FastAPI Performance Optimization: A Guide to Efficiency Improvement from Code to Deployment
To optimize the performance of FastAPI, it is necessary to systematically advance from five aspects: code, asynchronous processing, database, caching, and deployment. At the code level: prioritize using `async def` to handle I/O-intensive tasks (such as asynchronous database queries), use generators or pagination to reduce memory usage, and leverage parameter validation to filter invalid requests. For asynchronous programming, distinguish task types: use asynchronous frameworks for I/O-intensive tasks, and submit CPU-intensive tasks to a thread pool via `ThreadPoolExecutor`. The core of database optimization includes connection pool reuse, index optimization (to avoid full table scans), batch operations (e.g., `bulk_insert`), and lazy loading. Caching strategies are suitable for frequently accessed data: use in-memory caching with `cachetools` for simple scenarios, and Redis distributed caching for multi-instance deployments. At the deployment end, use Gunicorn + Uvicorn for multi-process/threading, Nginx reverse proxy for static resources, and containerization (Docker) with K8s for elastic scaling. Optimization should first identify bottlenecks, advance step-by-step from code to deployment, prioritize high-cost-effectiveness issues (such as indexes and caching), and continuously monitor and iterate.
Read MoreAsynchronous Dependency Injection in FastAPI: Techniques for Managing Asynchronous Tasks' Dependencies
FastAPI's Dependency Injection (DI) is a core tool for managing resource sharing and reuse, especially in asynchronous scenarios, avoiding code duplication and coupling. At its core, dependencies are declared using `Depends()`, where functions only need to declare the required resources, and resource acquisition is handled externally. At a basic level, synchronous dependencies use regular functions (e.g., `get_sync_db`), while asynchronous dependencies use `async def` (e.g., `get_async_db`), with FastAPI automatically handling `await` calls. For example, the asynchronous route function `read_users` injects an asynchronous database connection via `db=Depends(get_async_db)`. Advanced techniques include nested dependencies (e.g., combining authentication dependencies with database dependencies) and passing dependencies in background tasks. Pitfalls to watch for include forgetting `await`, cyclic dependencies, and type mismatches. Mastering these concepts enables efficient construction of decoupled, scalable asynchronous applications, enhancing development efficiency through rational resource reuse.
Read MoreFastAPI Query Parameters: How to Implement Parameter Filtering with Query and Path
In FastAPI, parameter handling is a core aspect. Query parameters (after the question mark in the URL) and path parameters (within the URL path) need to be processed using the `Query` and `Path` utilities, respectively. Query parameters can have default values (e.g., setting `age` to default to 18), be marked as required (using `...`), and apply validation rules such as `min_length` and `gt` (greater than), which restrict string lengths or numeric ranges. For example, `Query(..., min_length=5, max_length=50)` ensures a string parameter meets length constraints. Path parameters are handled with `Path` for validation, such as ensuring `user_id` is a positive integer (`gt=0`). Both utilities support type conversion, range filtering, and automatically generate Swagger documentation. `Query` is primarily for optional parameters (e.g., setting `name` to default to `None`) and required validation, while `Path` focuses on type validation for path parameters (e.g., integers). Proper use of these utilities enhances interface robustness, reduces invalid data, and simplifies parameter processing logic.
Read MoreAdvanced FastAPI Path Parameters: Dynamic Routing and Parameter Validation
FastAPI supports dynamic routing and parameter validation with flexibility and robustness. Basic usage includes paths like `/users/{user_id}`, where parameters can be automatically type-identified (e.g., `int`), and conversion failures return a 422 error. For advanced dynamic routing, it supports automatic type conversion, optional parameters (`Optional` with default values), and regex restrictions (`Path.pattern`), such as for order codes requiring an 8-character combination of uppercase letters/numbers (`^[A-Z0-9]{8}$`). Advanced parameter validation is achieved by setting ranges (`ge`/`le`) via `Path` or using enumeration types, e.g., product IDs must be `ge=1, le=99`, and order types restricted to the enum values `pending/completed/cancelled`. By combining dynamic routing with validation, a general interface is constructed, reducing manual validation code. The Swagger documentation (`/docs`) allows intuitive testing of these rules.
