make(pdf_info_list, md_make_mode, drop_mode, img_buket_path, parse_type, lang)
+        md_content = union_make(pdf_info_list, md_make_mode, drop_mode, img_buket_path)
         return md_content
 
 

magic_pdf/pipe/OCRPipe.py

@@ -10,8 +10,10 @@ from magic_pdf.user_api import parse_ocr_pdf
 class OCRPipe(AbsPipe):
 
     def __init__(self, pdf_bytes: bytes, model_list: list, image_writer: AbsReaderWriter, is_debug: bool = False,
-                 start_page_id=0, end_page_id=None, lang=None):
-        super().__init__(pdf_bytes, model_list, image_writer, is_debug, start_page_id, end_page_id, lang)
+                 start_page_id=0, end_page_id=None, lang=None,
+                 layout_model=None, formula_enable=None, table_enable=None):
+        super().__init__(pdf_bytes, model_list, image_writer, is_debug, start_page_id, end_page_id, lang,
+                         layout_model, formula_enable, table_enable)
 
     def pipe_classify(self):
         pass
@@ -19,12 +21,14 @@ class OCRPipe(AbsPipe):
     def pipe_analyze(self):
         self.model_list = doc_analyze(self.pdf_bytes, ocr=True,
                                       start_page_id=self.start_page_id, end_page_id=self.end_page_id,
-                                      lang=self.lang)
+                                      lang=self.lang, layout_model=self.layout_model,
+                                      formula_enable=self.formula_enable, table_enable=self.table_enable)
 
     def pipe_parse(self):
         self.pdf_mid_data = parse_ocr_pdf(self.pdf_bytes, self.model_list, self.image_writer, is_debug=self.is_debug,
                                           start_page_id=self.start_page_id, end_page_id=self.end_page_id,
-                                          lang=self.lang)
+                                          lang=self.lang, layout_model=self.layout_model,
+                                          formula_enable=self.formula_enable, table_enable=self.table_enable)
 
     def pipe_mk_uni_format(self, img_parent_path: str, drop_mode=DropMode.WHOLE_PDF):
         result = super().pipe_mk_uni_format(img_parent_path, drop_mode)

magic_pdf/pipe/TXTPipe.py

@@ -11,8 +11,10 @@ from magic_pdf.user_api import parse_txt_pdf
 class TXTPipe(AbsPipe):
 
     def __init__(self, pdf_bytes: bytes, model_list: list, image_writer: AbsReaderWriter, is_debug: bool = False,
-                 start_page_id=0, end_page_id=None, lang=None):
-        super().__init__(pdf_bytes, model_list, image_writer, is_debug, start_page_id, end_page_id, lang)
+                 start_page_id=0, end_page_id=None, lang=None,
+                 layout_model=None, formula_enable=None, table_enable=None):
+        super().__init__(pdf_bytes, model_list, image_writer, is_debug, start_page_id, end_page_id, lang,
+                         layout_model, formula_enable, table_enable)
 
     def pipe_classify(self):
         pass
@@ -20,12 +22,14 @@ class TXTPipe(AbsPipe):
     def pipe_analyze(self):
         self.model_list = doc_analyze(self.pdf_bytes, ocr=False,
                                       start_page_id=self.start_page_id, end_page_id=self.end_page_id,
-                                      lang=self.lang)
+                                      lang=self.lang, layout_model=self.layout_model,
+                                      formula_enable=self.formula_enable, table_enable=self.table_enable)
 
     def pipe_parse(self):
         self.pdf_mid_data = parse_txt_pdf(self.pdf_bytes, self.model_list, self.image_writer, is_debug=self.is_debug,
                                           start_page_id=self.start_page_id, end_page_id=self.end_page_id,
-                                          lang=self.lang)
+                                          lang=self.lang, layout_model=self.layout_model,
+                                          formula_enable=self.formula_enable, table_enable=self.table_enable)
 
