324 lines
9.5 KiB
Python
324 lines
9.5 KiB
Python
from flask import Flask, render_template, request
|
|
from flask_socketio import SocketIO, emit, join_room
|
|
|
|
import eventlet
|
|
|
|
import openai
|
|
|
|
from openai.error import InvalidRequestError
|
|
|
|
import os
|
|
|
|
import time
|
|
|
|
import boto3
|
|
import json
|
|
|
|
|
|
app = Flask(__name__)
|
|
app.config["SECRET_KEY"] = "your_secret_key"
|
|
socketio = SocketIO(app, async_mode="eventlet")
|
|
|
|
|
|
from flask_sqlalchemy import SQLAlchemy
|
|
|
|
app.config["SQLALCHEMY_DATABASE_URI"] = "sqlite:///chat.db"
|
|
app.config["SQLALCHEMY_TRACK_MODIFICATIONS"] = False
|
|
db = SQLAlchemy(app)
|
|
|
|
|
|
# Create an argument parser for aws profile.
|
|
import argparse
|
|
|
|
parser = argparse.ArgumentParser()
|
|
parser.add_argument("--profile", help="AWS profile name")
|
|
args = parser.parse_args()
|
|
profile_name = args.profile
|
|
|
|
|
|
class Message(db.Model):
|
|
id = db.Column(db.Integer, primary_key=True)
|
|
username = db.Column(db.String(128), nullable=False)
|
|
content = db.Column(db.String(1024), nullable=False)
|
|
room = db.Column(db.String(128), nullable=False)
|
|
|
|
def __init__(self, username, content, room):
|
|
self.username = username
|
|
self.content = content
|
|
self.room = room
|
|
|
|
|
|
# Create the database and tables
|
|
with app.app_context():
|
|
db.create_all()
|
|
|
|
|
|
@app.route("/")
|
|
def index():
|
|
return render_template("index.html")
|
|
|
|
|
|
@app.route("/chat/<room>")
|
|
def chat(room):
|
|
return render_template("chat.html", room=room)
|
|
|
|
|
|
@socketio.on("join")
|
|
def on_join(data):
|
|
room = data["room"]
|
|
join_room(room)
|
|
|
|
# Fetch previous messages from the database
|
|
previous_messages = Message.query.filter_by(room=room).all()
|
|
|
|
# Send the history of messages only to the newly connected client.
|
|
# The reason for using `request.sid` here is to target the specific session (or client) that
|
|
# just connected, so only they receive the backlog of messages, rather than broadcasting
|
|
# this information to all clients in the room.
|
|
for message in previous_messages:
|
|
emit(
|
|
"previous_messages",
|
|
{
|
|
"id": message.id,
|
|
"username": message.username,
|
|
"message": message.content,
|
|
},
|
|
room=request.sid,
|
|
)
|
|
|
|
# Broadcast to all clients in the room that a new user has joined.
|
|
# Here, `room=room` ensures the message is sent to everyone in that specific room.
|
|
emit(
|
|
"message",
|
|
{"id": None, "content": f"{data['username']} has joined the room."},
|
|
room=room,
|
|
)
|
|
|
|
|
|
@socketio.on("message")
|
|
def handle_message(data):
|
|
# Save the message to the database
|
|
new_message = Message(
|
|
username=data["username"], content=data["message"], room=data["room"]
|
|
)
|
|
db.session.add(new_message)
|
|
db.session.commit()
|
|
|
|
emit(
|
|
"message",
|
|
{"id": new_message.id, "content": f"{data['username']}: {data['message']}"},
|
|
room=data["room"],
|
|
)
|
|
|
|
if "claude" in data["message"] or "gpt" in data["message"]:
|
|
# Emit a temporary message indicating that llm is processing
|
|
emit(
|
|
"message",
|
|
{"id": None, "content": f"<span id='processing'>Processing...</span>"},
|
|
room=data["room"],
|
|
)
|
|
|
|
if "claude" in data["message"]:
|
|
eventlet.spawn(chat_claude, data["username"], data["room"], data["message"])
|
|
if "gpt-3" in data["message"]:
|
|
eventlet.spawn(chat_gpt, data["username"], data["room"], data["message"])
|
|
if "gpt-4" in data["message"]:
|
|
eventlet.spawn(
|
|
chat_gpt,
|
|
data["username"],
|
|
data["room"],
|
|
data["message"],
|
|
model_name="gpt-4",
|
|
)
|
|
|
|
|
|
@socketio.on("delete_message")
|
|
def handle_delete_message(data):
|
|
msg_id = data["message_id"]
|
|
# Delete the message from the database
|
|
message = db.session.query(Message).filter(Message.id == msg_id).one_or_none()
|
|
if message:
|
|
db.session.delete(message)
|
|
db.session.commit()
|
|
|
|
# Notify all clients in the room to remove the message from their DOM
|
|
emit("message_deleted", {"message_id": msg_id}, room=data["room"])
|
|
|
|
|
|
def chat_claude(username, room, message, model_name="anthropic.claude-v2"):
|
|
with app.app_context():
|
|
# claude has a 100,000 token context window for prompts.
|
|
all_messages = (
|
|
Message.query.filter_by(room=room).order_by(Message.id.desc()).all()
|
|
)
|
|
|
|
chat_history = ""
|
|
|
|
for msg in reversed(all_messages):
|
|
if msg.username in ["gpt-3.5-turbo", "anthropic.claude-v2", "gpt-4"]:
|
|
chat_history += f"Assistant: {msg.username}: {msg.content}\n\n"
|
|
else:
|
|
chat_history += f"Human: {msg.username}: {msg.content}\n\n"
