feat: set default memory messages limit to infinite (#5002)

api/core/memory/token_buffer_memory.py

@@ -1,3 +1,5 @@
+from typing import Optional
+
 from core.app.app_config.features.file_upload.manager import FileUploadConfigManager
 from core.file.message_file_parser import MessageFileParser
 from core.model_manager import ModelInstance
@@ -19,7 +21,7 @@ class TokenBufferMemory:
         self.model_instance = model_instance
 
     def get_history_prompt_messages(self, max_token_limit: int = 2000,
-                                    message_limit: int = 10) -> list[PromptMessage]:
+                                    message_limit: Optional[int] = None) -> list[PromptMessage]:
         """
         Get history prompt messages.
         :param max_token_limit: max token limit
@@ -28,10 +30,15 @@ class TokenBufferMemory:
         app_record = self.conversation.app
 
         # fetch limited messages, and return reversed
-        messages = db.session.query(Message).filter(
+        query = db.session.query(Message).filter(
             Message.conversation_id == self.conversation.id,
             Message.answer != ''
-        ).order_by(Message.created_at.desc()).limit(message_limit).all()
+        ).order_by(Message.created_at.desc())
+
+        if message_limit and message_limit > 0:
+            messages = query.limit(message_limit).all()
+        else:
+            messages = query.all()
 
         messages = list(reversed(messages))
         message_file_parser = MessageFileParser(
@@ -93,7 +100,7 @@ class TokenBufferMemory:
     def get_history_prompt_text(self, human_prefix: str = "Human",
                                 ai_prefix: str = "Assistant",
                                 max_token_limit: int = 2000,
-                                message_limit: int = 10) -> str:
+                                message_limit: Optional[int] = None) -> str:
         """
         Get history prompt text.
         :param human_prefix: human prefix

api/core/model_runtime/model_providers/openai/llm/gpt-3.5-turbo.yaml

@@ -9,7 +9,7 @@ features:
   - stream-tool-call
 model_properties:
   mode: chat
-  context_size: 4096
+  context_size: 16385
 parameter_rules:
   - name: temperature
     use_template: temperature

api/core/prompt/prompt_transform.py

@@ -77,7 +77,7 @@ class PromptTransform:
             max_token_limit=max_token_limit,
             message_limit=memory_config.window.size
             if (memory_config.window.enabled
-               and memory_config.window.size is not None
-               and memory_config.window.size > 0)
-            else 10
+                and memory_config.window.size is not None
+                and memory_config.window.size > 0)
+            else None
         )