fix: Update Dockerfile and application configuration for improved performance and compatibility

Dockerfile

@@ -0,0 +1,22 @@
+FROM python:3.10-slim
+
+WORKDIR /app
+ENV PYTHONDONTWRITEBYTECODE=1 \
+    PYTHONUNBUFFERED=1
+
+RUN apt-get update && apt-get install -y --no-install-recommends \
+        ffmpeg \
+        libsndfile1 \
+    && rm -rf /var/lib/apt/lists/*
+
+COPY requirements.txt .
+RUN pip install --no-cache-dir --upgrade pip \
+ && pip install --no-cache-dir -r requirements.txt
+
+COPY . .
+
+# Gradio defaults to 7860; change if you override `demo.launch(server_port=...)`
+ENV PORT=7860
+EXPOSE 7860
+
+CMD ["python", "app.py"]

app.py

@@ -15,7 +15,7 @@ import json
 from pydub import AudioSegment
 import ast
 from backend.utils import get_device
-if get_device() == "mps":
+if get_device() == "cpu":
     load_dotenv(override=True)
 
 
@@ -476,5 +476,11 @@ with gr.Blocks(css=custom_css, theme=gr.themes.Soft(primary_hue="orange", second
         show_progress="hidden"
     )
 
-demo.queue(default_concurrency_limit=10)
-demo.launch(debug=True, show_error=True, share=True)
+demo.queue(default_concurrency_limit=100)
+demo.launch(
+    debug=True,
+    show_error=True,
+    share=False,
+    server_name="0.0.0.0",
+    server_port=int(os.getenv("PORT", 7860)),
+)

backend/Dockerfile

@@ -1,44 +0,0 @@
-# Use an official Python runtime as a parent image
-FROM python:3.10-slim
-
-# Set the working directory in the container
-WORKDIR /app
-
-# Prevent python from writing pyc files to disc
-ENV PYTHONDONTWRITEBYTECODE 1
-# Ensure python output is sent straight to terminal (useful for logs)
-ENV PYTHONUNBUFFERED 1
-
-# Install system dependencies if any (e.g., if specific libraries needed them)
-# RUN apt-get update && apt-get install -y --no-install-recommends some-package && rm -rf /var/lib/apt/lists/*
-
-# Copy the requirements file into the container at /app
-COPY requirements.txt .
-# Install any needed packages specified in requirements.txt
-# Using --no-cache-dir to reduce image size
-RUN pip install --no-cache-dir --upgrade pip && \
-    pip install --no-cache-dir -r requirements.txt
-
-# Copy the rest of the backend application code into the container at /app
-COPY . .
-
-# Make port 8000 available to the world outside this container
-EXPOSE 8000
-
-# Define environment variables (placeholders, will be overridden by compose/env_file)
-ENV MONGO_URL=""
-ENV OPENAI_API_KEY=""
-ENV JAI_API_KEY=""
-ENV JAI_BASE_URL=""
-ENV TYPHOON_API_KEY=""
-ENV TYPHOON_BASE_URL=""
-ENV GEMMA_API_KEY=""
-ENV GEMMA_BASE_URL=""
-ENV LANGFUSE_SECRET_KEY=""
-ENV LANGFUSE_PUBLIC_KEY=""
-ENV LANGFUSE_HOST=""
-ENV PORT=8000
-ENV HOST=0.0.0.0
-ENV GEMINI_API_KEY=""
-# Run uvicorn server when the container launches
-CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "8000"]

backend/functions.py

@@ -14,7 +14,7 @@ from utils import get_device
 # from transformers import AutoTokenizer # No longer needed for reranker
 
 # Load environment variables
-if get_device() == "mps":
+if get_device() == "cpu":
     load_dotenv(override=True)
 
 # Set up logging
@@ -34,13 +34,13 @@ class MongoHybridSearch:
         Initialize MongoDB connection and embedder.
         """
         try:
-            self.client = AsyncIOMotorClient(mongo_uri)
-            self.database = self.client[database_name]
-            # Consider making collection name configurable
-            self.collection = self.database["homeshopping"]
-            # self.collection_fact = self.database["SCG_financial_report_jai"]
-            self.llm_analyzer = models.LLMFinanceAnalyzer()
-            self.embedder = models.Embedder() # Instantiate Embedder class from models
+            # self.client = AsyncIOMotorClient(mongo_uri)
+            # self.database = self.client[database_name]
+            # # Consider making collection name configurable
+            # self.collection = self.database["homeshopping"]
+            # # self.collection_fact = self.database["SCG_financial_report_jai"]
+            # self.llm_analyzer = models.LLMFinanceAnalyzer()
+            # self.embedder = models.Embedder() # Instantiate Embedder class from models
             logger.info("MongoHybridSearch initialized successfully.")
         except Exception as e:
             logger.error(f"Failed to initialize MongoHybridSearch: {e}")

backend/main.py

@@ -17,7 +17,7 @@ sys.path.append(os.path.abspath('./backend'))
 from models import LLMFinanceAnalyzer
 from functions import MongoHybridSearch
 from utils import get_device
-if get_device() == "mps":
+if get_device() == "cpu":
     load_dotenv(override=True)
 
 logging.basicConfig(level=logging.INFO, format="%(asctime)s - %(name)s - %(levelname)s - %(message)s")

backend/models.py

@@ -20,7 +20,7 @@ from systemprompt import (
     get_non_rag_prompt,
 )
 from utils import get_device
-if get_device() == "mps":
+if get_device() == "cpu":
     load_dotenv(override=True)
 logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(name)s - %(levelname)s - %(message)s')
 logger = logging.getLogger(__name__)

backend/tools.py

@@ -4,7 +4,7 @@ from dotenv import load_dotenv
 import os
 import requests
 from utils import get_device
-if get_device() == "mps":
+if get_device() == "cpu":
     load_dotenv(override=True)
 
 

backend/tts.py

@@ -67,6 +67,7 @@ def synthesize_text(text: str, lang = 'th' , speed = 2.0):
                 if first_chunk:
                     samples = samples[600:] # Optionally drop start of first chunk
                     first_chunk = False
+                
                 yield (24000, samples)
     except Exception as e:
         print(f"Error during TTS synthesis for text '{text}': {e}")

backend/utils.py

@@ -28,7 +28,7 @@ def get_device():
         return "mps"
     else:
         return "cpu"
-if get_device() == "mps":
+if get_device() == "cpu":
     load_dotenv(override=True)
 
 device = get_device()

requirements.txt

@@ -27,7 +27,7 @@ google-cloud-speech
 # sentence-transformers==3.4.1
 https://huggingface.co/jts-ai-team/modifiedfastrtc/resolve/main/fastrtc-0.0.24-py3-none-any.whl
 onnxruntime==1.21.0
-gradio==5.23.2
+gradio==5.44.1
 
 
 nemo_toolkit==2.4.0