add: n_messages on /api/ask to set the maximum number of messages to …

…include

CHANGELOG.md

@@ -5,6 +5,10 @@
     *Andrea Sponziello* 
 ### **Copyrigth**: *Tiledesk SRL*
 
+## [2024-07-29]
+### 0.2.9
+- add: n_messages on /api/ask to set the maximum number of messages to include 
+
 ## [2024-07-27]
 ### 0.2.8
 - add: history on /api/ask 

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "tilellm"
-version = "0.2.8"
+version = "0.2.9"
 description = "tiledesk for RAG"
 authors = ["Gianluca Lorenzo <[email protected]>"]
 repository = "https://github.com/Tiledesk/tiledesk-llm"

tilellm/controller/controller.py

@@ -199,7 +199,7 @@ async def ask_to_llm(question, chat_model=None):
         qa_prompt = ChatPromptTemplate.from_messages(
             [
                 ("system", question.system_context),
-                MessagesPlaceholder("chat_history_a"),
+                MessagesPlaceholder("chat_history_a", n_messages=question.n_messages),
                 ("human", "{input}"),
             ]
         )

tilellm/models/item_model.py

@@ -124,6 +124,7 @@ class QuestionToLLM(BaseModel):
     debug: bool = Field(default_factory=lambda: False)
     system_context: str = Field(default="You are a helpful AI bot. Always reply in the same language of the question.")
     chat_history_dict: Optional[Dict[str, ChatEntry]] = None
+    n_messages: int = Field(default_factory=lambda: None)
 
     @field_validator("temperature")
     def temperature_range(cls, v):
@@ -132,6 +133,13 @@ def temperature_range(cls, v):
             raise ValueError("Temperature must be between 0.0 and 1.0.")
         return v
 
+    @field_validator("n_messages")
+    def n_messages_range(cls, v):
+        """Ensures n_messages is within greater than 0"""
+        if not v > 0:
+            raise ValueError("n_messages must be greater than 0")
+        return v
+
     @field_validator("max_tokens")
     def max_tokens_range(cls, v):
         """Ensures max_tokens is a positive integer."""