Add reference-based Word and PowerPoint transformations

application/single_app/config.py

@@ -101,7 +101,7 @@
 EXECUTOR_TYPE = 'thread'
 EXECUTOR_MAX_WORKERS = 30
 SESSION_TYPE = 'filesystem'
-VERSION = "0.261.322"
+VERSION = "0.261.323"
 IS_DEVELOPMENT = is_development_env_enabled()
 
 # Opt-out for deployments where App Service Easy Auth is active but the platform
@@ -195,7 +195,7 @@ def _split_env_list(raw_value, lowercase=False):
 
 # Base allowed extensions (always available)
 BASE_ALLOWED_EXTENSIONS = {'txt', 'doc', 'docm', 'html', 'md', 'json', 'xml', 'yaml', 'yml', 'log'}
-DOCUMENT_EXTENSIONS = {'pdf', 'docx', 'pptx', 'ppt'}
+DOCUMENT_EXTENSIONS = {'pdf', 'docx', 'dotx', 'pptx', 'potx', 'ppt'}
 TABULAR_EXTENSIONS = {'csv', 'xlsx', 'xls', 'xlsm'}
 SCHEMA_EXTENSIONS = {'xsd'}
 VISIO_EXTENSIONS = {'vsdx'}

application/single_app/content_screening/publication.py

@@ -21,6 +21,7 @@
 )
 from content_screening.exports import build_clean_artifact, structured_sheets, table_schema_text
 from content_screening.extraction import heartbeat_publication, is_publication, publication_chunks
+from functions_document_transform_sources import OFFICE_REFERENCE_MIME_TYPES, TEMPLATE_EXTENSIONS, native_template_pages
 
 
 def _document_helpers():
@@ -185,7 +186,8 @@
     else:
         file_name = document.get("file_name") or scan["original_file_name"]
         content = storage.read_bytes(scan["source_ref"], subject)
-        content_type = mimetypes.guess_type(file_name)[0] or "application/octet-stream"
+        extension = "." + file_name.rsplit(".", 1)[-1].lower()
+        content_type = OFFICE_REFERENCE_MIME_TYPES.get(extension) or mimetypes.guess_type(file_name)[0] or "application/octet-stream"
     metadata = _metadata_from_units(units, file_name)
     settings = helpers.get_settings()
     max_characters = min(
@@ -199,7 +201,14 @@
     else:
         chunks = publication_chunks(units, max_characters)
     if not chunks:
-        raise ScreeningValidationError("No reviewed content is available to publish.")
+        extension = "." + file_name.rsplit(".", 1)[-1].lower()
+        if extension not in TEMPLATE_EXTENSIONS or scan.get("sanitized") or not scan.get("source_ref"):
+            raise ScreeningValidationError("No reviewed content is available to publish.")
+        # An empty native template can be released, but only after the normal review and
+        # original-byte binding. Never fabricate a searchable narrative to clear the hold.
+        pages, _supplemental, _structure = native_template_pages(content, extension)
+        if any(page["content"].strip() for page in pages):
+            raise ScreeningValidationError("Reviewed template content is missing.")
     scope_args = {
         "group_id": subject.scope_id if subject.scope_type == "group" else None,
         "public_workspace_id": subject.scope_id if subject.scope_type == "public" else None,

application/single_app/functions_chat_upload_extraction.py

@@ -146,6 +146,15 @@
     details = build_workspace_upload_details(document, len(ordered_chunks))
     body = indexed_text or render_vision_analysis_text(details["vision_analysis"])
 
+    if not body:
+        source = document if isinstance(document, dict) else {}
+        native_template_ready = (
+            str(source.get("file_name") or filename).lower().endswith((".dotx", ".potx"))
+            and isinstance(source.get("office_template_structure"), dict)
+            and not is_workspace_upload_failed(source) and not is_workspace_upload_processing(source)
+        )
+        if native_template_ready:
+            body = "Native Office template retained for structural inspection. No narrative content was indexed."
     if not body:
         source = document if isinstance(document, dict) else {}
         if is_workspace_upload_failed(source):