     def pipe_mk_uni_format(self, img_parent_path: str, drop_mode=DropMode.WHOLE_PDF):
         result = super().pipe_mk_uni_format(img_parent_path, drop_mode)

magic_pdf/pipe/UNIPipe.py

@@ -14,9 +14,11 @@ from magic_pdf.user_api import parse_union_pdf, parse_ocr_pdf
 class UNIPipe(AbsPipe):
 
     def __init__(self, pdf_bytes: bytes, jso_useful_key: dict, image_writer: AbsReaderWriter, is_debug: bool = False,
-                 start_page_id=0, end_page_id=None, lang=None):
+                 start_page_id=0, end_page_id=None, lang=None,
+                 layout_model=None, formula_enable=None, table_enable=None):
         self.pdf_type = jso_useful_key["_pdf_type"]
-        super().__init__(pdf_bytes, jso_useful_key["model_list"], image_writer, is_debug, start_page_id, end_page_id, lang)
+        super().__init__(pdf_bytes, jso_useful_key["model_list"], image_writer, is_debug, start_page_id, end_page_id,
+                         lang, layout_model, formula_enable, table_enable)
         if len(self.model_list) == 0:
             self.input_model_is_empty = True
         else:
@@ -29,18 +31,21 @@ class UNIPipe(AbsPipe):
         if self.pdf_type == self.PIP_TXT:
             self.model_list = doc_analyze(self.pdf_bytes, ocr=False,
                                           start_page_id=self.start_page_id, end_page_id=self.end_page_id,
-                                          lang=self.lang)
+                                          lang=self.lang, layout_model=self.layout_model,
+                                          formula_enable=self.formula_enable, table_enable=self.table_enable)
         elif self.pdf_type == self.PIP_OCR:
             self.model_list = doc_analyze(self.pdf_bytes, ocr=True,
                                           start_page_id=self.start_page_id, end_page_id=self.end_page_id,
-                                          lang=self.lang)
+                                          lang=self.lang, layout_model=self.layout_model,
+                                          formula_enable=self.formula_enable, table_enable=self.table_enable)
 
     def pipe_parse(self):
         if self.pdf_type == self.PIP_TXT:
             self.pdf_mid_data = parse_union_pdf(self.pdf_bytes, self.model_list, self.image_writer,
                                                 is_debug=self.is_debug, input_model_is_empty=self.input_model_is_empty,
                                                 start_page_id=self.start_page_id, end_page_id=self.end_page_id,
-                                                lang=self.lang)
+                                                lang=self.lang, layout_model=self.layout_model,
+                                                formula_enable=self.formula_enable, table_enable=self.table_enable)
         elif self.pdf_type == self.PIP_OCR:
             self.pdf_mid_data = parse_ocr_pdf(self.pdf_bytes, self.model_list, self.image_writer,
                                               is_debug=self.is_debug,

magic_pdf/pre_proc/ocr_detect_all_bboxes.py

 from loguru import logger
 
 from magic_pdf.libs.boxbase import get_minbox_if_overlap_by_ratio, calculate_overlap_area_in_bbox1_area_ratio, \
-    calculate_iou
+    calculate_iou, calculate_vertical_projection_overlap_ratio
 from magic_pdf.libs.drop_tag import DropTag
 from magic_pdf.libs.ocr_content_type import BlockType
 from magic_pdf.pre_proc.remove_bbox_overlap import remove_overlap_between_bbox_for_block
@@ -60,29 +60,34 @@ def ocr_prepare_bboxes_for_layout_split(img_blocks, table_blocks, discarded_bloc
     return all_bboxes, all_discarded_blocks, drop_reasons
 
