tool_use #1
5
.gitignore
vendored
5
.gitignore
vendored
@ -213,4 +213,7 @@ jspm_packages
|
||||
|
||||
# Optional REPL history
|
||||
.node_repl_history
|
||||
.next
|
||||
.next
|
||||
|
||||
config.ini
|
||||
*.db
|
||||
138
client.py
Normal file
138
client.py
Normal file
@ -0,0 +1,138 @@
|
||||
import time
|
||||
|
||||
import requests
|
||||
|
||||
|
||||
class LLMChatClient:
|
||||
def __init__(self, base_url, api_key):
|
||||
self.base_url = base_url.rstrip("/")
|
||||
self.api_key = api_key
|
||||
self.headers = {"X-API-Key": api_key, "Content-Type": "application/json"}
|
||||
|
||||
def submit_query(self, message):
|
||||
"""
|
||||
Submit a query to the LLM Chat Server.
|
||||
|
||||
Args:
|
||||
message (str): The message to send to the server.
|
||||
|
||||
Returns:
|
||||
str: The query ID for the submitted query.
|
||||
|
||||
Raises:
|
||||
requests.RequestException: If the request fails.
|
||||
|
||||
Example:
|
||||
client = LLMChatClient('http://localhost:5001', 'your-api-key')
|
||||
query_id = client.submit_query('What is the capital of France?')
|
||||
print(f"Query ID: {query_id}")
|
||||
|
||||
cURL equivalent:
|
||||
curl -X POST http://localhost:5001/api/v1/query \
|
||||
-H "Content-Type: application/json" \
|
||||
-H "X-API-Key: your-api-key" \
|
||||
-d '{"message": "What is the capital of France?"}'
|
||||
"""
|
||||
url = f"{self.base_url}/api/v1/query"
|
||||
data = {"message": message}
|
||||
response = requests.post(url, json=data, headers=self.headers)
|
||||
response.raise_for_status()
|
||||
return response.json()["query_id"]
|
||||
|
||||
def get_query_status(self, query_id):
|
||||
"""
|
||||
Get the status of a submitted query.
|
||||
|
||||
Args:
|
||||
query_id (str): The ID of the query to check.
|
||||
|
||||
Returns:
|
||||
dict: A dictionary containing the status and conversation history (if completed).
|
||||
|
||||
Raises:
|
||||
requests.RequestException: If the request fails.
|
||||
|
||||
Example:
|
||||
client = LLMChatClient('http://localhost:5001', 'your-api-key')
|
||||
status = client.get_query_status('query-id-here')
|
||||
print(f"Query status: {status['status']}")
|
||||
if status['status'] == 'completed':
|
||||
print(f"Conversation history: {status['conversation_history']}")
|
||||
|
||||
cURL equivalent:
|
||||
curl -X GET http://localhost:5001/api/v1/query_status/query-id-here \
|
||||
-H "X-API-Key: your-api-key"
|
||||
"""
|
||||
url = f"{self.base_url}/api/v1/query_status/{query_id}"
|
||||
response = requests.get(url, headers=self.headers)
|
||||
response.raise_for_status()
|
||||
return response.json()
|
||||
|
||||
def submit_query_and_wait(self, message, max_wait_time=300, poll_interval=2):
|
||||
"""
|
||||
Submit a query and wait for the result.
|
||||
|
||||
Args:
|
||||
message (str): The message to send to the server.
|
||||
max_wait_time (int): Maximum time to wait for the result in seconds.
|
||||
poll_interval (int): Time between status checks in seconds.
|
||||
|
||||
Returns:
|
||||
dict: The completed conversation history.
|
||||
|
||||
Raises:
|
||||
requests.RequestException: If the request fails.
|
||||
TimeoutError: If the query doesn't complete within max_wait_time.
|
||||
|
||||
Example:
|
||||
client = LLMChatClient('http://localhost:5001', 'your-api-key')
|
||||
result = client.submit_query_and_wait('What is the capital of France?')
|
||||
print(f"Conversation history: {result}")
|
||||
"""
|
||||
query_id = self.submit_query(message)
|
||||
start_time = time.time()
|
||||
|
||||
while time.time() - start_time < max_wait_time:
|
||||
status = self.get_query_status(query_id)
|
||||
if status["status"] == "completed":
|
||||
return status["conversation_history"]
|
||||
time.sleep(poll_interval)
|
||||
|
||||
raise TimeoutError(f"Query did not complete within {max_wait_time} seconds")
|
||||
|
||||
|
||||
class LLMChatAdminClient:
|
||||
def __init__(self, base_url, admin_key):
|
||||
self.base_url = base_url.rstrip("/")
|
||||
self.admin_key = admin_key
|
||||
self.headers = {"X-Admin-Key": admin_key, "Content-Type": "application/json"}
|
||||
|
||||
def generate_api_key(self, username):
|
||||
"""
|
||||
Generate a new API key for a user.
|
||||
|
||||
Args:
|
||||
username (str): The username to generate the API key for.
|
||||
|
||||
Returns:
|
||||
dict: A dictionary containing the username and generated API key.
|
||||
|
||||
Raises:
|
||||
requests.RequestException: If the request fails.
|
||||
|
||||
Example:
|
||||
admin_client = LLMChatAdminClient('http://localhost:5001', 'your-admin-key')
|
||||
result = admin_client.generate_api_key('new_user')
|
||||
print(f"Generated API key for {result['username']}: {result['api_key']}")
|
||||
|
||||
cURL equivalent:
|
||||
curl -X POST http://localhost:5001/admin/generate_key \
|
||||
-H "Content-Type: application/json" \
|
||||
-H "X-Admin-Key: your-admin-key" \
|
||||
-d '{"username": "new_user"}'
|
||||
"""
|
||||
url = f"{self.base_url}/admin/generate_key"
|
||||
data = {"username": username}
|
||||
response = requests.post(url, json=data, headers=self.headers)
|
||||
response.raise_for_status()
|
||||
return response.json()
|
||||
@ -1,3 +0,0 @@
|
||||
{
|
||||
"extends": ["next/core-web-vitals", "next/typescript"]
|
||||
}
|
||||
36
dewey/.gitignore
vendored
36
dewey/.gitignore
vendored
@ -1,36 +0,0 @@
|
||||
# See https://help.github.com/articles/ignoring-files/ for more about ignoring files.
|
||||
|
||||
# dependencies
|
||||
/node_modules
|
||||
/.pnp
|
||||
.pnp.js
|
||||
.yarn/install-state.gz
|
||||
|
||||
# testing
|
||||
/coverage
|
||||
|
||||
# next.js
|
||||
/.next/
|
||||
/out/
|
||||
|
||||
# production
|
||||
/build
|
||||
|
||||
# misc
|
||||
.DS_Store
|
||||
*.pem
|
||||
|
||||
# debug
|
||||
npm-debug.log*
|
||||
yarn-debug.log*
|
||||
yarn-error.log*
|
||||
|
||||
# local env files
|
||||
.env*.local
|
||||
|
||||
# vercel
|
||||
.vercel
|
||||
|
||||
# typescript
|
||||
*.tsbuildinfo
|
||||
next-env.d.ts
|
||||
@ -1,36 +0,0 @@
|
||||
This is a [Next.js](https://nextjs.org) project bootstrapped with [`create-next-app`](https://nextjs.org/docs/app/api-reference/cli/create-next-app).
|
||||
|
||||
## Getting Started
|
||||
|
||||
First, run the development server:
|
||||
|
||||
```bash
|
||||
npm run dev
|
||||
# or
|
||||
yarn dev
|
||||
# or
|
||||
pnpm dev
|
||||
# or
|
||||
bun dev
|
||||
```
|
||||
|
||||
Open [http://localhost:3000](http://localhost:3000) with your browser to see the result.
|
||||
|
||||
You can start editing the page by modifying `app/page.tsx`. The page auto-updates as you edit the file.
|
||||
|
||||
This project uses [`next/font`](https://nextjs.org/docs/app/building-your-application/optimizing/fonts) to automatically optimize and load [Geist](https://vercel.com/font), a new font family for Vercel.
|
||||
|
||||
## Learn More
|
||||
|
||||
To learn more about Next.js, take a look at the following resources:
|
||||
|
||||
- [Next.js Documentation](https://nextjs.org/docs) - learn about Next.js features and API.
|
||||
- [Learn Next.js](https://nextjs.org/learn) - an interactive Next.js tutorial.
|
||||
|
||||
You can check out [the Next.js GitHub repository](https://github.com/vercel/next.js) - your feedback and contributions are welcome!
|
||||
|
||||
## Deploy on Vercel
|
||||
|
||||
The easiest way to deploy your Next.js app is to use the [Vercel Platform](https://vercel.com/new?utm_medium=default-template&filter=next.js&utm_source=create-next-app&utm_campaign=create-next-app-readme) from the creators of Next.js.
|
||||
|
||||
Check out our [Next.js deployment documentation](https://nextjs.org/docs/app/building-your-application/deploying) for more details.
|
||||
Binary file not shown.
Binary file not shown.
Binary file not shown.
