Refactor code structure for improved readability and maintainability

addavid888 · addavid888 · commit bdc8c966e4ec · 2025-12-04T20:56:32.000+07:00
diff --git a/content/2-Proposal/_index.md b/content/2-Proposal/_index.md
@@ -6,84 +6,94 @@ chapter: false
 pre: " <b> 2. </b> "
 ---
 
-# Rafilm: AI-Powered Movie Logging & Recommendation Platform  
-## A Serverless AWS Solution for Intelligent Movie Discovery  
+# Rafilm: AI-Powered Movie Logging & Recommendation Platform
 
-### 1. Executive Summary  
-Rafilm is a Letterboxd-inspired movie logging and recommendation platform designed to help general users track their watched films, share reviews, and discover new favorites through AI-powered recommendations. Built as part of the AWS First Cloud Journey (FCJ) internship, Rafilm integrates Amazon Personalize and Bedrock to deliver tailored movie suggestions and conversational recommendations via a chatbot interface.  
+## A Serverless AWS Solution for Intelligent Movie Discovery
 
-The platform runs fully on AWS Serverless architecture, featuring Amplify-hosted Next.js frontend, Lambda-based backend services connected through API Gateway, and DynamoDB for scalable user and movie data storage. TMDb provides external movie data integration, while Amazon Cognito manages user authentication. Rafilm aims to demonstrate a scalable, intelligent, and cost-efficient architecture capable of supporting multi-user access and interactive experiences.  
+### 1. Executive Summary
 
-### 2. Problem Statement  
+Rafilm is a Letterboxd-inspired movie logging and recommendation platform designed to help general users track their watched films, share reviews, and discover new favorites through AI-powered recommendations. Built as part of the AWS First Cloud Journey (FCJ) internship, Rafilm integrates Amazon Personalize and Bedrock to deliver tailored movie suggestions and conversational recommendations via a chatbot interface.
 
-#### What’s the Problem?  
-While existing movie platforms like Letterboxd and IMDb offer robust logging and social features, they lack **personalized recommendation systems** and **interactive discovery experiences**. Users often rely on external sources or generic trending lists to find what to watch next, leading to irrelevant or repetitive suggestions.  
+The platform runs fully on AWS Serverless architecture, featuring Amplify-hosted Next.js frontend, Lambda-based backend services connected through API Gateway, and DynamoDB for scalable user and movie data storage. TMDb provides external movie data integration, while Amazon Cognito manages user authentication. Rafilm aims to demonstrate a scalable, intelligent, and cost-efficient architecture capable of supporting multi-user access and interactive experiences.
 
-#### The Solution  
-Rafilm integrates a **custom recommendation pipeline** powered by **Amazon Personalize**, combined with a **Bedrock LLM chatbot** that interprets user preferences and generates conversational movie recommendations. Users can log movies, write reviews, and receive curated suggestions—all within one seamless experience. Unlike Letterboxd, Rafilm focuses on data-driven personalization and AI-assisted interaction rather than pure social networking.  
+### 2. Problem Statement
 
-#### Benefits and Return on Investment  
-By leveraging AWS Serverless services, Rafilm achieves near-zero maintenance cost, pay-per-use scalability, and real-time AI-driven personalization. For the FCJ internship, the project serves as both a **technical showcase** and a **learning artifact** for integrating AI services in serverless architectures. Projected cost remains under $1/month during testing, with AWS Free Tier coverage for most usage.  
+#### What’s the Problem?
 
