Refactor AI chat model for live streaming and precise responses

This commit introduces several changes to the AI chat model in the `aicodebot/cli.py` file. The main changes include:

- The introduction of two new temperature settings: `PRECISE_TEMPERATURE` and `CREATIVE_TEMPERATURE`. These settings allow for more control over the randomness of the AI's responses.
- The use of the `rich.live.Live` class for live streaming of the AI's responses. This is implemented in the `alignment`, `debug`, `fun_fact`, `review`, and `sidekick` commands.
- The creation of a `RichLiveCallbackHandler` class that updates the live stream with each new token generated by the AI.
- Adjustments to the `max_tokens` parameter in the `fun_fact` and `sidekick` commands.

These changes aim to improve the user experience by providing real-time feedback from the AI and allowing for more precise responses. 🤖💬

Author: TechNickAI

aicodebot/cli.py

@@ -2,29 +2,32 @@
 from aicodebot.agents import get_agent
 from aicodebot.helpers import exec_and_get_output, get_token_length, git_diff_context
 from dotenv import load_dotenv
+from langchain.callbacks.base import BaseCallbackHandler
 from langchain.chains import LLMChain
 from langchain.chat_models import ChatOpenAI
 from langchain.prompts import load_prompt
 from openai.api_resources import engine
 from pathlib import Path
 from rich.console import Console
+from rich.live import Live
 from rich.markdown import Markdown
 from rich.style import Style
 import click, datetime, openai, os, random, subprocess, sys, tempfile, webbrowser
 
 # ----------------------------- Default settings ----------------------------- #
 
 DEFAULT_MAX_TOKENS = 512
-DEFAULT_TEMPERATURE = 0.1
+PRECISE_TEMPERATURE = 0
+CREATIVE_TEMPERATURE = 0.7
 DEFAULT_SPINNER = "point"
 
 # ----------------------- Setup for rich console output ---------------------- #
+
 console = Console()
 bot_style = Style(color="#30D5C8")
 error_style = Style(color="#FF0000")
 warning_style = Style(color="#FFA500")
 
-
 # -------------------------- Top level command group ------------------------- #
 
 
@@ -55,14 +58,21 @@ def alignment(verbose):
 
     # Set up the language model
     model = get_llm_model(get_token_length(prompt.template))
-    llm = ChatOpenAI(model=model, temperature=DEFAULT_TEMPERATURE, max_tokens=DEFAULT_MAX_TOKENS, verbose=verbose)
 
-    # Set up the chain
-    chain = LLMChain(llm=llm, prompt=prompt, verbose=verbose)
+    with Live(Markdown(""), auto_refresh=True) as live:
+        llm = ChatOpenAI(
+            model=model,
+            temperature=CREATIVE_TEMPERATURE,
+            max_tokens=DEFAULT_MAX_TOKENS,
+            verbose=verbose,
+            streaming=True,
+            callbacks=[RichLiveCallbackHandler(live)],
+        )
 
-    with console.status("Generating an inspirational message", spinner=DEFAULT_SPINNER):
-        response = chain.run({})
-        console.print(Markdown(response), style=bot_style)
+        # Set up the chain
+        chain = LLMChain(llm=llm, prompt=prompt, verbose=verbose)
+
+        chain.run({})
 
 
 @cli.command()
@@ -109,13 +119,13 @@ def commit(verbose, response_token_size, yes, skip_pre_commit):
         console.print(f"Diff context token size: {request_token_size}, using model: {model}")
 
     # Set up the language model
-    llm = ChatOpenAI(model=model, temperature=DEFAULT_TEMPERATURE, max_tokens=DEFAULT_MAX_TOKENS, verbose=verbose)
+    llm = ChatOpenAI(model=model, temperature=PRECISE_TEMPERATURE, max_tokens=DEFAULT_MAX_TOKENS, verbose=verbose)
 
     # Set up the chain
     chain = LLMChain(llm=llm, prompt=prompt, verbose=verbose)
 
     console.print("The following files will be committed:\n" + files)
-    with console.status("Generating the commit message", spinner=DEFAULT_SPINNER):
+    with console.status("Examining the diff and generating the commit message", spinner=DEFAULT_SPINNER):
         response = chain.run(diff_context)
 
     # Write the commit message to a temporary file
@@ -168,14 +178,20 @@ def debug(command, verbose):
 
     # Set up the language model
     model = get_llm_model(get_token_length(error_output) + get_token_length(prompt.template))
-    llm = ChatOpenAI(model=model, temperature=DEFAULT_TEMPERATURE, max_tokens=DEFAULT_MAX_TOKENS, verbose=verbose)
 
