Merge pull request #76 from dongcartney92/patch-1

get_mime_type bugfix
Filimoa · Nov 7, 2024 · 421b0ef · 421b0ef
2 parents 0986c20 + e997dd5
commit 421b0ef
Showing 1 changed file with 3 additions and 2 deletions.
diff --git a/src/openparse/text/pdfminer/core.py b/src/openparse/text/pdfminer/core.py
@@ -10,6 +10,7 @@
     LTTextContainer,
     LTTextLine,
 )
+from pdfminer.psparser import PSLiteral
 from pydantic import BaseModel, model_validator
 
 from openparse.pdf import Pdf
@@ -64,8 +65,8 @@ def _extract_chars(text_line: LTTextLine) -> List[CharElement]:
 
 
 def get_mime_type(pdf_object: LTImage) -> Optional[str]:
-    subtype = pdf_object.stream.attrs.get("Subtype", {"name": None}).name
-    filter_ = pdf_object.stream.attrs.get("Filter", {"name": None}).name
+    subtype = pdf_object.stream.attrs.get("Subtype", PSLiteral(None)).name
+    filter_ = pdf_object.stream.attrs.get("Filter", PSLiteral(None)).name
     if subtype == "Image":
         if filter_ == "DCTDecode":
             return "image/jpeg"