Read MoreFastAPI + SQLite: Quickly Build a Lightweight Database API Service
This article introduces the process of quickly building a "Student Information Management" database API service using FastAPI and SQLite. First, dependencies such as FastAPI, Uvicorn, and SQLAlchemy are installed via `pip`. SQLAlchemy's ORM is used to define student data models (including id, name, and age fields) and session management, with Pydantic models for data validation. The core implementation includes CRUD operations (create, read single/all students, update, delete). FastAPI routes bind HTTP methods (POST/GET/PUT/DELETE) to generate student management API endpoints. The database configuration uses SQLite, an embedded database that requires no additional server, with data stored in the `test.db` file. After starting the service with Uvicorn, FastAPI automatically generates Swagger UI documentation for easy testing. This lightweight and easy-to-use solution supports asynchronous operations, making it suitable for small-to-medium projects. Future expansions include multi-table associations or migration to PostgreSQL/MySQL.
Read MoreFastAPI + Frontend Interaction: Practical JavaScript Call to FastAPI API
This article introduces the interaction methods between a FastAPI backend and frontend JavaScript. The core principle is that the frontend calls the backend API through HTTP requests, and the backend returns JSON data after processing, which the frontend then renders and displays. Prerequisites: The backend needs to install FastAPI and uvicorn, while the frontend only requires HTML + JS. The backend writes main.py to implement three interfaces: GET (/api/hello) returns a message, GET (/api/items/{item_id}) with parameters returns product information, and POST (/api/submit) receives data and provides feedback. CORSMiddleware is configured to handle cross-origin issues (allowing all origins during development and specifying domains in production). On the frontend, the fetch API is used to call the interfaces. Three buttons respectively trigger the requests, parse the JSON, and display the results. During runtime, start the backend service and open the frontend page to test. Key knowledge points include cross-origin configuration, HTTP methods (GET/POST), JSON data exchange, and error handling. For advanced exploration, one can look into Axios, frontend frameworks, and data validation.
Read MoreFastAPI Deployment Guide: A Comprehensive Process from Local Development to Cloud Server Deployment
This article introduces the complete deployment process of FastAPI, from local development to cloud server deployment. First, install FastAPI and Uvicorn locally, write a simple interface (e.g., `main.py`), and test it with `uvicorn`. Next, purchase a Linux cloud server (e.g., Ubuntu), obtain information such as the IP and username, and connect remotely via SSH. The server needs to install Python 3 and dependencies, create a project directory, upload the code, generate `requirements.txt` and install dependencies. For production environment configuration, set up a systemd service, enable auto-start at boot (`fastapi.service`), and open the firewall port 8000. Nginx reverse proxy is recommended, with HTTPS configured through Certbot. After deployment, maintain the service via logs, and re-upload and restart the code when updated. For complex projects, Docker containerization deployment can be adopted. The core process: local debugging → server preparation → environment setup → service startup → security configuration → maintenance, ensuring the API stably serves external requests.
Read MoreFastAPI + Pydantic: Best Practices for Data Model Definition and Serialization
Combining FastAPI with Pydantic is an efficient combination for data processing in modern web development, where Pydantic focuses on data validation and serialization, and FastAPI provides high performance, automatic documentation, and asynchronous support. Base models are defined by inheriting from `BaseModel`, with field types specified via Python annotations. Fields without default values are required, while optional types are indicated using `| None` or `Optional`. Pydantic automatically validates types and formats, throwing detailed error messages for input mistakes. It also supports `Field` for custom constraints (e.g., length, range, regex). Models can be bidirectionally converted with dictionaries/JSON. In FastAPI, they can be directly used as request/response bodies, automatically validating request data and returning structured responses. Best practices include: using consistent naming styles for field aliases, handling complex structures with nested models, reusing code via model inheritance, and `extra="ignore"` to disregard unknown fields. Mastering these techniques enables robust data processing, reduces repetitive code, and enhances API reliability, making it suitable for quickly building efficient, type-safe web services.
