From a07a18715041a26a4602bbcee6d853c526e4ddbe Mon Sep 17 00:00:00 2001
From: Maxim Lysak <mly@zurich.ibm.com>
Date: Wed, 16 Oct 2024 12:22:19 +0200
Subject: [PATCH] Added and fixed origin for msword and mspowerpoint backend

Signed-off-by: Maxim Lysak <mly@zurich.ibm.com>
---
 docling/backend/mspowerpoint_backend.py |  6 +++++-
 docling/backend/msword_backend.py       | 19 ++++++++++++++++++-
 2 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/docling/backend/mspowerpoint_backend.py b/docling/backend/mspowerpoint_backend.py
index fba3e31d..91f1d11f 100644
--- a/docling/backend/mspowerpoint_backend.py
+++ b/docling/backend/mspowerpoint_backend.py
@@ -91,8 +91,12 @@ class MsPowerpointDocumentBackend(DeclarativeDocumentBackend, PaginatedDocumentB
             mimetype="application/vnd.ms-powerpoint",
             binary_hash=self.document_hash,
         )
+        if len(fname) > 0:
+            docname = Path(fname).stem
+        else:
+            docname = "stream"
         doc = DoclingDocument(
-            description=DescriptionItem(), name="name_without_extension", origin=origin
+            description=DescriptionItem(), name=docname, origin=origin
         )  # must add origin information
         doc = self.walk_linear(self.pptx_obj, doc)
 
diff --git a/docling/backend/msword_backend.py b/docling/backend/msword_backend.py
index 182e31a9..0de32682 100644
--- a/docling/backend/msword_backend.py
+++ b/docling/backend/msword_backend.py
@@ -8,6 +8,7 @@ from docling_core.types.doc import (
     DescriptionItem,
     DocItemLabel,
     DoclingDocument,
+    DocumentOrigin,
     GroupLabel,
     TableCell,
     TableData,
@@ -84,7 +85,23 @@ class MsWordDocumentBackend(DeclarativeDocumentBackend):
 
     def convert(self) -> DoclingDocument:
         # Parses the DOCX into a structured document model.
-        doc = DoclingDocument(description=DescriptionItem(), name="dummy")
+
+        fname = ""
+        if isinstance(self.path_or_stream, Path):
+            fname = self.path_or_stream.name
+
+        origin = DocumentOrigin(
+            filename=fname,
+            mimetype="application/vnd.openxmlformats-officedocument.wordprocessingml.document",
+            binary_hash=self.document_hash,
+        )
+        if len(fname) > 0:
+            docname = Path(fname).stem
+        else:
+            docname = "stream"
+        doc = DoclingDocument(
+            description=DescriptionItem(), name=docname, origin=origin
+        )
         if self.is_valid():
             assert self.docx_obj is not None
             doc = self.walk_linear(self.docx_obj.element.body, self.docx_obj, doc)