 
-def ocr_prepare_bboxes_for_layout_split_v2(img_blocks, table_blocks, discarded_blocks, text_blocks,
-                                        title_blocks, interline_equation_blocks, page_w, page_h):
-    all_bboxes = []
-    all_discarded_blocks = []
-    for image in img_blocks:
-        x0, y0, x1, y1 = image['bbox']
-        all_bboxes.append([x0, y0, x1, y1, None, None, None, BlockType.Image, None, None, None, None, image["score"]])
-
-    for table in table_blocks:
-        x0, y0, x1, y1 = table['bbox']
-        all_bboxes.append([x0, y0, x1, y1, None, None, None, BlockType.Table, None, None, None, None, table["score"]])
+def add_bboxes(blocks, block_type, bboxes):
+    for block in blocks:
+        x0, y0, x1, y1 = block['bbox']
+        if block_type in [
+            BlockType.ImageBody, BlockType.ImageCaption, BlockType.ImageFootnote,
+            BlockType.TableBody, BlockType.TableCaption, BlockType.TableFootnote
+        ]:
+            bboxes.append([x0, y0, x1, y1, None, None, None, block_type, None, None, None, None, block["score"], block["group_id"]])
+        else:
+            bboxes.append([x0, y0, x1, y1, None, None, None, block_type, None, None, None, None, block["score"]])
 
-    for text in text_blocks:
-        x0, y0, x1, y1 = text['bbox']
-        all_bboxes.append([x0, y0, x1, y1, None, None, None, BlockType.Text, None, None, None, None, text["score"]])
 
-    for title in title_blocks:
-        x0, y0, x1, y1 = title['bbox']
-        all_bboxes.append([x0, y0, x1, y1, None, None, None, BlockType.Title, None, None, None, None, title["score"]])
+def ocr_prepare_bboxes_for_layout_split_v2(
+        img_body_blocks, img_caption_blocks, img_footnote_blocks,
+        table_body_blocks, table_caption_blocks, table_footnote_blocks,
+        discarded_blocks, text_blocks, title_blocks, interline_equation_blocks, page_w, page_h
+):
+    all_bboxes = []
 
-    for interline_equation in interline_equation_blocks:
-        x0, y0, x1, y1 = interline_equation['bbox']
-        all_bboxes.append([x0, y0, x1, y1, None, None, None, BlockType.InterlineEquation, None, None, None, None, interline_equation["score"]])
+    add_bboxes(img_body_blocks, BlockType.ImageBody, all_bboxes)
+    add_bboxes(img_caption_blocks, BlockType.ImageCaption, all_bboxes)
+    add_bboxes(img_footnote_blocks, BlockType.ImageFootnote, all_bboxes)
+    add_bboxes(table_body_blocks, BlockType.TableBody, all_bboxes)
+    add_bboxes(table_caption_blocks, BlockType.TableCaption, all_bboxes)
+    add_bboxes(table_footnote_blocks, BlockType.TableFootnote, all_bboxes)
+    add_bboxes(text_blocks, BlockType.Text, all_bboxes)
+    add_bboxes(title_blocks, BlockType.Title, all_bboxes)
+    add_bboxes(interline_equation_blocks, BlockType.InterlineEquation, all_bboxes)
 
     '''block嵌套问题解决'''
     '''文本框与标题框重叠，优先信任文本框'''
@@ -96,13 +101,23 @@ def ocr_prepare_bboxes_for_layout_split_v2(img_blocks, table_blocks, discarded_b
     '''interline_equation框被包含在文本类型框内，且interline_equation比文本区块小很多时信任文本框，这时需要舍弃公式框'''
     # 通过后续大框套小框逻辑删除
 
-    '''discarded_blocks中只保留宽度超过1/3页面宽度的，高度超过10的，处于页面下半50%区域的（限定footnote）'''
+    '''discarded_blocks'''
+    all_discarded_blocks = []
+    add_bboxes(discarded_blocks, BlockType.Discarded, all_discarded_blocks)
+
+    '''footnote识别：宽度超过1/3页面宽度的，高度超过10的，处于页面下半50%区域的'''
+    footnote_blocks = []
     for discarded in discarded_blocks:
         x0, y0, x1, y1 = discarded['bbox']
-        all_discarded_blocks.append([x0, y0, x1, y1, None, None, None, BlockType.Discarded, None, None, None, None, discarded["score"]])
-        # 将footnote加入到all_bboxes中，用来计算layout
-        # if (x1 - x0) > (page_w / 3) and (y1 - y0) > 10 and y0 > (page_h / 2):
-        #     all_bboxes.append([x0, y0, x1, y1, None, None, None, BlockType.Footnote, None, None, None, None, discarded["score"]])
+        if (x1 - x0) > (page_w / 3) and (y1 - y0) > 10 and y0 > (page_h / 2):
+            footnote_blocks.append([x0, y0, x1, y1])
+
+    '''移除在footnote下面的任何框'''
+    need_remove_blocks = find_blocks_under_footnote(all_bboxes, footnote_blocks)
+    if len(need_remove_blocks) > 0:
+        for block in need_remove_blocks:
+            all_bboxes.remove(block)
+            all_discarded_blocks.append(block)
 
     '''经过以上处理后，还存在大框套小框的情况，则删除小框'''
     all_bboxes = remove_overlaps_min_blocks(all_bboxes)
@@ -113,6 +128,20 @@ def ocr_prepare_bboxes_for_layout_split_v2(img_blocks, table_blocks, discarded_b
     return all_bboxes, all_discarded_blocks
 
 
+def find_blocks_under_footnote(all_bboxes, footnote_blocks):
+    need_remove_blocks = []
+    for block in all_bboxes:
+        block_x0, block_y0, block_x1, block_y1 = block[:4]
+        for footnote_bbox in footnote_blocks:
+            footnote_x0, footnote_y0, footnote_x1, footnote_y1 = footnote_bbox
+            # 如果footnote的纵向投影覆盖了block的纵向投影的80%且block的y0大于等于footnote的y1
+            if block_y0 >= footnote_y1 and calculate_vertical_projection_overlap_ratio((block_x0, block_y0, block_x1, block_y1), footnote_bbox) >= 0.8:
+                if block not in need_remove_blocks:
+                    need_remove_blocks.append(block)
+                    break
+    return need_remove_blocks
+
+
 def fix_interline_equation_overlap_text_blocks_with_hi_iou(all_bboxes):
     # 先提取所有text和interline block
     text_blocks = []

magic_pdf/pre_proc/ocr_dict_merge.py

@@ -49,7 +49,7 @@ def merge_spans_to_line(spans):
                 continue
 
             # 如果当前的span与当前行的最后一个span在y轴上重叠，则添加到当前行
-            if __is_overlaps_y_exceeds_threshold(span['bbox'], current_line[-1]['bbox'], 0.6):
+            if __is_overlaps_y_exceeds_threshold(span['bbox'], current_line[-1]['bbox'], 0.5):
                 current_line.append(span)
             else:
                 # 否则，开始新行
@@ -153,6 +153,11 @@ def fill_spans_in_blocks(blocks, spans, radio):
             'type': block_type,
             'bbox': block_bbox,
         }
+        if block_type in [
+            BlockType.ImageBody, BlockType.ImageCaption, BlockType.ImageFootnote,
+            BlockType.TableBody, BlockType.TableCaption, BlockType.TableFootnote
+        ]:
+            block_dict["group_id"] = block[-1]
         block_spans = []
         for span in spans:
             span_bbox = span['bbox']
@@ -201,6 +206,27 @@ def fix_block_spans(block_with_spans, img_blocks, table_blocks):
     return fix_blocks
 
