feat: add rate limiter, create client

.env.template.sh

@@ -17,5 +17,8 @@ export DEEPGRAM_API_KEY=<your-deepgram-api-key>
 export GROQ_API_KEY=<your-groq-api-key>
 
 # Rate Limiter
-export RATE_LIMIT_MAX_REQUESTS=5
-export RATE_LIMIT_TIME_WINDOW=60
+export RATE_LIMIT_MAX_REQUESTS_RTC=5
+export RATE_LIMIT_TIME_WINDOW_RTC=60
+
+export RATE_LIMIT_MAX_REQUESTS_API=5
+export RATE_LIMIT_TIME_WINDOW_API=60

run.sh

@@ -0,0 +1,7 @@
+#!/bin/bash
+
+# Run Flask API server
+python -m src.main &
+
+# Run LiveKit agent
+python -c "from src.voice_agent import entrypoint, prewarm; from livekit.agents import cli, WorkerOptions; cli.run_app(WorkerOptions(entrypoint_fnc=entrypoint, prewarm_fnc=prewarm))" dev

src/__main__.py

@@ -1,18 +1,13 @@
 from flask import Flask
+from flask_cors import CORS
 from src.api.routes import api_bp
-from src.voice_agent import entrypoint, prewarm
-from livekit.agents import cli, WorkerOptions
-import threading
+from src.config import API_PORT, API_ORIGINS
 
 app = Flask(__name__)
 
 app.register_blueprint(api_bp, url_prefix='/api')
+CORS(app, origins=API_ORIGINS)
 
-def start_livekit_agent():
-    cli.run_app(WorkerOptions(entrypoint_fnc=entrypoint, prewarm_fnc=prewarm))
 
 if __name__ == "__main__":
-    agent_thread = threading.Thread(target=start_livekit_agent)
-    agent_thread.start()
-
-    app.run(host="0.0.0.0", port=5000)
+    app.run(host="0.0.0.0", port=API_PORT)

src/api/routes.py

@@ -1,11 +1,20 @@
-from flask import Blueprint, jsonify
+from flask import Blueprint, jsonify, request, abort
+from src.config import RATE_LIMIT_MAX_REQUESTS_API, RATE_LIMIT_TIME_WINDOW_API
 from src.token_service import create_access_token
 import random
+from src.rate_limiter import RateLimiter
 
 api_bp = Blueprint('api', __name__)
 
+# Initialize the rate limiter for IPs with desired limits
+rate_limiter = RateLimiter(max_requests=RATE_LIMIT_MAX_REQUESTS_API, time_window=RATE_LIMIT_TIME_WINDOW_API)  # 1 request per minute
+
 @api_bp.route("/getToken", methods=["GET"])
 def get_token():
+    ip_address = request.remote_addr
+    if not rate_limiter.is_allowed(ip_address):
+        abort(429, description="Too many requests from this IP. Please try again later.")
+
     participant_identity = f"user_{random.randint(1, 10000)}"
     room_name = "voice_assistant_room"
     participant_token = create_access_token(participant_identity, room_name)

src/config.py

@@ -6,11 +6,15 @@
 CARTESIA_API_KEY = os.getenv("CARTESIA_API_KEY")
 CARTESIA_API_URL = os.getenv("CARTESIA_API_URL", "https://api.cartesia.ai/voices")
 
-RATE_LIMIT_MAX_REQUESTS = int(os.getenv("RATE_LIMIT_MAX_REQUESTS", 5))
-RATE_LIMIT_TIME_WINDOW = int(os.getenv("RATE_LIMIT_TIME_WINDOW", 60))
+RATE_LIMIT_MAX_REQUESTS_RTC = int(os.getenv("RATE_LIMIT_MAX_REQUESTS_RTC", 5))
+RATE_LIMIT_TIME_WINDOW_RTC = int(os.getenv("RATE_LIMIT_TIME_WINDOW_RTC", 60))
+
+RATE_LIMIT_MAX_REQUESTS_API = int(os.getenv("RATE_LIMIT_MAX_REQUESTS_API", 5))
+RATE_LIMIT_TIME_WINDOW_API = int(os.getenv("RATE_LIMIT_TIME_WINDOW_API", 60))
 
 LIVEKIT_API_KEY = os.getenv("LIVEKIT_API_KEY")
 LIVEKIT_API_SECRET = os.getenv("LIVEKIT_API_SECRET")
 LIVEKIT_URL = os.getenv("LIVEKIT_URL")
 
 API_PORT = os.getenv("API_PORT", 5000)
+API_ORIGINS = os.getenv("API_ORIGINS")

src/voice_agent.py

@@ -8,10 +8,10 @@
 from livekit.agents.log import logger
 from livekit.plugins import silero, deepgram, openai, cartesia
 from src.rate_limiter import RateLimiter
-from src.config import CARTESIA_API_KEY, CARTESIA_API_URL
-from typing import List, Dict, Any
+from src.config import CARTESIA_API_KEY, CARTESIA_API_URL, RATE_LIMIT_MAX_REQUESTS_RTC, RATE_LIMIT_TIME_WINDOW_RTC
+from typing import List, Any
 