-    # Set up the chain
-    chat_chain = LLMChain(llm=llm, prompt=prompt, verbose=verbose)
+    with Live(Markdown(""), auto_refresh=True) as live:
+        llm = ChatOpenAI(
+            model=model,
+            temperature=PRECISE_TEMPERATURE,
+            max_tokens=DEFAULT_MAX_TOKENS,
+            verbose=verbose,
+            streaming=True,
+            callbacks=[RichLiveCallbackHandler(live)],
+        )
 
-    with console.status("Debugging", spinner=DEFAULT_SPINNER):
-        response = chat_chain.run(error_output)
-        console.print(Markdown(response), style=bot_style)
+        # Set up the chain
+        chain = LLMChain(llm=llm, prompt=prompt, verbose=verbose)
+        chain.run(error_output)
 
     sys.exit(process.returncode)
 
@@ -191,16 +207,22 @@ def fun_fact(verbose):
 
     # Set up the language model
     model = get_llm_model(get_token_length(prompt.template))
-    llm = ChatOpenAI(model=model, temperature=0.9, max_tokens=250, verbose=verbose)
 
-    # Set up the chain
-    chat_chain = LLMChain(llm=llm, prompt=prompt, verbose=verbose)
+    with Live(Markdown(""), auto_refresh=True) as live:
+        llm = ChatOpenAI(
+            model=model,
+            temperature=PRECISE_TEMPERATURE,
+            max_tokens=DEFAULT_MAX_TOKENS / 2,
+            verbose=verbose,
+            streaming=True,
+            callbacks=[RichLiveCallbackHandler(live)],
+        )
+
+        # Set up the chain
+        chain = LLMChain(llm=llm, prompt=prompt, verbose=verbose)
 
-    with console.status("Fetching a fun fact", spinner=DEFAULT_SPINNER):
-        # Select a random year so that we get a different answer each time
         year = random.randint(1942, datetime.datetime.utcnow().year)
-        response = chat_chain.run(f"programming and artificial intelligence in the year {year}")
-        console.print(Markdown(response), style=bot_style)
+        chain.run(f"programming and artificial intelligence in the year {year}")
 
 
 @cli.command
@@ -225,15 +247,20 @@ def review(commit, verbose):
     if verbose:
         console.print(f"Diff context token size: {request_token_size}, using model: {model}")
 
-    # Set up the language model
-    llm = ChatOpenAI(model=model, temperature=DEFAULT_TEMPERATURE, max_tokens=response_token_size, verbose=verbose)
+    with Live(Markdown(""), auto_refresh=True) as live:
+        llm = ChatOpenAI(
+            model=model,
+            temperature=PRECISE_TEMPERATURE,
+            max_tokens=response_token_size,
+            verbose=verbose,
+            streaming=True,
+            callbacks=[RichLiveCallbackHandler(live)],
+        )
 
-    # Set up the chain
-    chain = LLMChain(llm=llm, prompt=prompt, verbose=verbose)
+        # Set up the chain
+        chain = LLMChain(llm=llm, prompt=prompt, verbose=verbose)
 
-    with console.status("Reviewing code", spinner=DEFAULT_SPINNER):
-        response = chain.run(diff_context)
-        console.print(Markdown(response), style=bot_style)
+        chain.run(diff_context)
 
 
 @cli.command
@@ -254,7 +281,7 @@ def sidekick(task, verbose):
     setup_environment()
 
     model = get_llm_model()
-    llm = ChatOpenAI(model=model, temperature=DEFAULT_TEMPERATURE, max_tokens=3500, verbose=verbose)
+    llm = ChatOpenAI(model=model, temperature=PRECISE_TEMPERATURE, max_tokens=2000, verbose=verbose)
 
     agent = get_agent("sidekick", llm, verbose)
 
@@ -358,5 +385,16 @@ def get_llm_model(token_size=0):
             raise click.ClickException("🛑 The context is too large to for the Model. 😞")
 
 
+class RichLiveCallbackHandler(BaseCallbackHandler):
+    buffer = []
+
+    def __init__(self, live):
+        self.live = live
+
+    def on_llm_new_token(self, token, **kwargs):
+        self.buffer.append(token)
+        self.live.update(Markdown("".join(self.buffer)))
+
+
 if __name__ == "__main__":
     cli()