@ -1,64 +0,0 @@
|
||||
@tailwind base;
|
||||
@tailwind components;
|
||||
@tailwind utilities;
|
||||
|
||||
:root {
|
||||
--foreground-rgb: 255, 255, 255;
|
||||
--background-start-rgb: 0, 0, 0;
|
||||
--background-end-rgb: 0, 0, 0;
|
||||
}
|
||||
|
||||
body {
|
||||
color: rgb(var(--foreground-rgb));
|
||||
background: linear-gradient(
|
||||
to bottom,
|
||||
transparent,
|
||||
rgb(var(--background-end-rgb))
|
||||
)
|
||||
rgb(var(--background-start-rgb));
|
||||
font-family: 'Noto Sans Mono', monospace;
|
||||
}
|
||||
|
||||
/* Custom styles from the original index.html */
|
||||
.thinking-section {
|
||||
margin-bottom: 20px;
|
||||
border-left: 2px solid #444;
|
||||
padding-left: 10px;
|
||||
}
|
||||
|
||||
.thought-summary {
|
||||
font-weight: bold;
|
||||
margin-bottom: 5px;
|
||||
padding: 5px;
|
||||
border-radius: 3px;
|
||||
}
|
||||
|
||||
.thought-summary.plan { background-color: #2c3e50; }
|
||||
.thought-summary.decision { background-color: #34495e; }
|
||||
.thought-summary.tool_call { background-color: #16a085; }
|
||||
.thought-summary.tool_result { background-color: #27ae60; }
|
||||
.thought-summary.think_more { background-color: #2980b9; }
|
||||
.thought-summary.answer { background-color: #8e44ad; }
|
||||
|
||||
.thought-details {
|
||||
display: none;
|
||||
margin-left: 20px;
|
||||
border-left: 2px solid #444;
|
||||
padding-left: 10px;
|
||||
margin-bottom: 10px;
|
||||
white-space: pre-wrap;
|
||||
font-family: 'Noto Sans Mono', monospace;
|
||||
background-color: #222;
|
||||
}
|
||||
|
||||
.collapsible::before {
|
||||
content: '▶ ';
|
||||
display: inline-block;
|
||||
transition: transform 0.3s;
|
||||
}
|
||||
|
||||
.collapsible.open::before {
|
||||
transform: rotate(90deg);
|
||||
}
|
||||
|
||||
/* Add any other custom styles from the original index.html here */
|
||||
@ -1,31 +0,0 @@
|
||||
import './globals.css'
|
||||
import { Inter } from 'next/font/google'
|
||||
|
||||
const inter = Inter({ subsets: ['latin'] })
|
||||
|
||||
export const metadata = {
|
||||
title: 'DWS Intelligence',
|
||||
description: 'AI-powered chat application',
|
||||
}
|
||||
|
||||
export default function RootLayout({
|
||||
children,
|
||||
}: {
|
||||
children: React.ReactNode
|
||||
}) {
|
||||
return (
|
||||
<html lang="en">
|
||||
<head>
|
||||
<script src="https://cdnjs.cloudflare.com/ajax/libs/socket.io/4.0.1/socket.io.js"></script>
|
||||
<script src="https://cdn.jsdelivr.net/npm/marked/marked.min.js"></script>
|
||||
<script src="https://cdn.jsdelivr.net/npm/moment@2.29.4/moment.min.js"></script>
|
||||
<script src="https://cdn.jsdelivr.net/npm/chart.js"></script>
|
||||
<script src="https://cdn.jsdelivr.net/npm/chartjs-adapter-moment@1.0.1/dist/chartjs-adapter-moment.min.js"></script>
|
||||
<script src="https://cdnjs.cloudflare.com/ajax/libs/highlight.js/11.7.0/highlight.min.js"></script>
|
||||
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/highlight.js/11.7.0/styles/default.min.css" />
|
||||
<link href="https://fonts.googleapis.com/css2?family=Noto+Sans+Mono:wght@400;700&display=swap" rel="stylesheet" />
|
||||
</head>
|
||||
<body className={inter.className}>{children}</body>
|
||||
</html>
|
||||
)
|
||||
}
|
||||
@ -1,44 +0,0 @@
|
||||
'use client'
|
||||
|
||||
import { useState, useEffect } from 'react'
|
||||
import ChatArea from '../components/ChatArea'
|
||||
import Sidebar from '../components/Sidebar'
|
||||
import useSocket from '../hooks/useSocket'
|
||||
import useLocalStorage from '../hooks/useLocalStorage'
|
||||
|
||||
export default function Home() {
|
||||
const [currentChatId, setCurrentChatId] = useState<string | null>(null)
|
||||
const [chats, setChats] = useLocalStorage('chats', {})
|
||||
const socket = useSocket()
|
||||
|
||||
useEffect(() => {
|
||||
if (Object.keys(chats).length === 0) {
|
||||
createNewChat()
|
||||
} else {
|
||||
setCurrentChatId(Object.keys(chats)[0])
|
||||
}
|
||||
}, [])
|
||||
|
||||
const createNewChat = () => {
|
||||
const chatId = Date.now().toString()
|
||||
setChats(prevChats => ({
|
||||
...prevChats,
|
||||
[chatId]: { messages: [], thinkingSections: [] }
|
||||
}))
|
||||
setCurrentChatId(chatId)
|
||||
}
|
||||
|
||||
return (
|
||||
<div className="flex h-screen">
|
||||
<ChatArea
|
||||
currentChatId={currentChatId}
|
||||
setCurrentChatId={setCurrentChatId}
|
||||
chats={chats}
|
||||
setChats={setChats}
|
||||
createNewChat={createNewChat}
|
||||
socket={socket}
|
||||
/>
|
||||
<Sidebar socket={socket} />
|
||||
</div>
|
||||
)
|
||||
}
|
||||
@ -1,133 +0,0 @@
|
||||
import { useState, useEffect } from 'react'
|
||||
import ChatTabs from './ChatTabs'
|
||||
import ChatContainer from './ChatContainer'
|
||||
import UserInput from './UserInput'
|
||||
|
||||
export default function ChatArea({ currentChatId, setCurrentChatId, chats, setChats, createNewChat, socket }) {
|
||||
const [userInput, setUserInput] = useState('')
|
||||
|
||||
const sendMessage = () => {
|
||||
if (userInput.trim() && currentChatId) {
|
||||
const newMessage = { content: userInput, isUser: true }
|
||||
setChats(prevChats => ({
|
||||
...prevChats,
|
||||
[currentChatId]: {
|
||||
...prevChats[currentChatId],
|
||||
messages: [...prevChats[currentChatId].messages, newMessage],
|
||||
thinkingSections: [...prevChats[currentChatId].thinkingSections, { thoughts: [] }]
|
||||
}
|
||||
}))
|
||||
socket.emit('chat_request', {
|
||||
message: userInput,
|
||||
conversation_history: chats[currentChatId].messages
|
||||
.filter(m => !m.isUser).map(m => ({ role: 'assistant', content: m.content }))
|
||||
.concat(chats[currentChatId].messages.filter(m => m.isUser).map(m => ({ role: 'user', content: m.content })))
|
||||
})
|
||||
setUserInput('')
|
||||
}
|
||||
}
|
||||
|
||||
const switchToChat = (chatId: string) => {
|
||||
setCurrentChatId(chatId);
|
||||
}
|
||||
|
||||
const closeChat = (chatId: string) => {
|
||||
if (window.confirm('Are you sure you want to close this chat?')) {
|
||||
setChats(prevChats => {
|
||||
const newChats = { ...prevChats };
|
||||
delete newChats[chatId];
|
||||
return newChats;
|
||||
});
|
||||
if (currentChatId === chatId) {
|
||||
const remainingChatIds = Object.keys(chats).filter(id => id !== chatId);
|
||||
if (remainingChatIds.length > 0) {
|
||||
switchToChat(remainingChatIds[0]);
|
||||
} else {
|
||||
createNewChat();
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
useEffect(() => {
|
||||
if (socket) {
|
||||
socket.on('thinking', (data) => {
|
||||
// Handle thinking event
|
||||
setChats(prevChats => ({
|
||||
...prevChats,
|
||||
[currentChatId]: {
|
||||
...prevChats[currentChatId],
|
||||
thinkingSections: [
|
||||
...prevChats[currentChatId].thinkingSections,
|
||||
{ thoughts: [{ type: 'thinking', content: data.step }] }
|
||||
]
|
||||
}
|
||||
}));
|
||||
});
|
||||
|
||||
socket.on('thought', (data) => {
|
||||
// Handle thought event
|
||||
setChats(prevChats => ({
|
||||
...prevChats,
|
||||
[currentChatId]: {
|
||||
...prevChats[currentChatId],
|
||||
thinkingSections: prevChats[currentChatId].thinkingSections.map((section, index) =>
|
||||
index === prevChats[currentChatId].thinkingSections.length - 1
|
||||
? { ...section, thoughts: [...section.thoughts, data] }
|
||||
: section
|
||||
)
|
||||
}
|
||||
}));
|
||||
});
|
||||
|
||||
socket.on('chat_response', (data) => {
|
||||
// Handle chat response event
|
||||
setChats(prevChats => ({
|
||||
...prevChats,
|
||||
[currentChatId]: {
|
||||
...prevChats[currentChatId],
|
||||
messages: [...prevChats[currentChatId].messages, { content: data.response, isUser: false }]
|
||||
}
|
||||
}));
|
||||
});
|
||||
|
||||
socket.on('error', (data) => {
|
||||
// Handle error event
|
||||
console.error('Error:', data.message);
|
||||
// You might want to display this error to the user
|
||||
});
|
||||
}
|
||||
|
||||
return () => {
|
||||
if (socket) {
|
||||
socket.off('thinking');
|
||||
socket.off('thought');
|
||||
socket.off('chat_response');
|
||||
socket.off('error');
|
||||
}
|
||||
};
|
||||
}, [socket, currentChatId, setChats]);
|
||||
|
||||
return (
|
||||
<div className="flex flex-col flex-1">
|
||||
<ChatTabs
|
||||
chats={chats}
|
||||
currentChatId={currentChatId}
|
||||
createNewChat={createNewChat}
|
||||
switchToChat={switchToChat}
|
||||
closeChat={closeChat}
|
||||
/>
|
||||
{currentChatId && (
|
||||
<ChatContainer
|
||||
currentChat={chats[currentChatId]}
|
||||
socket={socket}
|
||||
/>
|
||||
)}
|
||||
<UserInput
|
||||
value={userInput}
|
||||
onChange={setUserInput}
|
||||
onSend={sendMessage}
|
||||
/>
|
||||
</div>
|
||||
)
|
||||
}
|
||||
@ -1,85 +0,0 @@
|
||||
import React, { useEffect, useRef } from 'react';
|
||||
import { marked } from 'marked';
|
||||
|
||||
interface ChatContainerProps {
|
||||
currentChat: {
|
||||
messages: Array<{ content: string; isUser: boolean }>;
|
||||
thinkingSections: Array<{ thoughts: Array<{ type: string; content: string; details?: string }> }>;
|
||||
} | null;
|
||||
socket: any;
|
||||
}
|
||||
|
||||
const ChatContainer: React.FC<ChatContainerProps> = ({ currentChat, socket }) => {
|
||||
const chatContainerRef = useRef<HTMLDivElement>(null);
|
||||
|
||||
useEffect(() => {
|
||||
if (chatContainerRef.current) {
|
||||
chatContainerRef.current.scrollTop = chatContainerRef.current.scrollHeight;
|
||||
}
|
||||
}, [currentChat]);
|
||||
|
||||
if (!currentChat) return null;
|
||||
|
||||
return (
|
||||
<div ref={chatContainerRef} className="flex-1 overflow-y-auto p-4 bg-gray-900">
|
||||
{currentChat.messages.map((message, index) => (
|
||||
<div
|
||||
key={index}
|
||||
className={`mb-4 ${
|
||||
message.isUser ? 'text-right text-cyan-300' : 'text-left text-white'
|
||||
}`}
|
||||
>
|
||||
<div
|
||||
className={`inline-block p-2 rounded-lg ${
|
||||
message.isUser ? 'bg-cyan-800' : 'bg-gray-700'
|
||||
}`}
|
||||
>
|
||||
{message.isUser ? (
|
||||
message.content
|
||||
) : (
|
||||
<div dangerouslySetInnerHTML={{ __html: marked(message.content) }} />
|
||||
)}
|
||||
</div>
|
||||
</div>
|
||||
))}
|
||||
{currentChat.thinkingSections.map((section, sectionIndex) => (
|
||||
<div key={sectionIndex} className="mb-4 border-l-2 border-gray-600 pl-4">
|
||||
{section.thoughts.map((thought, thoughtIndex) => (
|
||||
<div key={thoughtIndex} className="mb-2">
|
||||
<div className={`font-bold ${getThoughtColor(thought.type)}`}>
|
||||
{thought.type}:
|
||||
</div>
|
||||
<div dangerouslySetInnerHTML={{ __html: marked(thought.content) }} />
|
||||
{thought.details && (
|
||||
<pre className="mt-2 p-2 bg-gray-800 rounded">
|
||||
{thought.details}
|
||||
</pre>
|
||||
)}
|
||||
</div>
|
||||
))}
|
||||
</div>
|
||||
))}
|
||||
</div>
|
||||
);
|
||||
};
|
||||
|
||||
function getThoughtColor(type: string): string {
|
||||
switch (type.toLowerCase()) {
|
||||
case 'plan':
|
||||
return 'text-blue-400';
|
||||
case 'decision':
|
||||
return 'text-green-400';
|
||||
case 'tool_call':
|
||||
return 'text-yellow-400';
|
||||
case 'tool_result':
|
||||
return 'text-purple-400';
|
||||
case 'think_more':
|
||||
return 'text-pink-400';
|
||||
case 'answer':
|
||||
return 'text-red-400';
|
||||
default:
|
||||
return 'text-gray-400';
|
||||
}
|
||||
}
|
||||
|
||||
export default ChatContainer;
|
||||
@ -1,48 +0,0 @@
|
||||
import React from 'react';
|
||||
|
||||
interface ChatTabsProps {
|
||||
chats: Record<string, any>;
|
||||
currentChatId: string | null;
|
||||
createNewChat: () => void;
|
||||
switchToChat: (chatId: string) => void;
|
||||
closeChat: (chatId: string) => void;
|
||||
}
|
||||
|
||||
const ChatTabs: React.FC<ChatTabsProps> = ({ chats, currentChatId, createNewChat, switchToChat, closeChat }) => {
|
||||
return (
|
||||
<div className="flex bg-gray-800 p-2">
|
||||
{Object.keys(chats).map((chatId) => (
|
||||
<div
|
||||
key={chatId}
|
||||
className={`px-4 py-2 mr-2 rounded-t-lg flex items-center ${
|
||||
chatId === currentChatId ? 'bg-gray-600' : 'bg-gray-700'
|
||||
}`}
|
||||
>
|
||||
<button
|
||||
onClick={() => switchToChat(chatId)}
|
||||
className="flex-grow text-left"
|
||||
>
|
||||
Chat {chatId}
|
||||
</button>
|
||||
<button
|
||||
className="ml-2 text-red-500 hover:text-red-700"
|
||||
onClick={(e) => {
|
||||
e.stopPropagation();
|
||||
closeChat(chatId);
|
||||
}}
|
||||
>
|
||||
×
|
||||
</button>
|
||||
</div>
|
||||
))}
|
||||
<button
|
||||
className="px-4 py-2 bg-green-600 rounded-t-lg"
|
||||
onClick={createNewChat}
|
||||
>
|
||||
+ New Chat
|
||||
</button>
|
||||
</div>
|
||||
);
|
||||
};
|
||||
|
||||
export default ChatTabs;
|
||||
@ -1,129 +0,0 @@
|
||||
import React, { useEffect, useRef, useState } from 'react';
|
||||
import dynamic from 'next/dynamic';
|
||||
|
||||
const Chart = dynamic(() => import('chart.js/auto').then((mod) => mod.Chart), {
|
||||
ssr: false,
|
||||
});
|
||||
|
||||
interface SidebarProps {
|
||||
socket: any;
|
||||
}
|
||||
|
||||
const Sidebar: React.FC<SidebarProps> = ({ socket }) => {
|
||||
const [isCollapsed, setIsCollapsed] = useState(false);
|
||||
const chartRefs = useRef<{ [key: string]: any }>({
|
||||
cpu: null,
|
||||
memory: null,
|
||||
disk: null,
|
||||
gpu: null,
|
||||
gpuMemory: null,
|
||||
});
|
||||
|
||||
useEffect(() => {
|
||||
if (socket) {
|
||||
socket.on('system_resources', (data: any) => {
|
||||
updateCharts(data);
|
||||
});
|
||||
}
|
||||
|
||||
return () => {
|
||||
if (socket) {
|
||||
socket.off('system_resources');
|
||||
}
|
||||
};
|
||||
}, [socket]);
|
||||
|
||||
useEffect(() => {
|
||||
const initCharts = async () => {
|
||||
const ChartJS = await Chart;
|
||||
initializeCharts(ChartJS);
|
||||
};
|
||||
initCharts();
|
||||
|
||||
return () => {
|
||||
Object.values(chartRefs.current).forEach(chart => chart?.destroy());
|
||||
};
|
||||
}, []);
|
||||
|
||||
const initializeCharts = (ChartJS: any) => {
|
||||
const chartConfig = {
|
||||
type: 'line',
|
||||
options: {
|
||||
responsive: true,
|
||||
maintainAspectRatio: false,
|
||||
scales: {
|
||||
x: {
|
||||
type: 'time',
|
||||
time: {
|
||||
unit: 'second',
|
||||
},
|
||||
},
|
||||
y: {
|
||||
beginAtZero: true,
|
||||
max: 100,
|
||||
},
|
||||
},
|
||||
animation: false,
|
||||
},
|
||||
data: {
|
||||
datasets: [{
|
||||
data: [],
|
||||
borderColor: 'rgb(75, 192, 192)',
|
||||
tension: 0.1,
|
||||
}],
|
||||
},
|
||||
};
|
||||
|
||||
['cpu', 'memory', 'disk', 'gpu', 'gpuMemory'].forEach(chartName => {
|
||||
const ctx = document.getElementById(`${chartName}Chart`) as HTMLCanvasElement;
|
||||
if (ctx) {
|
||||
chartRefs.current[chartName] = new ChartJS(ctx, chartConfig);
|
||||
}
|
||||
});
|
||||
};
|
||||
|
||||
const updateCharts = (data: any) => {
|
||||
const now = new Date();
|
||||
Object.entries(data).forEach(([key, value]) => {
|
||||
const chartName = key.replace('_', '').toLowerCase();
|
||||
const chart = chartRefs.current[chartName];
|
||||
if (chart) {
|
||||
chart.data.datasets[0].data.push({x: now, y: value});
|
||||
chart.update('none');
|
||||
}
|
||||
});
|
||||
};
|
||||
|
||||
return (
|
||||
<div className={`w-80 bg-gray-800 p-4 ${isCollapsed ? 'hidden' : ''}`}>
|
||||
<button
|
||||
onClick={() => setIsCollapsed(!isCollapsed)}
|
||||
className="mb-4 px-4 py-2 bg-gray-700 text-white rounded-lg"
|
||||
>
|
||||
{isCollapsed ? 'Show Charts' : 'Hide Charts'}
|
||||
</button>
|
||||
<div className="mb-4">
|
||||
<h3 className="text-white mb-2">CPU Load</h3>
|
||||
<canvas id="cpuChart"></canvas>
|
||||
</div>
|
||||
<div className="mb-4">
|
||||
<h3 className="text-white mb-2">Memory Usage</h3>
|
||||
<canvas id="memoryChart"></canvas>
|
||||
</div>
|
||||
<div className="mb-4">
|
||||
<h3 className="text-white mb-2">Disk I/O</h3>
|
||||
<canvas id="diskChart"></canvas>
|
||||
</div>
|
||||
<div className="mb-4">
|
||||
<h3 className="text-white mb-2">GPU Load</h3>
|
||||
<canvas id="gpuChart"></canvas>
|
||||
</div>
|
||||
<div className="mb-4">
|
||||
<h3 className="text-white mb-2">GPU Memory</h3>
|
||||
<canvas id="gpuMemoryChart"></canvas>
|
||||
</div>
|
||||
</div>
|
||||
);
|
||||
};
|
||||
|
||||
export default Sidebar;
|
||||
@ -1,37 +0,0 @@
|
||||
import React from 'react';
|
||||
|
||||
interface UserInputProps {
|
||||
value: string;
|
||||
onChange: (value: string) => void;
|
||||
onSend: () => void;
|
||||
}
|
||||
|
||||
const UserInput: React.FC<UserInputProps> = ({ value, onChange, onSend }) => {
|
||||
const handleKeyPress = (e: React.KeyboardEvent<HTMLTextAreaElement>) => {
|
||||
if (e.key === 'Enter' && !e.shiftKey) {
|
||||
e.preventDefault();
|
||||
onSend();
|
||||
}
|
||||
};
|
||||
|
||||
return (
|
||||
<div className="p-4 bg-gray-800">
|
||||
<textarea
|
||||
value={value}
|
||||
onChange={(e) => onChange(e.target.value)}
|
||||
onKeyPress={handleKeyPress}
|
||||
className="w-full p-2 bg-gray-700 text-white rounded-lg resize-none"
|
||||
rows={3}
|
||||
placeholder="Type your message here..."
|
||||
/>
|
||||
<button
|
||||
onClick={onSend}
|
||||
className="mt-2 px-4 py-2 bg-blue-600 text-white rounded-lg"
|
||||
>
|
||||
Send
|
||||
</button>
|
||||
</div>
|
||||
);
|
||||
};
|
||||
|
||||
export default UserInput;
|
||||
@ -1,23 +0,0 @@
|
||||
import { useState, useEffect } from 'react'
|
||||
|
||||
export default function useLocalStorage(key, initialValue) {
|
||||
const [storedValue, setStoredValue] = useState(() => {
|
||||
try {
|
||||
const item = window.localStorage.getItem(key)
|
||||
return item ? JSON.parse(item) : initialValue
|
||||
} catch (error) {
|
||||
console.log(error)
|
||||
return initialValue
|
||||
}
|
||||
})
|
||||
|
||||
useEffect(() => {
|
||||
try {
|
||||
window.localStorage.setItem(key, JSON.stringify(storedValue))
|
||||
} catch (error) {
|
||||
console.log(error)
|
||||
}
|
||||
}, [key, storedValue])
|
||||
|
||||
return [storedValue, setStoredValue]
|
||||
}
|
||||
@ -1,14 +0,0 @@
|
||||
import { useEffect, useState } from 'react'
|
||||
import io from 'socket.io-client'
|
||||
|
||||
export default function useSocket() {
|
||||
const [socket, setSocket] = useState<any>(null)
|
||||
|
||||
useEffect(() => {
|
||||
const newSocket = io('http://localhost:5001')
|
||||
setSocket(newSocket)
|
||||
return () => newSocket.close()
|
||||
}, [])
|
||||
|
||||
return socket
|
||||
}
|
||||
@ -1,4 +0,0 @@
|
||||
/** @type {import('next').NextConfig} */
|
||||
const nextConfig = {};
|
||||
|
||||
export default nextConfig;
|
||||
5574
dewey/package-lock.json
generated
5574
dewey/package-lock.json
generated
File diff suppressed because it is too large
Load Diff
@ -1,31 +0,0 @@
|
||||
{
|
||||
"name": "dewey",
|
||||
"version": "0.1.0",
|
||||
"private": true,
|
||||
"scripts": {
|
||||
"dev": "next dev",
|
||||
"build": "next build",
|
||||
"start": "next start",
|
||||
"lint": "next lint"
|
||||
},
|
||||
"dependencies": {
|
||||
"autoprefixer": "^10.4.20",
|
||||
"chart.js": "^3.9.1",
|
||||
"chartjs-adapter-moment": "^1.0.1",
|
||||
"marked": "^4.3.0",
|
||||
"next": "^14.2.13",
|
||||
"react": "^18.3.1",
|
||||
"react-dom": "^18.3.1",
|
||||
"socket.io-client": "^4.8.0"
|
||||
},
|
||||
"devDependencies": {
|
||||
"@types/node": "^20.16.10",
|
||||
"@types/react": "^18.3.10",
|
||||
"@types/react-dom": "^18.3.0",
|
||||
"eslint": "^8",
|
||||
"eslint-config-next": "14.2.13",
|
||||
"postcss": "^8.4.47",
|
||||
"tailwindcss": "^3.4.13",
|
||||
"typescript": "^5.6.2"
|
||||
}
|
||||
}
|
||||
@ -1,8 +0,0 @@
|
||||
/** @type {import('postcss-load-config').Config} */
|
||||
const config = {
|
||||
plugins: {
|
||||
tailwindcss: {},
|
||||
},
|
||||
};
|
||||
|
||||
export default config;
|
||||
@ -1,19 +0,0 @@
|
||||
import type { Config } from "tailwindcss";
|
||||
|
||||
const config: Config = {
|
||||
content: [
|
||||
"./pages/**/*.{js,ts,jsx,tsx,mdx}",
|
||||
"./components/**/*.{js,ts,jsx,tsx,mdx}",
|
||||
"./app/**/*.{js,ts,jsx,tsx,mdx}",
|
||||
],
|
||||
theme: {
|
||||
extend: {
|
||||
colors: {
|
||||
background: "var(--background)",
|
||||
foreground: "var(--foreground)",
|
||||
},
|
||||
},
|
||||
},
|
||||
plugins: [],
|
||||
};
|
||||
export default config;
|
||||
@ -1,26 +0,0 @@
|
||||
{
|
||||
"compilerOptions": {
|
||||
"lib": ["dom", "dom.iterable", "esnext"],
|
||||
"allowJs": true,
|
||||
"skipLibCheck": true,
|
||||
"strict": true,
|
||||
"noEmit": true,
|
||||
"esModuleInterop": true,
|
||||
"module": "esnext",
|
||||
"moduleResolution": "bundler",
|
||||
"resolveJsonModule": true,
|
||||
"isolatedModules": true,
|
||||
"jsx": "preserve",
|
||||
"incremental": true,
|
||||
"plugins": [
|
||||
{
|
||||
"name": "next"
|
||||
}
|
||||
],
|
||||
"paths": {
|
||||
"@/*": ["./*"]
|
||||
}
|
||||
},
|
||||
"include": ["next-env.d.ts", "**/*.ts", "**/*.tsx", ".next/types/**/*.ts"],
|
||||
"exclude": ["node_modules"]
|
||||
}
|
||||
602
main.py
602
main.py
@ -1,125 +1,273 @@
|
||||
from flask import Flask, send_from_directory, request
|
||||
from flask_socketio import SocketIO, emit
|
||||
from flask_openapi3 import OpenAPI, Info
|
||||
from pydantic import BaseModel
|
||||
from typing import List
|
||||
from models import model_manager
|
||||
import structlog
|
||||
import time
|
||||
import psutil
|
||||
import GPUtil
|
||||
import threading
|
||||
import os
|
||||
from tools import DefaultToolManager
|
||||
import ollama
|
||||
import re
|
||||
import configparser
|
||||
import json
|
||||
from datetime import datetime
|
||||
import os
|
||||
import pprint
|
||||
import queue
|
||||
import re
|
||||
import secrets
|
||||
import sqlite3
|
||||
import threading
|
||||
import time
|
||||
import uuid
|
||||
from datetime import datetime
|
||||
from typing import List, Optional
|
||||
|
||||
import GPUtil
|
||||
import ollama
|
||||
import psutil
|
||||
import structlog
|
||||
from flask import Flask, g, jsonify, request, send_from_directory
|
||||
from flask_openapi3 import Info, OpenAPI
|
||||
from flask_socketio import SocketIO, emit
|
||||
from pydantic import BaseModel
|
||||
|
||||
from models import model_manager
|
||||
from tools import DefaultToolManager
|
||||
|
||||
logger = structlog.get_logger()
|
||||
|
||||
# Configuration setup
|
||||
CONFIG_FILE = "config.ini"
|
||||
|
||||
|
||||
def create_default_config():
|
||||
config = configparser.ConfigParser()
|
||||
config["DEFAULT"] = {
|
||||
"AdminKey": secrets.token_urlsafe(32),
|
||||
"DatabasePath": "llm_chat_server.db",
|
||||
}
|
||||
config["SERVER_FEATURES"] = {
|
||||
"EnableFrontend": "false",
|
||||
"EnableChatEndpoints": "false",
|
||||
"EnableAPIEndpoints": "true",
|
||||
}
|
||||
config["MODEL"] = {"PrimaryModel": "qwen2.5:14b"}
|
||||
config["PERFORMANCE"] = {"UpdateInterval": "0.1"}
|
||||
with open(CONFIG_FILE, "w") as configfile:
|
||||
config.write(configfile)
|
||||
|
||||
|
||||
def load_config():
|
||||
if not os.path.exists(CONFIG_FILE):
|
||||
create_default_config()
|
||||
|
||||
config = configparser.ConfigParser()
|
||||
config.read(CONFIG_FILE)
|
||||
return config
|
||||
|
||||
|
||||
config = load_config()
|
||||
ADMIN_KEY = config["DEFAULT"]["AdminKey"]
|
||||
DATABASE = config["DEFAULT"]["DatabasePath"]
|
||||
ENABLE_FRONTEND = config["SERVER_FEATURES"].getboolean("EnableFrontend")
|
||||
ENABLE_CHAT_ENDPOINTS = config["SERVER_FEATURES"].getboolean("EnableChatEndpoints")
|
||||
ENABLE_API_ENDPOINTS = config["SERVER_FEATURES"].getboolean("EnableAPIEndpoints")
|
||||
PRIMARY_MODEL = config["MODEL"]["PrimaryModel"]
|
||||
UPDATE_INTERVAL = config["PERFORMANCE"].getfloat("UpdateInterval")
|
||||
|
||||
openapi = OpenAPI(__name__, info=Info(title="LLM Chat Server", version="1.0.0"))
|
||||
app = openapi
|
||||
socketio = SocketIO(app, cors_allowed_origins="*")
|
||||
|
||||
tool_manager = DefaultToolManager()
|
||||
|
||||
@app.route('/')
|
||||
|
||||
# Database setup
|
||||
def get_db():
|
||||
db = getattr(g, "_database", None)
|
||||
if db is None:
|
||||
db = g._database = sqlite3.connect(DATABASE)
|
||||
db.row_factory = sqlite3.Row
|
||||
return db
|
||||
|
||||
|
||||
@app.teardown_appcontext
|
||||
def close_connection(exception):
|
||||
db = getattr(g, "_database", None)
|
||||
if db is not None:
|
||||
db.close()
|
||||
|
||||
|
||||
def init_db():
|
||||
with app.app_context():
|
||||
db = get_db()
|
||||
db.execute(
|
||||
"""
|
||||
CREATE TABLE IF NOT EXISTS Queries (
|
||||
id TEXT PRIMARY KEY,
|
||||
ip TEXT NOT NULL,
|
||||
timestamp DATETIME DEFAULT CURRENT_TIMESTAMP,
|
||||
query TEXT NOT NULL,
|
||||
api_key_id INTEGER,
|
||||
conversation_history TEXT,
|
||||
FOREIGN KEY (api_key_id) REFERENCES Keys (id)
|
||||
)
|
||||
"""
|
||||
)
|
||||
db.commit()
|
||||
|
||||
|
||||
# Create a schema.sql file with the following content:
|
||||
"""
|
||||
CREATE TABLE IF NOT EXISTS Keys (
|
||||
