From: Trenton Holmes <holmes.trenton@gmail.com>
Date: Wed, 14 Sep 2022 14:48:12 +0000 (-0700)
Subject: Ensure the tika parse function gets a string, not a PathLike
X-Git-Tag: v1.9.0-beta.rc3~7^2
X-Git-Url: http://git.ipfire.org/gitweb/gitweb.cgi?a=commitdiff_plain;h=refs%2Fpull%2F1591%2Fhead;p=thirdparty%2Fpaperless-ngx.git

Ensure the tika parse function gets a string, not a PathLike
---

diff --git a/src/paperless_tika/parsers.py b/src/paperless_tika/parsers.py
index e706e3aa5f..1cfb1eecbb 100644
--- a/src/paperless_tika/parsers.py
+++ b/src/paperless_tika/parsers.py
@@ -1,4 +1,5 @@
 import os
+from pathlib import Path
 
 import dateutil.parser
 import requests
@@ -28,6 +29,11 @@ class TikaDocumentParser(DocumentParser):
 
     def extract_metadata(self, document_path, mime_type):
         tika_server = settings.TIKA_ENDPOINT
+
+        # tika does not support a PathLike, only strings
+        # ensure this is a string
+        document_path = str(document_path)
+
         try:
             parsed = parser.from_file(document_path, tika_server)
         except Exception as e:
@@ -47,10 +53,14 @@ class TikaDocumentParser(DocumentParser):
             for key in parsed["metadata"]
         ]
 
-    def parse(self, document_path, mime_type, file_name=None):
+    def parse(self, document_path: Path, mime_type, file_name=None):
         self.log("info", f"Sending {document_path} to Tika server")
         tika_server = settings.TIKA_ENDPOINT
 
+        # tika does not support a PathLike, only strings
+        # ensure this is a string
+        document_path = str(document_path)
+
         try:
             parsed = parser.from_file(document_path, tika_server)
         except Exception as err: