Update OpenAI ChatGPT 3.5 turbo token limit

Author: darrenburns

.pre-commit-config.yaml

@@ -31,15 +31,3 @@ repos:
     hooks:
       - id: ruff
         args: [ --fix, --exit-non-zero-on-fix ]
-  - repo: https://github.com/psf/black
-    rev: 23.3.0
-    hooks:
-      - id: black
-  - repo: local
-    hooks:
-      - id: mypy
-        name: mypy
-        entry: 'mypy'
-        language: system
-        files: elia_chat
-        types: [python, file]

.python-version

@@ -0,0 +1 @@
+3.11

elia_chat/database/models.py

@@ -1,20 +1,18 @@
-from __future__ import annotations
-
 import pathlib
 from datetime import datetime
-from typing import Any
+from typing import Any, Optional
 
 from sqlalchemy import Column, DateTime, func, JSON, desc
 from sqlalchemy.orm import selectinload
-from sqlmodel import SQLModel, Field, create_engine, Session, select, Relationship
+from sqlmodel import Field, Relationship, Session, SQLModel, create_engine, select
 
 
 class MessageDao(SQLModel, table=True):
     __tablename__ = "message"
 
     id: int = Field(default=None, primary_key=True)
-    chat_id: int | None = Field(foreign_key="chat.id")
-    chat: ChatDao = Relationship(back_populates="messages")
+    chat_id: Optional[int] = Field(foreign_key="chat.id")
+    chat: "ChatDao" = Relationship(back_populates="messages")
     role: str
     content: str
     timestamp: datetime | None = Field(
@@ -23,7 +21,7 @@ class MessageDao(SQLModel, table=True):
     status: str | None
     end_turn: bool | None
     weight: float | None
-    meta: dict = Field(sa_column=Column(JSON), default={})
+    meta: dict[Any, Any] = Field(sa_column=Column(JSON), default={})
     recipient: str | None
 
 
@@ -39,7 +37,7 @@ class ChatDao(SQLModel, table=True):
     messages: list[MessageDao] = Relationship(back_populates="chat")
 
     @staticmethod
-    def all() -> list[ChatDao]:
+    def all() -> list["ChatDao"]:
         with Session(engine) as session:
             # Create a subquery that finds the maximum
             # (most recent) timestamp for each chat.
@@ -60,7 +58,7 @@ def all() -> list[ChatDao]:
             return list(results)
 
     @staticmethod
-    def from_id(chat_id: str) -> ChatDao:
+    def from_id(chat_id: str) -> "ChatDao":
         with Session(engine) as session:
             statement = (
                 select(ChatDao)

elia_chat/widgets/chat.py

@@ -27,9 +27,9 @@
 from elia_chat.widgets.chat_header import ChatHeader
 from elia_chat.widgets.chat_options import (
     DEFAULT_MODEL,
-    MODEL_MAPPING,
     ChatOptions,
     GPTModel,
+    get_model_by_name,
 )
 from elia_chat.widgets.chatbox import Chatbox
 
@@ -182,7 +182,7 @@ async def stream_agent_response(self) -> None:
         log.debug(
             f"Creating streaming response with model {self.chat_data.model_name!r}"
         )
-        selected_model: GPTModel = MODEL_MAPPING[self.chat_data.model_name]
+        selected_model: GPTModel = get_model_by_name(self.chat_data.model_name)
         llm: BaseChatModel = selected_model.model
         trimmed_messages = self.trim_messages(
             model=llm,

elia_chat/widgets/chat_options.py

@@ -4,7 +4,7 @@
 from typing import Dict
 
 from langchain.callbacks import AsyncIteratorCallbackHandler
-from langchain.chat_models import ChatOpenAI
+from langchain_openai import ChatOpenAI
 from langchain.chat_models.base import BaseChatModel
 from langchain.llms.base import LLM
 from rich.console import RenderableType
@@ -42,11 +42,11 @@ class GPTModel:
     description="The fastest ChatGPT model, great for most everyday tasks.",
     css_class="gpt35",
     model=ChatOpenAI(
-        model_name="gpt-3.5-turbo",
+        model="gpt-3.5-turbo",
         streaming=True,
         callbacks=[callback],
     ),
-    token_limit=4096,
+    token_limit=16385,
 )
 AVAILABLE_MODELS = [
     DEFAULT_MODEL,
@@ -59,7 +59,7 @@ class GPTModel:
         "complex tasks which require advanced reasoning.",
         css_class="gpt4",
         model=ChatOpenAI(
-            model_name="gpt-4-1106-preview",
+            model="gpt-4-turbo",
             streaming=True,
             callbacks=[callback],
         ),
@@ -69,6 +69,11 @@ class GPTModel:
 MODEL_MAPPING: Dict[str, GPTModel] = {model.name: model for model in AVAILABLE_MODELS}
 
 
+def get_model_by_name(model_name: str) -> GPTModel:
+    """Given the name of a model as a string, return the GPTModel."""
+    return MODEL_MAPPING[model_name]
+
+
 class ModelPanel(Static):
     class Selected(Message):
         def __init__(self, model: GPTModel):
@@ -172,7 +177,7 @@ def compose(self) -> ComposeResult:
                 model_set.focus()
                 for index, model in enumerate(AVAILABLE_MODELS):
                     model_panel = ModelPanel(
-                        model, id=model.name, classes=model.css_class
+                        model, id=model.css_class, classes=model.css_class
                     )
                     if index == 0:
                         model_panel.selected = True