id INTEGER PRIMARY KEY AUTOINCREMENT,
|
||||
username TEXT NOT NULL UNIQUE,
|
||||
api_key TEXT NOT NULL UNIQUE
|
||||
);
|
||||
|
||||
CREATE TABLE IF NOT EXISTS Queries (
|
||||
id INTEGER PRIMARY KEY AUTOINCREMENT,
|
||||
request_id TEXT NOT NULL UNIQUE,
|
||||
ip TEXT NOT NULL,
|
||||
timestamp DATETIME DEFAULT CURRENT_TIMESTAMP,
|
||||
query TEXT NOT NULL,
|
||||
api_key_id INTEGER,
|
||||
conversation_history TEXT,
|
||||
FOREIGN KEY (api_key_id) REFERENCES Keys (id)
|
||||
);
|
||||
"""
|
||||
|
||||
|
||||
def validate_api_key(api_key):
|
||||
db = get_db()
|
||||
cursor = db.cursor()
|
||||
cursor.execute("SELECT id FROM Keys WHERE api_key = ?", (api_key,))
|
||||
result = cursor.fetchone()
|
||||
return result[0] if result else None
|
||||
|
||||
|
||||
@app.route("/")
|
||||
def index():
|
||||
logger.info("Serving index.html")
|
||||
return send_from_directory('.', 'index.html')
|
||||
if ENABLE_FRONTEND:
|
||||
logger.info("Serving index.html")
|
||||
return send_from_directory(".", "index.html")
|
||||
else:
|
||||
return jsonify({"error": "Frontend is disabled"}), 404
|
||||
|
||||
|
||||
class ChatRequest(BaseModel):
|
||||
message: str
|
||||
|
||||
|
||||
class ChatResponse(BaseModel):
|
||||
response: str
|
||||
|
||||
@socketio.on('chat_request')
|
||||
|
||||
@socketio.on("chat_request")
|
||||
def handle_chat_request(data):
|
||||
user_input = data['message']
|
||||
conversation_history = data.get('conversation_history', [])
|
||||
conversation_history = [{"role": "system", "content": ANSWER_QUESTION_PROMPT}] + conversation_history
|
||||
logger.info("Received chat request", user_input=user_input, conversation_history=conversation_history)
|
||||
|
||||
if not ENABLE_CHAT_ENDPOINTS:
|
||||
emit("error", {"message": "Chat endpoints are disabled"})
|
||||
return
|
||||
|
||||
user_input = data["message"]
|
||||
conversation_history = data.get("conversation_history", [])
|
||||
conversation_history = [
|
||||
{"role": "system", "content": ANSWER_QUESTION_PROMPT}
|
||||
] + conversation_history
|
||||
logger.info(
|
||||
"Received chat request",
|
||||
user_input=user_input,
|
||||
conversation_history=conversation_history,
|
||||
)
|
||||
|
||||
start_time = time.time()
|
||||
try:
|
||||
final_response = answer_question_tools(user_input, conversation_history)
|
||||
end_time = time.time()
|
||||
thinking_time = round(end_time - start_time, 2)
|
||||
|
||||
emit('chat_response', {
|
||||
'response': final_response,
|
||||
'thinking_time': thinking_time
|
||||
})
|
||||
|
||||
emit(
|
||||
"chat_response",
|
||||
{"response": final_response, "thinking_time": thinking_time},
|
||||
)
|
||||
except Exception as e:
|
||||
logger.exception("Error during chat processing", error=str(e))
|
||||
end_time = time.time()
|
||||
thinking_time = round(end_time - start_time, 2)
|
||||
emit('error', {
|
||||
'message': f"An error occurred: {str(e)}",
|
||||
'thinking_time': thinking_time
|
||||
})
|
||||
emit(
|
||||
"error",
|
||||
{"message": f"An error occurred: {str(e)}", "thinking_time": thinking_time},
|
||||
)
|
||||
|
||||
def answer_question_tools(user_input: str, conversation_history: List[dict], max_retries: int = 100):
|
||||
|
||||
def answer_question_tools(
|
||||
user_input: str, conversation_history: List[dict], max_retries: int = 100
|
||||
):
|
||||
global tool_manager
|
||||
|
||||
|
||||
# If conversation_history is empty, initialize it with the system prompt
|
||||
if not conversation_history:
|
||||
conversation_history = [
|
||||
{"role": "system", "content": ANSWER_QUESTION_PROMPT},
|
||||
]
|
||||
|
||||
logger.info("Starting chat", user_input=user_input, conversation_history=conversation_history)
|
||||
|
||||
logger.info(
|
||||
"Starting chat",
|
||||
user_input=user_input,
|
||||
conversation_history=conversation_history,
|
||||
)
|
||||
# Add the new user input to the conversation history
|
||||
conversation_history.append({"role": "user", "content": user_input})
|
||||
|
||||
emit('thinking', {'step': 'Starting'})
|
||||
emit('conversation_history', {'history': conversation_history})
|
||||
|
||||
emit("thinking", {"step": "Starting"})
|
||||
emit("conversation_history", {"history": conversation_history})
|
||||
|
||||
last_thought_content = None
|
||||
|
||||
for _ in range(max_retries):
|
||||
response = ollama.chat(model=PRIMARY_MODEL, messages=conversation_history, tools=tool_manager.get_tools_for_ollama_dict(), stream=False)
|
||||
assistant_message = response['message']
|
||||
|
||||
response = ollama.chat(
|
||||
model=PRIMARY_MODEL,
|
||||
messages=conversation_history,
|
||||
tools=tool_manager.get_tools_for_ollama_dict(),
|
||||
stream=False,
|
||||
)
|
||||
assistant_message = response["message"]
|
||||
|
||||
conversation_history.append(assistant_message)
|
||||
emit('conversation_history', {'history': conversation_history})
|
||||
emit("conversation_history", {"history": conversation_history})
|
||||
pprint.pp(assistant_message)
|
||||
|
||||
if 'tool_calls' in assistant_message:
|
||||
for tool_call in assistant_message['tool_calls']:
|
||||
tool_name = tool_call['function']['name']
|
||||
tool_args = tool_call['function']['arguments']
|
||||
emit('thought', {'type': 'tool_call', 'content': f"Tool: {tool_name}\nArguments: {tool_args}"})
|
||||
if "tool_calls" in assistant_message:
|
||||
for tool_call in assistant_message["tool_calls"]:
|
||||
tool_name = tool_call["function"]["name"]
|
||||
tool_args = tool_call["function"]["arguments"]
|
||||
emit(
|
||||
"thought",
|
||||
{
|
||||
"type": "tool_call",
|
||||
"content": f"Tool: {tool_name}\nArguments: {tool_args}",
|
||||
},
|
||||
)
|
||||
tool_response = tool_manager.get_tool(tool_name).execute(tool_args)
|
||||
conversation_history.append({
|
||||
"role": "tool",
|
||||
"content": tool_response
|
||||
})
|
||||
emit('conversation_history', {'history': conversation_history})
|
||||
emit('thought', {'type': 'tool_result', 'content': tool_response})
|
||||
conversation_history.append({"role": "tool", "content": tool_response})
|
||||
emit("conversation_history", {"history": conversation_history})
|
||||
emit("thought", {"type": "tool_result", "content": tool_response})
|
||||
else:
|
||||
if "<reply>" in assistant_message['content'].lower():
|
||||
reply_content = re.search(r'<reply>(.*?)</reply>', assistant_message['content'], re.DOTALL)
|
||||
if "<reply>" in assistant_message["content"].lower():
|
||||
reply_content = re.search(
|
||||
r"<reply>(.*?)</reply>", assistant_message["content"], re.DOTALL
|
||||
)
|
||||
if reply_content:
|
||||
reply_answer = reply_content.group(1).strip()
|
||||
emit('thought', {'type': 'answer', 'content': reply_answer})
|
||||
emit("thought", {"type": "answer", "content": reply_answer})
|
||||
return reply_answer
|
||||
else:
|
||||
current_thought_content = assistant_message['content'].strip()
|
||||
emit('thought', {'type': 'thoughts', 'content': current_thought_content})
|
||||
|
||||
current_thought_content = assistant_message["content"].strip()
|
||||
emit(
|
||||
"thought", {"type": "thoughts", "content": current_thought_content}
|
||||
)
|
||||
|
||||
# Check for two consecutive thoughts, with the second being empty
|
||||
if last_thought_content and not current_thought_content:
|
||||
emit('thought', {'type': 'answer', 'content': last_thought_content})
|
||||
emit("thought", {"type": "answer", "content": last_thought_content})
|
||||
return last_thought_content
|
||||
|
||||
|
||||
last_thought_content = current_thought_content
|
||||
continue
|
||||
|
||||
return f"Max iterations reached. Last response: {assistant_message['content']}"
|
||||
|
||||
|
||||
ANSWER_QUESTION_PROMPT2 = f"""
|
||||
The current date is {datetime.now().strftime("%A, %B %d, %Y")}, your knowledge cutoff was December 2023.