-### 3. Solution Architecture  
+While existing movie platforms like Letterboxd and IMDb offer robust logging and social features, they lack **personalized recommendation systems** and **interactive discovery experiences**. Users often rely on external sources or generic trending lists to find what to watch next, leading to irrelevant or repetitive suggestions.
+
+#### The Solution
+
+Rafilm integrates a **custom recommendation pipeline** powered by **Amazon Personalize**, combined with a **Bedrock LLM chatbot** that interprets user preferences and generates conversational movie recommendations. Users can log movies, write reviews, and receive curated suggestions—all within one seamless experience. Unlike Letterboxd, Rafilm focuses on data-driven personalization and AI-assisted interaction rather than pure social networking.
+
+#### Benefits and Return on Investment
+
+By leveraging AWS Serverless services, Rafilm achieves near-zero maintenance cost, pay-per-use scalability, and real-time AI-driven personalization. For the FCJ internship, the project serves as both a **technical showcase** and a **learning artifact** for integrating AI services in serverless architectures. Projected cost remains under $1/month during testing, with AWS Free Tier coverage for most usage.
+
+### 3. Solution Architecture
 
 Rafilm employs a modular serverless architecture using AWS services for scalability, integration, and cost optimization.
 
-![architecture-diagram](/images/2-Proposal/platform_architecture.jpeg)
-
-#### AWS Services Used  
-- **AWS Amplify**: Hosts the Next.js frontend for movie browsing, logging, and chatbot interaction.  
-- **Amazon Cognito**: Handles user registration, login, and session management.  
-- **Amazon API Gateway**: Routes client requests to backend Lambda functions.  
-- **AWS Lambda**: Executes serverless business logic (e.g., CRUD for reviews, fetching TMDb data, triggering recommendations).  
-- **Amazon DynamoDB**: Stores user logs, movie interactions, and preferences.  
-- **Amazon Personalize**: Trains and serves personalized recommendation models.  
-- **Amazon Bedrock**: Provides conversational chatbot functionality for recommendation dialogue.  
-- **Amazon S3**: Stores static assets and backups for logs and model outputs.  
-
-#### Component Design  
-- **Frontend (Next.js)**: User-friendly interface for movie discovery, logging, and chat-based recommendations.  
-- **Backend (Lambda + API Gateway)**: Stateless logic layer handling user operations, movie fetching, and recommendation retrieval.  
-- **Data Layer (DynamoDB + S3)**: Stores structured user interactions and movie metadata for model training.  
-- **AI Layer (Personalize + Bedrock)**: Personalize analyzes historical user interactions; Bedrock chatbot provides natural language access to personalized results.  
-- **Authentication (Cognito)**: Securely manages multi-user access.  
-
-#### Architecture Overview  
-1. Users log in via Cognito and interact with the Next.js interface.  
-2. Actions such as logging or rating trigger API Gateway → Lambda → DynamoDB workflows.  
-3. The Bedrock chatbot accesses Personalize results to generate conversational movie suggestions.  
-4. Amplify hosts the frontend for seamless deployment and scalability.  
-
-### 4. Technical Implementation  
-
-#### Implementation Phases  
-1. **Architecture Design (Month 1):** Research AWS serverless and AI integration patterns; finalize architecture diagrams.  
-2. **Prototype Integration (Month 2):** Implement Amplify hosting, Cognito authentication, and Lambda-based backend APIs.  
-3. **Recommendation System (Month 3):** Connect Personalize and Bedrock for end-to-end AI recommendation and chatbot response.  
-4. **Testing & Deployment:** Conduct functional testing, optimize costs, and deploy production-ready version on Amplify.  
-
-#### Technical Requirements  
-- **Frontend:** Next.js + React hosted via AWS Amplify, using TMDb API for movie data.  
-- **Backend:** AWS Lambda (Node.js runtime) connected through API Gateway.  
-- **Database:** Amazon DynamoDB for scalable user and review data.  
-- **AI Components:** Amazon Personalize (user-item recommendations) and Bedrock (chatbot dialogue).  
-- **Authentication:** Amazon Cognito for secure, multi-user access.  
