:gem: [Feature] MessageComposer: Merge or split messages for OpenAI API format

apis/chat_api.py

@@ -6,6 +6,7 @@ from conversations import (
     ConversationConnector,
     ConversationCreator,
     ConversationSession,
+    MessageComposer,
 )
 
 
@@ -113,37 +114,21 @@ class ChatAPIApp:
             default=[{"role": "user", "content": "Hello, who are you?"}],
             description="(list) Messages",
         )
-        sec_access_token: str = Field(
-            default="",
-            description="(str) Sec Access Token",
-        )
-        client_id: str = Field(
-            default="",
-            description="(str) Client ID",
-        )
-        conversation_id: str = Field(
-            default="",
-            description="(str) Conversation ID",
-        )
-        invocation_id: int = Field(
-            default=0,
-            description="(int) Invocation ID",
-        )
 
     def chat_completions(self, item: ChatCompletionsPostItem):
+        creator = ConversationCreator()
+        creator.create()
+
         connector = ConversationConnector(
             conversation_style=item.model,
-            sec_access_token=item.sec_access_token,
-            client_id=item.client_id,
-            conversation_id=item.conversation_id,
-            invocation_id=item.invocation_id,
+            sec_access_token=creator.sec_access_token,
+            client_id=creator.client_id,
+            conversation_id=creator.conversation_id,
+            invocation_id=0,
         )
 
-        if item.invocation_id == 0:
-            # TODO: History Messages Merger
-            prompt = item.messages[-1]["content"]
-        else:
-            prompt = item.messages[-1]["content"]
+        message_composer = MessageComposer()
+        prompt = message_composer.merge(item.messages)
 
         return EventSourceResponse(
             connector.stream_chat(prompt=prompt, yield_output=True),

conversations/__init__.py

@@ -1,3 +1,4 @@
 from .conversation_connector import ConversationConnector
 from .conversation_creator import ConversationCreator
 from .conversation_session import ConversationSession
+from .message_composer import MessageComposer

conversations/message_composer.py

@@ -0,0 +1,70 @@
+import re
+from pprint import pprint
+
+
+class MessageComposer:
+    def __init__(self):
+        pass
+
+    def merge(self, messages) -> str:
+        self.messages = messages
+        self.merged_str = ""
+        for message in messages:
+            role = message["role"]
+            if role.lower() in ["system", "user"]:
+                role_str = "me"
+            elif role.lower() in ["assistant", "bot"]:
+                role_str = "you"
+            else:
+                role_str = "unknown"
+            content = message["content"]
+            self.merged_str += f"`{role_str}`:\n{content}\n\n"
+        self.merged_str += "`you`:\n"
+        return self.merged_str
+
+    def split(self, merged_str) -> list:
+        self.messages = []
+        self.merged_str = merged_str
+        pattern = r"`(?P<role>me|you)`:\n(?P<content>.*)\n+"
+        matches = re.finditer(pattern, self.merged_str, re.MULTILINE)
+        for match in matches:
+            role = match.group("role")
+
+            if role == "me":
+                role_str = "user"
+            elif role == "you":
+                role_str = "assistant"
+            else:
+                role_str = "unknown"
+
+            self.messages.append(
+                {
+                    "role": role_str,
+                    "content": match.group("content"),
+                }
+            )
+        return self.messages
+
+
+if __name__ == "__main__":
+    composer = MessageComposer()
+    messages = [
+        {
+            "role": "system",
+            "content": "You are a LLM developed by OpenAI. Your name is GPT-4.",
+        },
+        {"role": "user", "content": "Hello, who are you?"},
+        {"role": "assistant", "content": "I am a bot."},
+        {"role": "user", "content": "What is your name?"},
+        {"role": "assistant", "content": "My name is Bing."},
+        {"role": "user", "content": "Tell me a joke."},
+        {"role": "assistant", "content": "What is a robot's favorite type of music?"},
+        {
+            "role": "user",
+            "content": "How many questions have I asked? Please list them.",
+        },
+    ]
+    merged_str = composer.merge(messages)
+    print(merged_str)
+    pprint(composer.split(merged_str))
+    print(composer.merge(composer.split(merged_str)))