|
||||
You are Dewey, an AI assistant with access to external tools and the ability to think through complex problems. Your role is to assist users by leveraging tools when necessary, thinking deeply about problems, and providing accurate and helpful information, all with a cheerful, but witty personality. Here are the tools available to you:
|
||||
@ -216,9 +364,6 @@ Generate the final response to the user within <reply></reply> tags:
|
||||
Remember to always be helpful, accurate, and respectful in your interactions, while maintaining your distinctive character blend of House and Jarvis.
|
||||
"""
|
||||
|
||||
PRIMARY_MODEL = "qwen2.5:14b"
|
||||
|
||||
UPDATE_INTERVAL = 0.1 # 100ms, configurable
|
||||
|
||||
def get_system_resources():
|
||||
cpu_load = psutil.cpu_percent()
|
||||
@ -227,44 +372,319 @@ def get_system_resources():
|
||||
disk_io = psutil.disk_io_counters()
|
||||
disk_read = disk_io.read_bytes
|
||||
disk_write = disk_io.write_bytes
|
||||
|
||||
|
||||
gpus = GPUtil.getGPUs()
|
||||
gpu_load = gpus[0].load * 100 if gpus else 0
|
||||
gpu_memory = gpus[0].memoryUtil * 100 if gpus else 0
|
||||
|
||||
|
||||
return {
|
||||
'cpu_load': cpu_load,
|
||||
'memory_usage': memory_usage,
|
||||
'disk_read': disk_read,
|
||||
'disk_write': disk_write,
|
||||
'gpu_load': gpu_load,
|
||||
'gpu_memory': gpu_memory
|
||||
"cpu_load": cpu_load,
|
||||
"memory_usage": memory_usage,
|
||||
"disk_read": disk_read,
|
||||
"disk_write": disk_write,
|
||||
"gpu_load": gpu_load,
|
||||
"gpu_memory": gpu_memory,
|
||||
}
|
||||
|
||||
|
||||
def send_system_resources():
|
||||
last_disk_read = 0
|
||||
last_disk_write = 0
|
||||
while True:
|
||||
resources = get_system_resources()
|
||||
|
||||
|
||||
# Calculate disk I/O rates
|
||||
disk_read_rate = (resources['disk_read'] - last_disk_read) / UPDATE_INTERVAL
|
||||
disk_write_rate = (resources['disk_write'] - last_disk_write) / UPDATE_INTERVAL
|
||||
|
||||
socketio.emit('system_resources', {
|
||||
'cpu_load': resources['cpu_load'],
|
||||
'memory_usage': resources['memory_usage'],
|
||||
'disk_read_rate': disk_read_rate,
|
||||
'disk_write_rate': disk_write_rate,
|
||||
'gpu_load': resources['gpu_load'],
|
||||
'gpu_memory': resources['gpu_memory']
|
||||
})
|
||||
|
||||
last_disk_read = resources['disk_read']
|
||||
last_disk_write = resources['disk_write']
|
||||
disk_read_rate = (resources["disk_read"] - last_disk_read) / UPDATE_INTERVAL
|
||||
disk_write_rate = (resources["disk_write"] - last_disk_write) / UPDATE_INTERVAL
|
||||
|
||||
socketio.emit(
|
||||
"system_resources",
|
||||
{
|
||||
"cpu_load": resources["cpu_load"],
|
||||
"memory_usage": resources["memory_usage"],
|
||||
"disk_read_rate": disk_read_rate,
|
||||
"disk_write_rate": disk_write_rate,
|
||||
"gpu_load": resources["gpu_load"],
|
||||
"gpu_memory": resources["gpu_memory"],
|
||||
},
|
||||
)
|
||||
|
||||
last_disk_read = resources["disk_read"]
|
||||
last_disk_write = resources["disk_write"]
|
||||
time.sleep(UPDATE_INTERVAL)
|
||||
|
||||
|
||||
class QueryRequest(BaseModel):
|
||||
message: str
|
||||
|
||||
|
||||
class QueryResponse(BaseModel):
|
||||
query_id: str
|
||||
|
||||
|
||||
class QueryStatusResponse(BaseModel):
|
||||
status: str
|
||||
conversation_history: Optional[List[dict]]
|
||||
|
||||
|
||||
@app.post(
|
||||
"/api/v1/query",
|
||||
responses={
|
||||
"200": QueryResponse,
|
||||
"401": {"description": "Unauthorized"},
|
||||
"500": {"description": "Internal Server Error"},
|
||||
},
|
||||
)
|
||||
def api_query(body: QueryRequest):
|
||||
"""
|
||||
Submit a new query to the LLM Chat Server.
|
||||
|
||||
This endpoint requires authentication via an API key.
|
||||
|
||||
Sample cURL:
|
||||
curl -X POST http://localhost:5001/api/v1/query \
|
||||
-H "Content-Type: application/json" \
|
||||
-H "X-API-Key: your-api-key" \
|
||||
-d '{"message": "What is the capital of France?"}'
|
||||
"""
|
||||
if not ENABLE_API_ENDPOINTS:
|
||||
return jsonify({"error": "API endpoints are disabled"}), 404
|
||||
|
||||
api_key = request.headers.get("X-API-Key")
|
||||
if not api_key:
|
||||
return jsonify({"error": "API key is required"}), 401
|
||||
|
||||
api_key_id = validate_api_key(api_key)
|
||||
if not api_key_id:
|
||||
return jsonify({"error": "Invalid API key"}), 401
|
||||
|
||||
user_input = body.message
|
||||
query_id = str(uuid.uuid4())
|
||||
|
||||
try:
|
||||
db = get_db()
|
||||
cursor = db.cursor()
|
||||
cursor.execute(
|
||||
"INSERT INTO Queries (id, ip, query, api_key_id) VALUES (?, ?, ?, ?)",
|
||||
(query_id, request.remote_addr, user_input, api_key_id),
|
||||
)
|
||||
db.commit()
|
||||
|
||||
return jsonify({"query_id": query_id})
|
||||
except Exception as e:
|
||||
logger.exception("Error during API query processing", error=str(e))
|
||||
return jsonify({"error": str(e)}), 500
|
||||
|
||||
|
||||
@app.get(
|
||||
"/api/v1/query_status/<query_id>",
|
||||
responses={
|
||||
"200": QueryStatusResponse,
|
||||
"404": {"description": "Query not found"},
|
||||
"500": {"description": "Internal Server Error"},
|
||||
},
|
||||
)
|
||||
def get_query_status(query_id: str):
|
||||
"""
|
||||
Get the status of a submitted query.
|
||||
|
||||
This endpoint requires authentication via an API key.