-- **Automation:** AWS SDK & CloudFormation for provisioning; AWS SAM for deployment workflows.  
-
-### 5. Timeline & Milestones  
-
-| Phase | Duration | Key Deliverables |
-|-------|-----------|------------------|
-| Month 1 | Research & Architecture | AWS design finalizing |
-| Month 2 | Core Development | Amplify hosting, Cognito setup, Lambda API, DynamoDB schema |
-| Month 3 | AI Integration & Testing | Personalize training, Bedrock chatbot, system deployment |
-| Post-Launch | Continuous Improvement | Cost optimization, new features, UX refinement |
-
-### 6. Budget Estimation  
-
-**Estimated Monthly Cost (AWS Free Tier eligible):**  
-
-*To be updated*
+![architecture-diagram](/images/2-Proposal/solution-architect-rafilm.jpg)
+
+#### AWS Services Used
+
+- **AWS Amplify**: Hosts the Next.js frontend for movie browsing, logging, and chatbot interaction.
+- **Amazon Cognito**: Handles user registration, login, and session management.
+- **Amazon API Gateway**: Routes client requests to backend Lambda functions.
+- **AWS Lambda**: Executes serverless business logic (e.g., CRUD for reviews, fetching TMDb data, triggering recommendations).
+- **Amazon DynamoDB**: Stores user logs, movie interactions, and preferences.
+- **Amazon Personalize**: Trains and serves personalized recommendation models.
+- **Amazon Bedrock**: Provides conversational chatbot functionality for recommendation dialogue.
+- **Amazon S3**: Stores static assets and backups for logs and model outputs.
+
+#### Component Design
+
+- **Frontend (Next.js)**: User-friendly interface for movie discovery, logging, and chat-based recommendations.
+- **Backend (Lambda + API Gateway)**: Stateless logic layer handling user operations, movie fetching, and recommendation retrieval.
+- **Data Layer (DynamoDB + S3)**: Stores structured user interactions and movie metadata for model training.
+- **AI Layer (Personalize + Bedrock)**: Personalize analyzes historical user interactions; Bedrock chatbot provides natural language access to personalized results.
+- **Authentication (Cognito)**: Securely manages multi-user access.
+
+#### Architecture Overview
+
+1. Users log in via Cognito and interact with the Next.js interface.
+2. Actions such as logging or rating trigger API Gateway → Lambda → DynamoDB workflows.
+3. The Bedrock chatbot accesses Personalize results to generate conversational movie suggestions.
+4. Amplify hosts the frontend for seamless deployment and scalability.
+
+### 4. Technical Implementation
+
+#### Implementation Phases
+
+1. **Architecture Design (Month 1):** Research AWS serverless and AI integration patterns; finalize architecture diagrams.
+2. **Prototype Integration (Month 2):** Implement Amplify hosting, Cognito authentication, and Lambda-based backend APIs.
+3. **Recommendation System (Month 3):** Connect Personalize and Bedrock for end-to-end AI recommendation and chatbot response.
+4. **Testing & Deployment:** Conduct functional testing, optimize costs, and deploy production-ready version on Amplify.
+
+#### Technical Requirements
+
+- **Frontend:** Next.js + React hosted via AWS Amplify, using TMDb API for movie data.
+- **Backend:** AWS Lambda (Node.js runtime) connected through API Gateway.
+- **Database:** Amazon DynamoDB for scalable user and review data.
+- **AI Components:** Amazon Personalize (user-item recommendations) and Bedrock (chatbot dialogue).
+- **Authentication:** Amazon Cognito for secure, multi-user access.
+- **Automation:** AWS SDK & CloudFormation for provisioning; AWS SAM for deployment workflows.
+
+### 5. Timeline & Milestones
+
+| Phase       | Duration                 | Key Deliverables                                            |
+| ----------- | ------------------------ | ----------------------------------------------------------- |
+| Month 1     | Research & Architecture  | AWS design finalizing                                       |
+| Month 2     | Core Development         | Amplify hosting, Cognito setup, Lambda API, DynamoDB schema |
+| Month 3     | AI Integration & Testing | Personalize training, Bedrock chatbot, system deployment    |
+| Post-Launch | Continuous Improvement   | Cost optimization, new features, UX refinement              |
+
+### 6. Budget Estimation
+
+**Estimated Monthly Cost:**
+
+_To be updated_
 
