fix: mem-reader bug (#255)

* fix: mem-reader bug

* fix: test mem reader

Author: CaralHsi

src/memos/mem_reader/simple_struct.py

@@ -1,6 +1,7 @@
 import concurrent.futures
 import copy
 import json
+import os
 import re
 
 from abc import ABC
@@ -19,11 +20,12 @@
     SIMPLE_STRUCT_DOC_READER_PROMPT,
     SIMPLE_STRUCT_DOC_READER_PROMPT_ZH,
     SIMPLE_STRUCT_MEM_READER_EXAMPLE,
-    SIMPLE_STRUCT_MEM_READER_PROMPT_ZH,
-    SIMPLE_STRUCT_MEM_READER_PROMPT,
     SIMPLE_STRUCT_MEM_READER_EXAMPLE_ZH,
+    SIMPLE_STRUCT_MEM_READER_PROMPT,
+    SIMPLE_STRUCT_MEM_READER_PROMPT_ZH,
 )
 
+
 logger = log.get_logger(__name__)
 PROMPT_DICT = {
     "chat": {
@@ -207,7 +209,7 @@ def get_scene_data_info(self, scene_data: list, type: str) -> list[str]:
         elif type == "doc":
             for item in scene_data:
                 try:
-                    if not isinstance(item, str):
+                    if os.path.exists(item):
                         parsed_text = parser.parse(item)
                         results.append({"file": "pure_text", "text": parsed_text})
                     else:

tests/mem_reader/test_simple_structure.py

@@ -124,11 +124,13 @@ def test_get_scene_data_info_with_doc(self, mock_parser_factory):
         parser_instance.parse.return_value = "Parsed document text.\n"
         mock_parser_factory.from_config.return_value = parser_instance
 
-        scene_data = [{"fake_file_like": "should trigger parse"}]
-        result = self.reader.get_scene_data_info(scene_data, type="doc")
+        scene_data = ["/fake/path/to/doc.txt"]
+        with patch("os.path.exists", return_value=True):
+            result = self.reader.get_scene_data_info(scene_data, type="doc")
 
         self.assertIsInstance(result, list)
         self.assertEqual(result[0]["text"], "Parsed document text.\n")
+        parser_instance.parse.assert_called_once_with("/fake/path/to/doc.txt")
 
     def test_parse_json_result_success(self):
         """Test successful JSON parsing."""