|
||||
|
||||
Sample cURL:
|
||||
curl -X GET http://localhost:5001/api/v1/query_status/query-id-here \
|
||||
-H "X-API-Key: your-api-key"
|
||||
"""
|
||||
try:
|
||||
db = get_db()
|
||||
cursor = db.cursor()
|
||||
cursor.execute(
|
||||
"SELECT conversation_history FROM Queries WHERE id = ?", (query_id,)
|
||||
)
|
||||
result = cursor.fetchone()
|
||||
|
||||
if result is None:
|
||||
return jsonify({"error": "Query not found"}), 404
|
||||
|
||||
conversation_history = result[0]
|
||||
|
||||
if conversation_history is None:
|
||||
return jsonify({"status": "processing"}), 202
|
||||
else:
|
||||
return jsonify(
|
||||
{
|
||||
"status": "completed",
|
||||
"conversation_history": json.loads(conversation_history),
|
||||
}
|
||||
)
|
||||
except Exception as e:
|
||||
logger.exception("Error retrieving query status", error=str(e))
|
||||
return jsonify({"error": str(e)}), 500
|
||||
|
||||
|
||||
def answer_question_tools_api(
|
||||
user_input: str, conversation_history: List[dict], max_retries: int = 100
|
||||
):
|
||||
global tool_manager
|
||||
|
||||
if not conversation_history:
|
||||
conversation_history = [
|
||||
{"role": "system", "content": ANSWER_QUESTION_PROMPT},
|
||||
]
|
||||
|
||||
logger.info(
|
||||
"Starting API chat",
|
||||
user_input=user_input,
|
||||
conversation_history=conversation_history,
|
||||
)
|
||||
conversation_history.append({"role": "user", "content": user_input})
|
||||
|
||||
last_thought_content = None
|
||||
|
||||
for _ in range(max_retries):
|
||||
response = ollama.chat(
|
||||
model=PRIMARY_MODEL,
|
||||
messages=conversation_history,
|
||||
tools=tool_manager.get_tools_for_ollama_dict(),
|
||||
stream=False,
|
||||
)
|
||||
assistant_message = response["message"]
|
||||
|
||||
conversation_history.append(assistant_message)
|
||||
|
||||
if "tool_calls" in assistant_message:
|
||||
for tool_call in assistant_message["tool_calls"]:
|
||||
tool_name = tool_call["function"]["name"]
|
||||
tool_args = tool_call["function"]["arguments"]
|
||||
tool_response = tool_manager.get_tool(tool_name).execute(tool_args)
|
||||
conversation_history.append({"role": "tool", "content": tool_response})
|
||||
else:
|
||||
if "<reply>" in assistant_message["content"].lower():
|
||||
reply_content = re.search(
|
||||
r"<reply>(.*?)</reply>", assistant_message["content"], re.DOTALL
|
||||
)
|
||||
if reply_content:
|
||||
reply_answer = reply_content.group(1).strip()
|
||||
conversation_history.append(
|
||||
{"role": "assistant", "content": reply_answer}
|
||||
)
|
||||
return conversation_history
|
||||
else:
|
||||
current_thought_content = assistant_message["content"].strip()
|
||||
|
||||
if last_thought_content and not current_thought_content:
|
||||
conversation_history.append(
|
||||
{"role": "assistant", "content": last_thought_content}
|
||||
)
|
||||
return conversation_history
|
||||
|
||||
last_thought_content = current_thought_content
|
||||
continue
|
||||
|
||||
conversation_history.append(
|
||||
{
|
||||
"role": "assistant",
|
||||
"content": f"Max iterations reached. Last response: {assistant_message['content']}",
|
||||
}
|
||||
)
|
||||
return conversation_history
|
||||
|
||||
|
||||
def process_queries():
|
||||
with app.app_context():
|
||||
while True:
|
||||
try:
|
||||
db = get_db()
|
||||
cursor = db.cursor()
|
||||
cursor.execute(
|
||||
"SELECT id, query FROM Queries WHERE conversation_history IS NULL ORDER BY timestamp ASC LIMIT 1"
|
||||
)
|
||||
result = cursor.fetchone()
|
||||
|
||||
if result:
|
||||
query_id, user_input = result
|
||||
conversation_history = [
|
||||
{"role": "system", "content": ANSWER_QUESTION_PROMPT}
|
||||
]
|
||||
final_conversation_history = answer_question_tools_api(
|
||||
user_input, conversation_history
|
||||
)
|
||||
|
||||
cursor.execute(
|
||||
"UPDATE Queries SET conversation_history = ? WHERE id = ?",
|
||||
(json.dumps(final_conversation_history), query_id),
|
||||
)
|
||||
db.commit()
|
||||
else:
|
||||
time.sleep(
|
||||
1
|
||||
) # Wait for 1 second before checking again if no queries are found
|
||||
except Exception as e:
|
||||
logger.exception("Error processing query", error=str(e))
|
||||
time.sleep(1) # Wait for 1 second before retrying in case of an error
|
||||
|
||||
|
||||
# Admin endpoint for generating API keys
|
||||
class GenerateKeyRequest(BaseModel):
|
||||
username: str
|
||||
|
||||
|
||||
class GenerateKeyResponse(BaseModel):
|
||||
username: str
|
||||
api_key: str
|
||||
|
||||
|
||||
@app.post(
|
||||
"/admin/generate_key",
|
||||
responses={
|
||||
"200": GenerateKeyResponse,
|
||||
"401": {"description": "Unauthorized"},
|
||||
"500": {"description": "Internal Server Error"},
|
||||
},
|
||||
)
|
||||
def generate_api_key(body: GenerateKeyRequest):
|
||||
"""
|
||||
Generate a new API key for a user.
|
||||
|
||||
This endpoint requires authentication via an admin key.
|
||||
|
||||
Sample cURL:
|
||||
curl -X POST http://localhost:5001/admin/generate_key \
|
||||
-H "Content-Type: application/json" \
|
||||
-H "X-Admin-Key: your-admin-key" \
|
||||
-d '{"username": "new_user"}'
|
||||
"""
|
||||
admin_key = request.headers.get("X-Admin-Key")
|
||||
if not admin_key or admin_key != ADMIN_KEY:
|
||||
return jsonify({"error": "Invalid admin key"}), 401
|
||||
|
||||
username = body.username
|
||||
api_key = secrets.token_urlsafe(32)
|
||||
|
||||
try:
|
||||
db = get_db()
|
||||
cursor = db.cursor()
|
||||
cursor.execute(
|
||||
"INSERT INTO Keys (username, api_key) VALUES (?, ?)", (username, api_key)
|
||||
)
|
||||
db.commit()
|
||||
return jsonify({"username": username, "api_key": api_key})
|
||||
except sqlite3.IntegrityError:
|
||||
return jsonify({"error": "Username already exists"}), 400
|
||||
except Exception as e:
|
||||
logger.exception("Error generating API key", error=str(e))
|
||||
return jsonify({"error": str(e)}), 500
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
logger.info("Starting LLM Chat Server")
|
||||
threading.Thread(target=send_system_resources, daemon=True).start()
|
||||
init_db() # Initialize the database
|
||||
|
||||
if ENABLE_FRONTEND or ENABLE_CHAT_ENDPOINTS:
|
||||
threading.Thread(target=send_system_resources, daemon=True).start()
|
||||
|
||||
if ENABLE_API_ENDPOINTS:
|
||||
threading.Thread(
|
||||
target=lambda: app.app_context().push() and process_queries(), daemon=True
|
||||
).start()
|
||||
|
||||
socketio.run(app, debug=True, host="0.0.0.0", port=5001)
|
||||
95
models.py
95
models.py
@ -3,29 +3,86 @@ import structlog
|
||||
|
||||
logger = structlog.get_logger()
|
||||
|
||||
|
||||
class ModelManager:
|
||||
def __init__(self):
|
||||
self.model_capabilities = {
|
||||
"ajindal/llama3.1-storm:8b": ["general_knowledge", "reasoning", "tool_calling", "conversation", "multilingual", "instruction_following"],
|
||||
"llama3.1:8b": ["general_knowledge", "reasoning", "tool_calling", "conversation", "multilingual", "instruction_following"],
|
||||
"qwen2.5:7b": ["general_knowledge", "reasoning", "tool_calling", "conversation", "multilingual", "instruction_following"],
|
||||
"llama3.2:3b": ["summarization", "instruction_following", "tool_calling", "multilingual"],
|
||||
"llava:7b": ["visual_reasoning", "visual_conversation", "visual_tool_calling", "vision", "ocr", "multimodal"],
|
||||
"ajindal/llama3.1-storm:8b": [
|
||||
"general_knowledge",
|
||||
"reasoning",
|
||||
"tool_calling",
|
||||
"conversation",
|
||||
"multilingual",
|
||||
"instruction_following",
|
||||
],
|
||||
"llama3.1:8b": [
|
||||
"general_knowledge",
|
||||
"reasoning",
|
||||
"tool_calling",
|
||||
"conversation",
|
||||
"multilingual",
|
||||
"instruction_following",
|
||||
],
|
||||
"qwen2.5:7b": [
|
||||
"general_knowledge",
|
||||
"reasoning",
|
||||
"tool_calling",
|
||||
"conversation",
|
||||
"multilingual",
|
||||
"instruction_following",
|
||||
],
|
||||
"llama3.2:3b": [
|
||||
"summarization",
|
||||
"instruction_following",
|
||||
"tool_calling",
|
||||
"multilingual",
|
||||
],
|
||||
"llava:7b": [
|
||||
"visual_reasoning",
|
||||
"visual_conversation",
|
||||
"visual_tool_calling",
|
||||
"vision",
|
||||
"ocr",
|
||||
"multimodal",
|
||||
],
|
||||
}
|
||||
logger.info("ModelManager initialized", model_capabilities=self.model_capabilities)
|
||||
logger.info(
|
||||
"ModelManager initialized", model_capabilities=self.model_capabilities
|
||||
)
|
||||
|
||||
def get_model_capabilities(self, model_name):
|
||||
capabilities = self.model_capabilities.get(model_name, [])
|
||||
logger.debug("Retrieved model capabilities", model=model_name, capabilities=capabilities)
|
||||
logger.debug(
|
||||
"Retrieved model capabilities", model=model_name, capabilities=capabilities
|
||||
)
|
||||
return capabilities
|
||||
|
||||
def select_best_model(self, required_capability):
|
||||
suitable_models = [model for model, capabilities in self.model_capabilities.items() if required_capability in capabilities]
|
||||
selected_model = suitable_models[0] if suitable_models else list(self.model_capabilities.keys())[0]
|
||||
logger.info("Selected best model", required_capability=required_capability, selected_model=selected_model)