 
+def fix_block_spans_v2(block_with_spans):
+    """1、img_block和table_block因为包含caption和footnote的关系，存在block的嵌套关系
+    需要将caption和footnote的text_span放入相应img_block和table_block内的
+    caption_block和footnote_block中 2、同时需要删除block中的spans字段."""
+    fix_blocks = []
+    for block in block_with_spans:
+        block_type = block['type']
+
+        if block_type in [BlockType.Text, BlockType.Title,
+                          BlockType.ImageCaption, BlockType.ImageFootnote,
+                          BlockType.TableCaption, BlockType.TableFootnote
+                          ]:
+            block = fix_text_block(block)
+        elif block_type in [BlockType.InterlineEquation, BlockType.ImageBody, BlockType.TableBody]:
+            block = fix_interline_block(block)
+        else:
+            continue
+        fix_blocks.append(block)
+    return fix_blocks
+
+
 def fix_discarded_block(discarded_block_with_spans):
     fix_discarded_blocks = []
     for block in discarded_block_with_spans:

magic_pdf/resources/model_config/model_configs.yaml

-config:
-  device: cpu
-  layout: True
-  formula: True
-  table_config:
-    model: TableMaster
-    is_table_recog_enable: False
-    max_time: 400
-
 weights:
-  layout: Layout/model_final.pth
-  mfd: MFD/weights.pt
-  mfr: MFR/unimernet_small
+  layoutlmv3: Layout/LayoutLMv3/model_final.pth
+  doclayout_yolo: Layout/YOLO/doclayout_yolo_ft.pt
+  yolo_v8_mfd: MFD/YOLO/yolo_v8_ft.pt
+  unimernet_small: MFR/unimernet_small
   struct_eqtable: TabRec/StructEqTable
-  TableMaster: TabRec/TableMaster
\ No newline at end of file
+  tablemaster: TabRec/TableMaster
\ No newline at end of file

magic_pdf/tools/cli.py

@@ -52,7 +52,7 @@ without method specified, auto will be used by default.""",
     help="""
     Input the languages in the pdf (if known) to improve OCR accuracy.  Optional.
     You should input "Abbreviation" with language form url:
-    https://paddlepaddle.github.io/PaddleOCR/en/ppocr/blog/multi_languages.html#5-support-languages-and-abbreviations
+    https://paddlepaddle.github.io/PaddleOCR/latest/en/ppocr/blog/multi_languages.html#5-support-languages-and-abbreviations
     """,
     default=None,
 )

magic_pdf/tools/common.py

@@ -6,8 +6,8 @@ import click
 from loguru import logger
 
 import magic_pdf.model as model_config
-from magic_pdf.libs.draw_bbox import (draw_layout_bbox, draw_span_bbox,
-                                      draw_model_bbox, draw_line_sort_bbox)
+from magic_pdf.libs.draw_bbox import (draw_layout_bbox, draw_line_sort_bbox,
+                                      draw_model_bbox, draw_span_bbox)
 from magic_pdf.libs.MakeContentConfig import DropMode, MakeMode
 from magic_pdf.pipe.OCRPipe import OCRPipe
 from magic_pdf.pipe.TXTPipe import TXTPipe
@@ -46,10 +46,12 @@ def do_parse(
     start_page_id=0,
     end_page_id=None,
     lang=None,
+    layout_model=None,
+    formula_enable=None,
+    table_enable=None,
 ):
     if debug_able:
         logger.warning('debug mode is on')
-        # f_dump_content_list = True
         f_draw_model_bbox = True
         f_draw_line_sort_bbox = True
 