Read MorePractical Guide to FastAPI Middleware: Implementing Request Logging and Response Time Statistics
This article introduces the use of FastAPI middleware for uniformly handling request/response logic (such as logging and authentication) to avoid repetitive code. First, FastAPI and Uvicorn need to be installed. The core is to inherit Starlette's BaseHTTPMiddleware and implement the dispatch method to handle middleware logic: record the request start time, call call_next to obtain the response, calculate the response time, construct and output logs containing method, path, IP, response time, and status code. To add the middleware to the application, app.add_middleware must be called. During testing, define a simple route, access it, and the log will be output to the console. Extensions and optimizations can include using the logging module, recording request bodies, and exception handling. Middleware simplifies development, improves code maintainability, and is suitable for common logic such as authentication and CORS.
Read MoreDetailed Explanation of FastAPI Request Bodies: Defining Complex Data Structures with Pydantic
This article introduces Pydantic, a core tool in FastAPI for handling complex request bodies. Request bodies are used to pass complex data (e.g., JSON) in POST/PUT requests and are more suitable for structured data compared to query parameters. As the recommended data validation and parsing library by FastAPI, Pydantic allows defining data structures and automatically validates types and formats, reducing the need for manual parsing code. For basic models, such as a `User` class with `name` and `age` fields, FastAPI automatically parses the request body into an object. Nested models are implemented through sub-models (e.g., a user with an address). List types are supported using `List` and nested lists (e.g., an order containing multiple products). Pydantic automatically intercepts invalid data and returns a 422 error when there is a type mismatch. In summary, mastering Pydantic standardizes API development. It enables handling complex structures through nested models and list support, and combining with automatic validation enhances robustness. It is a key skill for FastAPI to process request bodies.
Read MoreFastAPI and Python Version Compatibility: Version Issues to Note for Beginners
Python version compatibility is crucial in FastAPI development. Mismatched versions can lead to installation failures, errors, or missing functionality. FastAPI supports a minimum Python version of 3.6 and is compatible with the latest stable version (e.g., 3.11). Python 3.9 or 3.10 is recommended for the best stability and ecosystem support. To check the Python version, use terminal commands: on Windows, run `python --version`; on Mac/Linux, use `python3 --version`. A version ≥3.6 meets the basic requirements; upgrades are necessary for versions 3.5 and below. For upgrading: On Windows, download the installer from the official website and check "Add Python to PATH". On Mac/Linux (e.g., Ubuntu), use the system package manager (`sudo apt install python3.10`) or pyenv for multi-version management. Key considerations for different versions: Versions below 3.5 cannot install FastAPI. Python 3.6 lacks support for some advanced syntax (e.g., complex type hints). Python 3.11 requires ensuring compatibility with dependencies like Pydantic. Common error solutions: If installation fails due to an outdated Python version, upgrade Python. For syntax errors, verify usage of features unsupported in lower versions (e.g., Python 3.6 does not support the 3.8+ walrus operator). If dependency import fails, downgrade Pydantic (e.g., `pip install pydantic==1.10`).
Read MoreDetailed Explanation of FastAPI Dependency Injection: Basic and Advanced Usage of Depends
Dependency Injection (DI) core is to inject dependencies (such as database connections) into functions automatically by the system, rather than having the functions acquire them themselves, thereby enhancing code reusability and decoupling. FastAPI implements this through `Depends`, which involves two steps: defining a dependency function (to produce a dependency object, e.g., simulating a database connection), and declaring the dependency in the path function using `Depends(dependency_function)`, where FastAPI automatically calls and injects the result. Dependency functions can accept path/query parameters, such as querying a user based on a `user_id`. Advanced usages include: nested dependencies (dependencies on other dependencies), caching dependencies with `lru_cache` (singleton), asynchronous dependencies (adapting to asynchronous path functions), and combining with Pydantic for parameter validation. Core advantages: code reusability, decoupling (path functions only focus on business logic), testability (dependencies can be replaced with mocks), and extensibility (adding new dependencies only requires modifying the dependency function). Mastering `Depends` enables a clearer and more robust API structure.
