Matches prefix to verify presence of DOCX,PPTX,XLSX files instead of standard file names

Author: srisudarsan

test_unstructured/file_utils/test_document_from_office365.docx

@@ -30,6 +30,7 @@
 
 is_in_docker = os.path.exists("/.dockerenv")
 
+
 # ================================================================================================
 # STRATEGY #1 - DIRECT DETECTION OF CFB/ZIP-BASED BINARY FILE TYPES (8 TYPES)
 # ================================================================================================
@@ -987,3 +988,9 @@ def test_json_content_type_is_disambiguated_for_ndjson():
     file_buffer.name = "filename.pdf"
     predicted_type = detect_filetype(file=file_buffer, content_type="application/json")
     assert predicted_type == FileType.NDJSON
+
+
+def test_office_files_when_document_archive_has_non_standard_prefix():
+
+    predicted_type = detect_filetype(file_path="test_document_from_office365.docx")
+    assert predicted_type == FileType.DOCX

test_unstructured/file_utils/test_filetype.py

@@ -747,13 +747,13 @@ def _file_type(self) -> FileType | None:
 
             filenames = zip.namelist()
 
-            if "word/document.xml" in filenames:
+            if any(re.match(r"word/document.*\.xml$", filename) for filename in filenames):
                 return FileType.DOCX
 
-            if "xl/workbook.xml" in filenames:
+            if any(re.match(r"xl/workbook.*\.xml$", filename) for filename in filenames):
                 return FileType.XLSX
 
-            if "ppt/presentation.xml" in filenames:
+            if any(re.match(r"ppt/presentation.*\.xml$", filename) for filename in filenames):
                 return FileType.PPTX
 
             # -- ODT and EPUB files place their MIME-type in `mimetype` in the archive root --