-rate_limiter = RateLimiter(max_requests=5, time_window=60)
+rate_limiter = RateLimiter(RATE_LIMIT_MAX_REQUESTS_RTC, RATE_LIMIT_TIME_WINDOW_RTC)
 
 def prewarm(proc: JobProcess):
     if not CARTESIA_API_KEY:
@@ -42,7 +42,7 @@ def update_tts_voice(tts, voice_data: dict):
 async def entrypoint(ctx: JobContext):
     try:
         initial_ctx = ChatContext(messages=[
-            ChatMessage(role="system", content="You are an AI voice agent tasked with generating a prompt for a specific role in a conversation.")
+            ChatMessage(role="system", content="You are a professional voice assistant designed to help with daily tasks efficiently and accurately. Your main purpose is to respond in a friendly but concise manner, providing relevant information, reminders, and updates as needed. ")
         ])
 
         cartesia_voices: List[dict[str, Any]] = ctx.proc.userdata.get("cartesia_voices")
@@ -69,17 +69,14 @@ async def entrypoint(ctx: JobContext):
         @ctx.room.on("participant_attributes_changed")
         def on_participant_attributes_changed(changed_attributes: dict[str, str], participant: rtc.Participant):
             user_id = participant.identity
-            if not rate_limiter.is_allowed(user_id):
-                logger.warning(f"Rate limit exceeded for user {user_id}")
-                asyncio.create_task(agent.say("You have exceeded the rate limit. Please try again later."))
-                return
-
+
             if participant.kind != rtc.ParticipantKind.PARTICIPANT_KIND_STANDARD:
                 return
 
             if "voice" in changed_attributes:
                 voice_id = participant.attributes.get("voice")
                 logger.info(f"Participant {participant.identity} requested voice change: {voice_id}")
+
                 if not voice_id:
                     return
 
@@ -93,6 +90,7 @@ def on_participant_attributes_changed(changed_attributes: dict[str, str], partic
                     if not (is_agent_speaking or is_user_speaking):
                         asyncio.create_task(agent.say("How do I sound now?", allow_interruptions=True))
 
+
         await ctx.connect()
 
         @agent.on("agent_started_speaking")
@@ -104,9 +102,18 @@ def agent_started_speaking():
         def agent_stopped_speaking():
             nonlocal is_agent_speaking
             is_agent_speaking = False
-
+        
         @agent.on("user_started_speaking")
         def user_started_speaking():
+            # Check if user exceeds the rate limit
+            logger.info(f"User {ctx.room.local_participant.identity} started speaking")
+            user_id = ctx.room.local_participant.identity
+            if not rate_limiter.is_allowed(user_id):
+                logger.warning(f"Rate limit exceeded for user {user_id}")
+
+                # Silently disconnect the participant
+                asyncio.create_task(ctx.room.disconnect())
+                return
             nonlocal is_user_speaking
             is_user_speaking = True
 
@@ -115,12 +122,14 @@ def user_stopped_speaking():
             nonlocal is_user_speaking
             is_user_speaking = False
 
+
+
         voices = [{"id": voice["id"], "name": voice["name"]} for voice in cartesia_voices]
         voices.sort(key=lambda x: x["name"])
         await ctx.room.local_participant.set_attributes({"voices": json.dumps(voices)})
 
         agent.start(ctx.room)
-        await agent.say("Thank you for calling Softcery Voice Agent", allow_interruptions=True)
+        await agent.say("Thank you for calling Softcery Voice Agent. How can I help you today?", allow_interruptions=True)
 
     except Exception as e:
         logger.error(f"An error occurred in the entrypoint: {e}")

web/.env.template.sh

@@ -0,0 +1,2 @@
+export VA_BACKEND_URL=<BACKEND_URL>
+export LIVEKIT_URL=<LIVEKIT_URL>

web/.gitignore

@@ -0,0 +1,24 @@
+# Logs
+logs
+*.log
+npm-debug.log*
+yarn-debug.log*
+yarn-error.log*
+pnpm-debug.log*
+lerna-debug.log*
+
+node_modules
+dist
+dist-ssr
+*.local
+.env.sh
+# Editor directories and files
+.vscode/*
+!.vscode/extensions.json
+.idea
+.DS_Store
+*.suo
+*.ntvs*
+*.njsproj
+*.sln
+*.sw?

web/eslint.config.js

@@ -0,0 +1,28 @@
+import js from '@eslint/js'
+import globals from 'globals'
+import reactHooks from 'eslint-plugin-react-hooks'
+import reactRefresh from 'eslint-plugin-react-refresh'
+import tseslint from 'typescript-eslint'
+
+export default tseslint.config(
+  { ignores: ['dist'] },
+  {
+    extends: [js.configs.recommended, ...tseslint.configs.recommended],
+    files: ['**/*.{ts,tsx}'],
+    languageOptions: {
+      ecmaVersion: 2020,
+      globals: globals.browser,
+    },
+    plugins: {
+      'react-hooks': reactHooks,
+      'react-refresh': reactRefresh,
+    },
+    rules: {
+      ...reactHooks.configs.recommended.rules,
+      'react-refresh/only-export-components': [
+        'warn',
+        { allowConstantExport: true },
+      ],
+    },
+  },
+)