Read MoreEnhancing FastAPI Documentation: Tips for Customizing Swagger UI
Swagger UI is the default API documentation tool for FastAPI, enabling interface visualization and testing. Customization can enhance its professionalism. For basic modifications, set parameters like title and description when creating the FastAPI instance to display personalized information on the `/docs` page. There are two styles customization methods: injecting CSS via middleware to quickly modify the background, navigation bar, etc.; or using static files to inject complex styles (e.g., logos). Sensitive information can be hidden by excluding fields with `Pydantic` model `Field(exclude=True)` or the interface `response_model_exclude`. Advanced techniques include layout adjustment, adding descriptions, and replacing buttons. The core implementation relies on basic information, CSS, and parameter control. Beginners can start with simple modifications while noting version compatibility.
Read MoreFastAPI Asynchronous Tasks: Handling Time-Consuming Operations with BackgroundTasks
In web development, directly handling time-consuming operations (such as sending emails and generating reports) in API endpoints will block user waiting and affect experience. FastAPI's `BackgroundTasks` can execute such tasks asynchronously after the request response, avoiding blocking. `BackgroundTasks` is a class provided by FastAPI that automatically executes background tasks after the request processing is completed without blocking the interface response. It only requires three steps: import `BackgroundTasks`, declare the `bg` parameter in the route function, and register the time-consuming function and parameters through `bg.add_task()`. Example: Simulating the generation of a large file (taking 5 seconds). After the user submits the request, the interface immediately returns success, and the file generation is completed asynchronously in the background. Key points: Tasks are executed after the response, support positional/keyword parameters and sequential execution, suitable for I/O-intensive tasks (such as file reading/writing), not suitable for CPU-intensive tasks; exceptions are not caught, and task failures need to be handled by oneself; unexecuted tasks will be lost when the application restarts or crashes, so it is not suitable for persistent tasks. `BackgroundTasks` is lightweight and easy to use, improving user experience through quick responses, and is suitable for non-critical path time-consuming operations.
Read MoreFastAPI Form Data Handling: Receiving multipart/form-data
To handle `multipart/form-data` format (for mixed form and file transmission) in FastAPI, tools like `Form`, `File`, or `UploadFile` are required. Text data is received using `Form`, where `Form(...)` marks required parameters (e.g., `name: str = Form(...)`), and optional parameters can be set with default values. For file uploads, there are two methods: `File` returns binary content (suitable for simple scenarios), while `UploadFile` provides metadata such as filename and MIME type (use the `read()` method if saving the file). In mixed scenarios, both `Form` and file tools must be used simultaneously. Testing can be done by submitting requests through FastAPI's built-in Swagger UI (`http://localhost:8000/docs`). Mastering these tools enables handling requirements for form submissions with both text and files.
Read MoreDetailed Explanation of FastAPI Status Codes: Usage Scenarios for 200, 404, 500, etc.
HTTP status codes are numeric codes returned by servers to indicate the result of request processing. Proper configuration in FastAPI helps clients understand the outcome of requests. There are two ways to set status codes in FastAPI: directly returning a tuple (data + status code), or using the `HTTPException` exception (recommended for error scenarios). Common core status codes and scenarios: 200 (Request successful, used for returning data in GET/PUT, etc.); 404 (Resource not found, when GET/DELETE requests cannot locate the target); 500 (Internal server error, requires exception handling to avoid exposure); 201 (POST request successfully created a resource, returns the new resource); 204 (No content, successful DELETE/PUT with no returned data); 400 (Bad request, e.g., format or required field issues); 401 (Unauthorized, user not logged in); 403 (Forbidden, authenticated but with insufficient permissions). Best practices: Map status codes to corresponding HTTP methods, e.g., 200/404 for GET, 201 for POST, 204 for DELETE. Correctly using status codes prevents client-side errors, and FastAPI's Swagger documentation aids in debugging.