 <!-- | Service | Usage | Monthly Cost |
 |----------|--------|--------------|
@@ -99,22 +109,23 @@ Rafilm employs a modular serverless architecture using AWS services for scalabil
 
 <!-- **Total Estimated Cost:** ≈ **$0.7/month** (≈ **$8.40/year**)   -->
 
-### 7. Risk Assessment  
+### 7. Risk Assessment
+
+| Risk                           | Probability | Impact | Mitigation                                    |
+| ------------------------------ | ----------- | ------ | --------------------------------------------- |
+| API rate limits from TMDb      | Medium      | Medium | Cache requests via Lambda                     |
+| Model training cost escalation | Low         | Medium | Use limited training dataset for testing      |
+| Chatbot latency                | Medium      | Low    | Optimize Bedrock model type and response size |
+| Authentication or token expiry | Medium      | Low    | Use short-lived JWTs and refresh tokens       |
+
+### 8. Expected Outcomes
 
-| Risk | Probability | Impact | Mitigation |
-|------|--------------|---------|------------|
-| API rate limits from TMDb | Medium | Medium | Cache requests via Lambda |
-| Model training cost escalation | Low | Medium | Use limited training dataset for testing |
-| Chatbot latency | Medium | Low | Optimize Bedrock model type and response size |
-| Authentication or token expiry | Medium | Low | Use short-lived JWTs and refresh tokens |
-| Data loss | Low | High | Backup DynamoDB to S3 regularly |
+#### Technical Improvements
 
-### 8. Expected Outcomes  
+- Demonstrates **serverless integration** of AI/ML and LLM services in real-world use.
+- Establishes a reusable AWS architecture for recommendation-based apps.
 
-#### Technical Improvements  
-- Demonstrates **serverless integration** of AI/ML and LLM services in real-world use.  
-- Establishes a reusable AWS architecture for recommendation-based apps.  
+#### Long-Term Value
 
-#### Long-Term Value  
-- Provides a foundation for future expansion into a **social movie discovery network**.  
-- Serves as an **AWS FCJ internship showcase project** highlighting scalability, personalization, and conversational AI.  
+- Provides a foundation for future expansion into a **social movie discovery network**.
+- Serves as an **AWS FCJ internship showcase project** highlighting scalability, personalization, and conversational AI.
diff --git a/content/2-Proposal/_index.vi.md b/content/2-Proposal/_index.vi.md
@@ -29,7 +29,7 @@ Bằng cách tận dụng dịch vụ Serverless của AWS, Rafilm đạt chi ph
 
 Rafilm sử dụng kiến trúc serverless mô-đun trên AWS nhằm đạt khả năng mở rộng, tích hợp và tối ưu chi phí.
 
-![sơ-đồ-kiến-trúc](/images/2-Proposal/platform_architecture.jpeg)
+![architecture-diagram](/images/2-Proposal/solution-architect-rafilm.jpg)
 
 #### Dịch vụ AWS sử dụng  
 - **AWS Amplify**: Host frontend Next.js cho việc duyệt phim, ghi nhật ký và tương tác chatbot.  
@@ -107,7 +107,6 @@ Rafilm sử dụng kiến trúc serverless mô-đun trên AWS nhằm đạt kh
 | Chi phí huấn luyện mô hình tăng | Thấp | Trung bình | Dùng tập dữ liệu giới hạn cho thử nghiệm |
 | Độ trễ của chatbot | Trung bình | Thấp | Tối ưu loại mô hình Bedrock và kích thước phản hồi |
 | Xác thực hoặc hết hạn token | Trung bình | Thấp | Dùng JWT thời gian sống ngắn và refresh token |
-| Mất dữ liệu | Thấp | Cao | Sao lưu DynamoDB sang S3 định kỳ |
 
 ### 8. Kết quả mong đợi  
 
diff --git a/static/images/2-Proposal/solution-architect-rafilm.jpg b/static/images/2-Proposal/solution-architect-rafilm.jpg
