feat: add DocConverter for legacy .doc files using unword parser

packages/markitdown/src/markitdown/_markitdown.py

@@ -29,6 +29,7 @@
     BingSerpConverter,
     PdfConverter,
     DocxConverter,
+    DocConverter,
     XlsxConverter,
     XlsConverter,
     PptxConverter,
@@ -192,6 +193,7 @@ def enable_builtins(self, **kwargs) -> None:
             self.register_converter(YouTubeConverter())
             self.register_converter(BingSerpConverter())
             self.register_converter(DocxConverter())
+            self.register_converter(DocConverter())
             self.register_converter(XlsxConverter())
             self.register_converter(XlsConverter())
             self.register_converter(PptxConverter())

packages/markitdown/src/markitdown/converters/__init__.py

@@ -11,6 +11,7 @@
 from ._bing_serp_converter import BingSerpConverter
 from ._pdf_converter import PdfConverter
 from ._docx_converter import DocxConverter
+from ._doc_converter import DocConverter
 from ._xlsx_converter import XlsxConverter, XlsConverter
 from ._pptx_converter import PptxConverter
 from ._image_converter import ImageConverter
@@ -34,6 +35,7 @@
     "BingSerpConverter",
     "PdfConverter",
     "DocxConverter",
+    "DocConverter",
     "XlsxConverter",
     "XlsConverter",
     "PptxConverter",

packages/markitdown/src/markitdown/converters/_doc_converter.py

@@ -0,0 +1,73 @@
+import sys
+from typing import Any, BinaryIO
+
+from .._stream_info import StreamInfo
+from .._base_converter import DocumentConverter, DocumentConverterResult
+from .._exceptions import MissingDependencyException, MISSING_DEPENDENCY_MESSAGE, FileConversionException
+
+#using unword dependency
+_dependency_exc_info = None
+unword = None
+try:
+    import unword as _unword
+    unword = _unword
+except ImportError:
+    _dependency_exc_info = sys.exc_info()
+
+ACCEPTED_MIME_TYPE_PREFIXES = [
+    "application/msword", "application/x-msword"
+]
+
+ACCEPTED_FILE_EXTENSIONS = [".doc"]
+
+
+class DocConverter(DocumentConverter):
+    """
+    Converts DOC (Word 97-2003) files to Markdown. Uses unword package 
+    as parser backend to extract body text with heading levels,
+    page breaks, and textbox contents.
+    """
+
+    def accepts(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,
+    ) -> bool:
+        mimetype = (stream_info.mimetype or "").lower()
+        extension = (stream_info.extension or "").lower()
+
+        if extension in ACCEPTED_FILE_EXTENSIONS:
+            return True
+
+        for prefix in ACCEPTED_MIME_TYPE_PREFIXES:
+            if mimetype.startswith(prefix):
+                return True
+
+        return False
+
+    def convert(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> DocumentConverterResult:
+        # Check: the dependencies
+        if _dependency_exc_info is not None:
+            raise MissingDependencyException(
+                MISSING_DEPENDENCY_MESSAGE.format(
+                    converter=type(self).__name__,
+                    extension=".doc",
+                    feature="doc",
+                )
+            ) from _dependency_exc_info[1].with_traceback(  # type: ignore[union-attr]
+                _dependency_exc_info[2]
+            )
+
+        try:
+            doc = unword.parse_doc(file_stream.read())
+        except Exception as e:
+            raise FileConversionException(f"Failed to parse .doc file: {e}") from e
+
+        title = getattr(doc, "title", None) or getattr(doc, "metadata", {}).get("title")
+        return DocumentConverterResult(markdown=doc.body_text.strip(), title=title)

packages/markitdown/src/markitdown/converters/_pptx_converter.py

@@ -4,6 +4,7 @@
 import io
 import re
 import html
+import hashlib
 
 from typing import BinaryIO, Any
 from operator import attrgetter
@@ -140,16 +141,38 @@ def get_shape_content(shape, **kwargs):
                     alt_text = re.sub(r"[\r\n\[\]]", " ", alt_text)
                     alt_text = re.sub(r"\s+", " ", alt_text).strip()
 
+                    output_dir = kwargs.get("output_dir")
+
                     # If keep_data_uris is True, use base64 encoding for images
                     if kwargs.get("keep_data_uris", False):
                         blob = shape.image.blob
                         content_type = shape.image.content_type or "image/png"
                         b64_string = base64.b64encode(blob).decode("utf-8")
                         md_content += f"\n![{alt_text}](data:{content_type};base64,{b64_string})\n"
                     else:
-                        # A placeholder name
-                        filename = re.sub(r"\W", "", shape.name) + ".jpg"
+                        #save image to disk to reference
+                        blob = shape.image.blob
+                        content_type = shape.image.content_type or "image/png"
+                        ext_map = {"jpeg": "jpg", "svg+xml": "svg"}
+                        raw_ext = content_type.split("/")[-1]
+                        ext = ext_map.get(raw_ext, raw_ext)
+
+                        #add filename collision handling
+                        suffix = hashlib.md5(blob).hexdigest()[:8]
+                        safe_name = re.sub(r"\W", "", shape.name) if shape.name else "image"
+                        filename = f"{safe_name}_{suffix}.{ext}"
+
+                        output_dir = kwargs.get("output_dir", ".")
+                        image_path = os.path.join(output_dir, filename)
+
+                        try:
+                            with open(image_path, "wb") as img_file:
+                                img_file.write(blob)
+                        except OSError as e:
+                            raise OSError(f"Failed to write image to '{image_path}'.") from e
+
                         md_content += "\n![" + alt_text + "](" + filename + ")\n"
+
 
                 # Tables
                 if self._is_table(shape):

packages/markitdown/tests/_test_vectors.py

@@ -31,6 +31,22 @@ class FileTestVector(object):
             "data:image/png;base64,iVBORw0KGgoAAAANSU",
         ],
     ),
+    FileTestVector(
+        filename="test.doc",
+        mimetype="application/msword",
+        charset=None,
+        url=None,
+        must_include=[
+            "93d437af-bc31-492f-a7fc-3cbc9b7c1710",
+            "fd0ed3e3-6373-4446-815c-4b979f6063a9",
+            "#Test for converting .doc files to MD format",
+            "Let’s test it!",
+        ],
+        must_not_include=[
+            "d0cf11e0",
+            "\x00"
+        ],
+    ),
     FileTestVector(
         filename="test.xlsx",
         mimetype="application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
@@ -68,7 +84,7 @@ class FileTestVector(object):
             "AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation",
             "a3f6004b-6f4f-4ea8-bee3-3741f4dc385f",  # chart title
             "2003",  # chart value
-            "![This phrase of the caption is Human-written.](Picture4.jpg)",
+            "![This phrase of the caption is Human-written.](Picture4_",
         ],
         must_not_include=["data:image/jpeg;base64,/9j/4AAQSkZJRgABAQE"],
     ),