@@ -64,13 +66,16 @@ def do_parse(
     if parse_method == 'auto':
         jso_useful_key = {'_pdf_type': '', 'model_list': model_list}
         pipe = UNIPipe(pdf_bytes, jso_useful_key, image_writer, is_debug=True,
-                       start_page_id=start_page_id, end_page_id=end_page_id, lang=lang)
+                       start_page_id=start_page_id, end_page_id=end_page_id, lang=lang,
+                       layout_model=layout_model, formula_enable=formula_enable, table_enable=table_enable)
     elif parse_method == 'txt':
         pipe = TXTPipe(pdf_bytes, model_list, image_writer, is_debug=True,
-                       start_page_id=start_page_id, end_page_id=end_page_id, lang=lang)
+                       start_page_id=start_page_id, end_page_id=end_page_id, lang=lang,
+                       layout_model=layout_model, formula_enable=formula_enable, table_enable=table_enable)
     elif parse_method == 'ocr':
         pipe = OCRPipe(pdf_bytes, model_list, image_writer, is_debug=True,
-                       start_page_id=start_page_id, end_page_id=end_page_id, lang=lang)
+                       start_page_id=start_page_id, end_page_id=end_page_id, lang=lang,
+                       layout_model=layout_model, formula_enable=formula_enable, table_enable=table_enable)
     else:
         logger.error('unknown parse method')
         exit(1)

magic_pdf/user_api.py

@@ -101,11 +101,19 @@ def parse_union_pdf(pdf_bytes: bytes, pdf_models: list, imageWriter: AbsReaderWr
     if pdf_info_dict is None or pdf_info_dict.get("_need_drop", False):
         logger.warning(f"parse_pdf_by_txt drop or error, switch to parse_pdf_by_ocr")
         if input_model_is_empty:
-            pdf_models = doc_analyze(pdf_bytes,
-                                     ocr=True,
-                                     start_page_id=start_page_id,
-                                     end_page_id=end_page_id,
-                                     lang=lang)
+            layout_model = kwargs.get("layout_model", None)
+            formula_enable = kwargs.get("formula_enable", None)
+            table_enable = kwargs.get("table_enable", None)
+            pdf_models = doc_analyze(
+                pdf_bytes,
+                ocr=True,
+                start_page_id=start_page_id,
+                end_page_id=end_page_id,
+                lang=lang,
+                layout_model=layout_model,
+                formula_enable=formula_enable,
+                table_enable=table_enable,
+            )
         pdf_info_dict = parse_pdf(parse_pdf_by_ocr)
         if pdf_info_dict is None:
             raise Exception("Both parse_pdf_by_txt and parse_pdf_by_ocr failed.")

magic_pdf/utils/annotations.py

+
+from loguru import logger
+
+
+def ImportPIL(f):
+    try:
+        import PIL  # noqa: F401
+    except ImportError:
+        logger.error('Pillow not installed, please install by pip.')
+        exit(1)
+    return f

next_docs/en/api.rst

+Data Api
+------------------
+
+.. toctree::
+   :maxdepth: 2
+
+   api/dataset.rst
+   api/data_reader_writer.rst
+   api/read_api.rst

next_docs/en/api/data_reader_writer.rst

+
+Data Reader Writer
+--------------------
+
+.. autoclass:: magic_pdf.data.data_reader_writer.DataReader
+   :members:
+   :inherited-members:
+
+.. autoclass:: magic_pdf.data.data_reader_writer.DataWriter
+   :members:
+   :inherited-members:
+
+.. autoclass:: magic_pdf.data.data_reader_writer.S3DataReader
+   :members:
+   :inherited-members:
+
+.. autoclass:: magic_pdf.data.data_reader_writer.S3DataWriter
+   :members:
+   :inherited-members:
+
+.. autoclass:: magic_pdf.data.data_reader_writer.FileBasedDataReader
+   :members:
+   :inherited-members:
+
+.. autoclass:: magic_pdf.data.data_reader_writer.FileBasedDataWriter
+   :members:
+   :inherited-members:
+
+.. autoclass:: magic_pdf.data.data_reader_writer.S3DataReader
+   :members:
+   :inherited-members:
+
+.. autoclass:: magic_pdf.data.data_reader_writer.S3DataWriter
+   :members:
+   :inherited-members:
+
+.. autoclass:: magic_pdf.data.data_reader_writer.MultiBucketS3DataReader
+   :members:
+   :inherited-members:
+
+.. autoclass:: magic_pdf.data.data_reader_writer.MultiBucketS3DataWriter
+   :members:
+   :inherited-members:
+