|
|
|
|
# append the new message.
|
|
chat_history += f"Human: {username}: {message}\n\nAssistant:"
|
|
|
|
# Initialize the Bedrock client using boto3 and profile name.
|
|
if profile_name:
|
|
session = boto3.Session(profile_name=profile_name)
|
|
client = session.client("bedrock-runtime", region_name="us-east-1")
|
|
else:
|
|
client = boto3.client("bedrock-runtime", region_name="us-east-1")
|
|
|
|
# Define the request parameters
|
|
params = {
|
|
"modelId": model_name,
|
|
"contentType": "application/json",
|
|
"accept": "*/*",
|
|
"body": json.dumps(
|
|
{
|
|
"prompt": chat_history,
|
|
"max_tokens_to_sample": 2048,
|
|
"temperature": 0,
|
|
"top_k": 250,
|
|
"top_p": 0.999,
|
|
"stop_sequences": ["\n\nHuman:"],
|
|
"anthropic_version": "bedrock-2023-05-31",
|
|
}
|
|
).encode(),
|
|
}
|
|
|
|
# Invoke the model with response stream
|
|
response = client.invoke_model_with_response_stream(**params)
|
|
|
|
# Process the event stream
|
|
buffer = ""
|
|
|
|
# save empty message, we need the ID when we chunk the response.
|
|
with app.app_context():
|
|
new_message = Message(username=model_name, content=buffer, room=room)
|
|
db.session.add(new_message)
|
|
db.session.commit()
|
|
msg_id = new_message.id
|
|
|
|
first_chunk = True
|
|
for event in response["body"]:
|
|
content = ""
|
|
|
|
if "chunk" in event:
|
|
chunk_data = json.loads(event["chunk"]["bytes"].decode())
|
|
content = chunk_data["completion"]
|
|
|
|
if content:
|
|
buffer += content # Accumulate content
|
|
|
|
if first_chunk:
|
|
socketio.emit(
|
|
"message_chunk",
|
|
{
|
|
"id": msg_id,
|
|
"content": f"{username} ({model_name}): {content}",
|
|
},
|
|
room=room,
|
|
)
|
|
first_chunk = False
|
|
else:
|
|
socketio.emit(
|
|
"message_chunk",
|
|
{"id": msg_id, "content": content},
|
|
room=room,
|
|
)
|
|
socketio.sleep(0) # Force immediate handling
|
|
|
|
# Save the entire completion to the database
|
|
with app.app_context():
|
|
new_message = (
|
|
db.session.query(Message).filter(Message.id == msg_id).one_or_none()
|
|
)
|
|
if new_message:
|
|
new_message.content = buffer
|
|
db.session.add(new_message)
|
|
db.session.commit()
|
|
|
|
socketio.emit("delete_processing_message", msg_id, room=room)
|
|
|
|
|
|
def chat_gpt(username, room, message, model_name="gpt-3.5-turbo"):
|
|
with app.app_context():
|
|
last_messages = (
|
|
Message.query.filter_by(room=room)
|
|
.order_by(Message.id.desc())
|
|
.limit(10)
|
|
.all()
|
|
)
|
|
|
|
chat_history = [
|
|
{
|
|
"role": "system"
|
|
if (
|
|
msg.username == "gpt-3.5-turbo"
|
|
or msg.username == "anthropic.claude-v2"
|
|
or msg.username == "gpt-4"
|
|
)
|
|
else "user",
|
|
"content": f"{msg.username}: {msg.content}",
|
|
}
|
|
for msg in reversed(last_messages)
|
|
]
|
|
|
|
chat_history.append({"role": "user", "content": message})
|
|
|
|
buffer = "" # Content buffer for accumulating the chunks
|
|
|
|
# save empty message, we need the ID when we chunk the response.
|
|
with app.app_context():
|
|
new_message = Message(username=model_name, content=buffer, room=room)
|
|
db.session.add(new_message)
|
|
db.session.commit()
|
|
msg_id = new_message.id
|
|
|
|
first_chunk = True
|
|
for chunk in openai.ChatCompletion.create(
|
|
model=model_name,
|
|
messages=chat_history,
|
|
temperature=0,
|
|
stream=True,
|
|
):
|
|
content = chunk["choices"][0].get("delta", {}).get("content")
|
|
|
|
if content:
|
|
buffer += content # Accumulate content
|
|
|
|
if first_chunk:
|
|
socketio.emit(
|
|
"message_chunk",
|
|
{
|
|
"id": msg_id,
|
|
"content": f"{username} ({model_name}): {content}",
|
|
},
|
|
room=room,
|
|
)
|
|
first_chunk = False
|
|
else:
|
|
socketio.emit(
|
|
"message_chunk",
|
|
{"id": msg_id, "content": content},
|
|
room=room,
|
|
)
|
|
socketio.sleep(0) # Force immediate handling
|
|
|
|
# Save the entire completion to the database
|
|
with app.app_context():
|
|
new_message = (
|
|
db.session.query(Message).filter(Message.id == msg_id).one_or_none()
|
|
)
|
|
if new_message:
|
|
new_message.content = buffer
|
|
db.session.add(new_message)
|
|
db.session.commit()
|
|
|
|
socketio.emit("delete_processing_message", msg_id, room=room)
|
|
|
|
|
|
if __name__ == "__main__":
|
|
socketio.run(app, host="0.0.0.0", port=5001)
|