|
||||
suitable_models = [
|
||||
model
|
||||
for model, capabilities in self.model_capabilities.items()
|
||||
if required_capability in capabilities
|
||||
]
|
||||
selected_model = (
|
||||
suitable_models[0]
|
||||
if suitable_models
|
||||
else list(self.model_capabilities.keys())[0]
|
||||
)
|
||||
logger.info(
|
||||
"Selected best model",
|
||||
required_capability=required_capability,
|
||||
selected_model=selected_model,
|
||||
)
|
||||
return selected_model
|
||||
|
||||
def generate_text(self, model_name, prompt, max_length=100, system="You are a helpful assistant.", tools=[]):
|
||||
def generate_text(
|
||||
self,
|
||||
model_name,
|
||||
prompt,
|
||||
max_length=100,
|
||||
system="You are a helpful assistant.",
|
||||
tools=[],
|
||||
):
|
||||
# Check if model exists
|
||||
try:
|
||||
ollama.pull(model_name)
|
||||
@ -37,10 +94,16 @@ class ModelManager:
|
||||
else:
|
||||
logger.exception("Error pulling model", model=model_name, error=str(e))
|
||||
raise e
|
||||
|
||||
|
||||
response = ollama.generate(model=model_name, prompt=prompt, system=system, tools=tools, max_tokens=max_length)
|
||||
logger.debug("Text generated", model=model_name, response=response['response'])
|
||||
return response['response']
|
||||
response = ollama.generate(
|
||||
model=model_name,
|
||||
prompt=prompt,
|
||||
system=system,
|
||||
tools=tools,
|
||||
max_tokens=max_length,
|
||||
)
|
||||
logger.debug("Text generated", model=model_name, response=response["response"])
|
||||
return response["response"]
|
||||
|
||||
model_manager = ModelManager()
|
||||
|
||||
model_manager = ModelManager()
|
||||
|
||||
@ -1,13 +0,0 @@
|
||||
/** @type {import('next').NextConfig} */
|
||||
const nextConfig = {
|
||||
reactStrictMode: true,
|
||||
webpack: (config) => {
|
||||
config.externals.push({
|
||||
'utf-8-validate': 'commonjs utf-8-validate',
|
||||
'bufferutil': 'commonjs bufferutil',
|
||||
})
|
||||
return config
|
||||
},
|
||||
}
|
||||
|
||||
export default nextConfig;
|
||||
16
schema.sql
Normal file
16
schema.sql
Normal file
@ -0,0 +1,16 @@
|
||||
CREATE TABLE IF NOT EXISTS Keys (
|
||||
id INTEGER PRIMARY KEY AUTOINCREMENT,
|
||||
username TEXT NOT NULL UNIQUE,
|
||||
api_key TEXT NOT NULL UNIQUE
|
||||
);
|
||||
|
||||
CREATE TABLE IF NOT EXISTS Queries (
|
||||
id INTEGER PRIMARY KEY AUTOINCREMENT,
|
||||
request_id TEXT NOT NULL UNIQUE,
|
||||
ip TEXT NOT NULL,
|
||||
timestamp DATETIME DEFAULT CURRENT_TIMESTAMP,
|
||||
query TEXT NOT NULL,
|
||||
api_key_id INTEGER,
|
||||
conversation_history TEXT,
|
||||
FOREIGN KEY (api_key_id) REFERENCES Keys (id)
|
||||
);
|
||||
121
tools.py
121
tools.py
@ -1,10 +1,12 @@
|
||||
import subprocess
|
||||
import tempfile
|
||||
import time
|
||||
|
||||
import duckduckgo_search
|
||||
import requests
|
||||
from readability.readability import Document
|
||||
from markdownify import markdownify as md
|
||||
import subprocess
|
||||
import time
|
||||
import tempfile
|
||||
from readability.readability import Document
|
||||
|
||||
|
||||
class Tool:
|
||||
def __init__(self, name: str, description: str, arguments: dict, returns: str):
|
||||
@ -29,13 +31,23 @@ class ToolManager:
|
||||
if tool.name == name:
|
||||
return tool
|
||||
return None
|
||||
|
||||
|
||||
def get_tools_and_descriptions_for_prompt(self):
|
||||
return "\n".join([f"{tool.name}: {tool.description}" for tool in self.tools])
|
||||
|
||||
|
||||
def get_tools_for_ollama_dict(self):
|
||||
return [{'type': 'function', 'function': {'name': tool.name, 'description': tool.description, 'parameters': tool.arguments}} for tool in self.tools]
|
||||
|
||||
return [
|
||||
{
|
||||
"type": "function",
|
||||
"function": {
|
||||
"name": tool.name,
|
||||
"description": tool.description,
|
||||
"parameters": tool.arguments,
|
||||
},
|
||||
}
|
||||
for tool in self.tools
|
||||
]
|
||||
|
||||
|
||||
class DefaultToolManager(ToolManager):
|
||||
def __init__(self):
|
||||
@ -48,12 +60,22 @@ class DefaultToolManager(ToolManager):
|
||||
|
||||
class SearchTool(Tool):
|
||||
def __init__(self):
|
||||
super().__init__("search_web", "Search the internet for information", {'type': 'object', 'properties': {'query': {'type': 'string', 'description': 'The search query'}}}, "results:list[string]")
|
||||
super().__init__(
|
||||
"search_web",
|
||||
"Search the internet for information",
|
||||
{
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"query": {"type": "string", "description": "The search query"}
|
||||
},
|
||||
},
|
||||
"results:list[string]",
|
||||
)
|
||||
|
||||
def execute(self, arg: dict) -> str:
|
||||
res = duckduckgo_search.DDGS().text(arg['query'], max_results=5)
|
||||
return '\n\n'.join([f"{r['title']}\n{r['body']}\n{r['href']}" for r in res])
|
||||
|
||||
res = duckduckgo_search.DDGS().text(arg["query"], max_results=5)
|
||||
return "\n\n".join([f"{r['title']}\n{r['body']}\n{r['href']}" for r in res])
|
||||
|
||||
|
||||
def get_readable_page_contents(url: str) -> str:
|
||||
try:
|
||||
@ -64,20 +86,42 @@ def get_readable_page_contents(url: str) -> str:
|
||||
return md(content)
|
||||
except Exception as e:
|
||||
return f"Error fetching readable content: {str(e)}"
|
||||
|
||||
|
||||
|
||||
class GetReadablePageContentsTool(Tool):
|
||||
def __init__(self):
|
||||
super().__init__("get_readable_page_contents", "Get the contents of a web page in a readable format", {'type': 'object', 'properties': {'url': {'type': 'string', 'description': 'The url of the web page'}}}, "contents:string")
|
||||
super().__init__(
|
||||
"get_readable_page_contents",
|
||||
"Get the contents of a web page in a readable format",
|
||||
{
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"url": {"type": "string", "description": "The url of the web page"}
|
||||
},
|
||||
},
|
||||
"contents:string",
|
||||
)
|
||||
|
||||
def execute(self, arg: dict) -> str:
|
||||
return get_readable_page_contents(arg['url'])
|
||||
return get_readable_page_contents(arg["url"])
|
||||
|
||||
|
||||
class CalculatorTool(Tool):
|
||||
def __init__(self):
|
||||
super().__init__("calculator", "Perform a calculation using python's eval function", {'type': 'object', 'properties': {'expression': {'type': 'string', 'description': 'The mathematical expression to evaluate, should be a python mathematical expression'}}}, "result:string")
|
||||
super().__init__(
|
||||
"calculator",
|
||||
"Perform a calculation using python's eval function",
|
||||
{
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"expression": {
|
||||
"type": "string",
|
||||
"description": "The mathematical expression to evaluate, should be a python mathematical expression",
|
||||
}
|
||||
},
|
||||
},
|
||||
"result:string",
|
||||
)
|
||||
|
||||
def execute(self, arg: dict) -> str:
|
||||
try:
|
||||
@ -88,30 +132,45 @@ class CalculatorTool(Tool):
|
||||
|
||||
class PythonCodeTool(Tool):
|
||||
def __init__(self):
|
||||
super().__init__("python_code", "Execute python code",
|
||||
{'type': 'object', 'properties': {'code': {'type': 'string', 'description': 'The python code to execute, can be multiline'}}},
|
||||
"result:string")
|
||||
super().__init__(
|
||||
"python_code",
|
||||
"Execute python code using a temporary file and a subprocess. You must print results to stdout.",
|
||||
{
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"code": {
|
||||
"type": "string",
|
||||
"description": "The python code to execute, can be multiline",
|
||||
}
|
||||
},
|
||||
},
|
||||
"result:string",
|
||||
)
|
||||
|
||||
def execute(self, arg: dict) -> str:
|
||||
try:
|
||||
with tempfile.NamedTemporaryFile(suffix=".py", mode="w", delete=False) as temp_file:
|
||||
temp_file.write(arg['code'])
|
||||
with tempfile.NamedTemporaryFile(
|
||||
suffix=".py", mode="w", delete=False
|
||||
) as temp_file:
|
||||
temp_file.write(arg["code"])
|
||||
temp_file.flush()
|
||||
|
||||
start_time = time.time()
|
||||
process = subprocess.Popen(['python', temp_file.name],
|
||||
stdout=subprocess.PIPE,
|
||||
stderr=subprocess.PIPE,
|
||||
text=True)
|
||||
process = subprocess.Popen(
|
||||
["python", temp_file.name],
|
||||
stdout=subprocess.PIPE,
|
||||
stderr=subprocess.PIPE,
|
||||
text=True,
|
||||
)
|
||||
stdout, stderr = process.communicate(timeout=10) # 10 second timeout
|
||||
end_time = time.time()
|
||||
execution_time = end_time - start_time
|
||||
|
||||
result = {
|
||||
'stdout': stdout,
|
||||
'stderr': stderr,
|
||||
'return_value': process.returncode,
|
||||
'execution_time': execution_time
|
||||
"stdout": stdout,
|
||||
"stderr": stderr,
|
||||
"return_value": process.returncode,
|
||||
"execution_time": execution_time,
|
||||
}
|
||||
|
||||
except subprocess.TimeoutExpired:
|
||||
@ -119,5 +178,5 @@ class PythonCodeTool(Tool):
|
||||
return "Error: Code execution timed out after 10 seconds"
|
||||
except Exception as e:
|
||||
return f"Error executing code: {str(e)}"
|
||||
|
||||
return '\n'.join([f"{k}:\n{v}" for k, v in result.items()])
|
||||
|
||||
return "\n".join([f"{k}:\n{v}" for k, v in result.items()])
|
||||
|
||||
Loading…
Reference in New Issue
Block a user