Read MoreFastAPI + Docker: Complete Steps for Containerized Deployment
This article introduces the method of containerizing a FastAPI application using Docker to solve the environment inconsistency issue in development and deployment. First, create a FastAPI application: write `main.py` (including root path and parameterized interfaces), install dependencies `fastapi` and `uvicorn`, and generate `requirements.txt`. Next, package it through a Dockerfile: base on the Python 3.9-slim image, set the working directory to `/app`, copy dependency files and install them, copy the code, and finally start the service with `uvicorn` (port 8000). Execute `docker build -t my-fastapi-app .` to build the image, then run the container with `docker run -p 8000:8000 my-fastapi-app`. For testing, access `http://localhost:8000` or the API documentation at `http://localhost:8000/docs`. Common issues like port conflicts require changing the port or stopping the program; code modifications need rebuilding the image and restarting the container. The advantages of containerization include environment consistency, rapid migration, and dependency isolation. Future optimizations can include Docker Compose, reverse proxies, etc.
Read MoreFastAPI Request Timeouts? A Guide to Asynchronous Processing and Performance Optimization
The essence of request timeouts in FastAPI is that the server processing time exceeds the client's waiting threshold. Common causes include user network lag, high server load, and interface-specific time-consuming operations (e.g., I/O operations), which can lead to poor user experience and request failures. Methods to set timeouts in FastAPI: At the route level, use the `timeout` parameter in `async def` functions (e.g., `@app.get("/slow-task", timeout=10)`); at the global level, implement via middleware (e.g., intercepting requests and setting `asyncio.wait_for` with a 10-second timeout). Asynchronous processing is key to improving speed: Asynchronous routes (`async def`) use `await` to call non-blocking operations (e.g., async database queries), while background tasks (`BackgroundTasks` or `asyncio.create_task`) handle non-immediate-return tasks (e.g., sending emails). Comprehensive performance optimization requires: Caching frequently accessed data (`lru_cache` or Redis), configuring database connection pools (e.g., `asyncpg`), optimizing databases (avoid N+1 queries, add indexes, batch operations), and deploying with multi-processing (Uvicorn multi-worker) or load balancing. Comprehensive optimization steps: Set reasonable timeout thresholds, process I/O-intensive tasks asynchronously, cache frequently accessed data, optimize database connections, and deploy appropriately.
Read MoreCommon Pitfalls in FastAPI: The Most Frequent Mistakes for Novice Developers
This article summarizes 8 common errors and their solutions in FastAPI development: 1. Parameter type confusion: Path parameters need explicit type declaration (e.g., `user_id: int`), query parameters are suitable for simple filtering, and complex data should use POST with Pydantic request body; 2. Pydantic models must correctly define types and inherit `BaseModel`, with field types matching input parameters; 3. Status codes should follow REST specifications (201 for resource creation, 204 for deletion); 4. CORS configuration requires `CORSMiddleware`, specifying frontend domain in production; 5. Use `asyncio.run_in_executor` when calling synchronous libraries from async functions; 6. Use `yield` for dependency injection to handle resource release, and import middleware from FastAPI's corresponding modules; 7. Routes must be registered with the app to generate documentation. It is recommended to refer to the official documentation, verify parameter types and status codes, and avoid issues like un释放ed resources.
