Adding conditions

chatsky/llm/conditions.py

@@ -6,9 +6,20 @@
 """
 
 from chatsky.script.core.message import Message
+from chatsky.script import Context
+from chatsky.pipeline import Pipeline
+import re
 
-def regex_search(pattern: str) -> bool:
-    pass
 
-def semantic_distance(target: str | Message, threshold: float) -> bool:
-    pass
+def regex_search(pattern: str):
+    def _(ctx: Context, _: Pipeline) -> bool:
+        return bool(re.search(pattern, ctx.last_request.text))
+
+    return _
+
+
+def semantic_distance(target: str | Message, threshold: float):
+    def _(ctx: Context, _: Pipeline) -> bool:
+        pass
+
+    return _

chatsky/llm/filters.py

@@ -0,0 +1,15 @@
+"""
+Filters.
+---------
+This module contains a collection of basic functions for history filtering to avoid cluttering LLMs context window.
+"""
+
+from chatsky.script.core.message import Message
+
+def is_important(msg: Message) -> bool:
+    if msg.misc["important"]:
+        return True
+    return False
+
+def from_the_model(msg: Message, model_name: str) -> bool:
+    return msg.annotation.__generated_by_model__ == model_name

chatsky/llm/wrapper.py

@@ -96,7 +96,7 @@ def llm_response(
         model_name,
         prompt="",
         history=5,
-        filter: Callable=None
+        filter_func: Callable=None
     ):
     """
     Basic function for receiving LLM responses.
@@ -105,28 +105,16 @@ def llm_response(
     :param model_name: Name of the model from the `Pipeline.models` dictionary.
     :param prompt: Prompt for the model.
     :param history: Number of messages to keep in history.
-    :param filter: filter function to filter messages that will go the models context.
+    :param filter_func: filter function to filter messages that will go the models context.
     """
     model = pipeline.get(model_name)
     history_messages = []
     if history == 0:
         return model.respond([prompt + "\n" + ctx.last_request.text])
     else:
-        for req, resp in filter(lambda x: filter(x[0], x[1]), zip(ctx.requests[-history:], ctx.responses[-history:])):
-            if req.attachments != []:
-                content = [{"type": "text", "text": prompt + "\n" + ctx.last_request.text}]
-                for image in ctx.last_request.attachments:
-                    if image is not Image:
-                        continue
-                    content.append(
-                        {"type": "image_url", "image_url": {"url": __attachment_to_content(image)}}
-                    )
-                req_message = HumanMessage(content=content)
-            else:
-                req_message = HumanMessage(req.text)
-
-            history_messages.append(req_message)
-            history_messages.append(SystemMessage(resp.text))
+        for req, resp in filter(lambda x: filter_func(x), zip(ctx.requests[-history:], ctx.responses[-history:])):
+            history_messages.append(message_to_langchain(req))
+            history_messages.append(message_to_langchain(resp, human=False))
         return model.respond(history_messages)
 
 
@@ -160,7 +148,7 @@ def __attachment_to_content(attachment: Image) -> str:
     return image_b64
 
 
-def message_to_langchain(message: Message):
+def message_to_langchain(message: Message, human=True):
     if message.attachments != []:
         content = [{"type": "text", "text": message.text}]
         for image in message.attachments:
@@ -169,4 +157,7 @@ def message_to_langchain(message: Message):
             content.append(
                 {"type": "image_url", "image_url": {"url": __attachment_to_content(image)}}
             )
-    return HumanMessage(content=content)
+    if human:
+        return HumanMessage(content=content)
+    else:
+        return SystemMessage(content=content)