Read MoreFastAPI Practical Case: Build a Simple Blog API with 50 Lines of Code
FastAPI is a modern, high-performance Python framework that supports asynchronous programming, type hints, and automatic API documentation, making it ideal for quickly building APIs. This article implements a simple blog API with CRUD functionality for articles using just 50 lines of code. First, install `fastapi` and `uvicorn`. Define `PostCreate` (request model) and `PostResponse` (response model) using `Pydantic`, and simulate an in-memory list `posts` to store articles. Five endpoints are implemented: `GET /posts` (retrieve all articles), `GET /posts/{post_id}` (single article), `POST /posts` (create, with 201 status code), `PUT /posts/{post_id}` (update), and `DELETE /posts/{post_id}` (with 204 status code). FastAPI's automatic parameter validation and status code handling are leveraged throughout. FastAPI automatically generates Swagger UI and ReDoc documentation, facilitating easy testing of the API. Key knowledge points include route definition, Pydantic data models, status codes, and automatic documentation. Potential extensions could include adding a database, user authentication, and pagination. This example demonstrates FastAPI's concise and efficient nature, making it an excellent starting point for beginners.
Read MoreFastAPI + Redis: Basic Application of Caching and State Management
Web development often faces challenges of handling rapid request responses and sharing temporary state across multiple requests. The combination of FastAPI (a high-performance asynchronous framework) and Redis (an in-memory database) effectively addresses these issues. FastAPI supports high concurrency, while Redis, with its fast read/write speed and expiration time features, can cache frequently accessed, low-updating data (such as Fibonacci calculation results) to reduce redundant computations and store temporary states (such as user access counters) for cross-request sharing. This article introduces environment setup (installing FastAPI, Redis client, and starting Redis), basic connections (using utility functions to manage Redis clients with `decode_responses=True` to ensure string results), and demonstrates applications through two examples: caching (Fibonacci calculation results) and state management (user access counts). It also mentions dependency injection for code optimization and notes on Redis persistence, connection pooling, and key naming conventions in production environments.
Read MoreFastAPI + JWT Authentication: Implementing Simple User Login Verification
This article introduces the complete process of implementing user login authentication using FastAPI and JWT, with the core steps as follows: 1. **Environment Preparation**: Install FastAPI, uvicorn, python-jose (for JWT handling), passlib[bcrypt] (for password hashing), and python-multipart (for form processing). 2. **Core Concepts**: JWT enables stateless authentication, FastAPI dependencies reuse verification logic, and passwords are stored using bcrypt hashing. 3. **Code Implementation**: - Configure JWT parameters (secret key, algorithm, expiration time) and simulate a user database. - Use passlib to generate password hashes and define utility functions for JWT generation and verification. - Define OAuth2 dependencies to extract tokens, create a login endpoint (/token) to validate users and return tokens, and protected endpoints (/users/me) to return user information after token validation. 4. **Running Tests**: Start the uvicorn service, use Swagger UI to test the login endpoint for obtaining a token, and then use the token to access protected endpoints. 5. **Key Knowledge Points**: Dependency reuse for verification logic, secure handling of JWT secrets (environment variables for production), and password hashing to prevent plaintext leaks. Through the above steps, the implementation
Read MoreFastAPI + Uvicorn: Basic Configuration for Local Development and Deployment
This article introduces the web development and deployment process using FastAPI with Uvicorn. FastAPI is a high-performance Python framework supporting asynchronous operations and automatic API documentation. Uvicorn, as an ASGI server, is the recommended deployment tool for FastAPI, and their combination enables efficient development. **Environment Installation**: First, create a virtual environment (e.g., `python -m venv venv`), activate it, and then install dependencies with `pip install fastapi uvicorn`. **Development Configuration**: Write `main.py`, define routes (e.g., root route `/` and parameterized route `/items/{item_id}`), and start the server with `uvicorn main:app --reload` for auto-reloading in development mode. Verify the interface by accessing `http://127.0.0.1:8000`. **Production Deployment**: Use the basic command `uvicorn main:app --host 0.0.0.0 --port 8000`. Specify the number of worker processes with `--workers` for multi-processing. Ensure the deployment server opens the port and manage processes via `nohup` or `systemd`. **Common Issues**: For port conflicts, change the port. If the service is unreachable, confirm `--host 0.0.0.0` and firewall settings. If installation fails, update pip or check Python version compatibility.
Read MoreFastAPI + SQLAlchemy: Rapidly Building Database-Driven APIs
This article introduces how to build a database-driven API using FastAPI + SQLAlchemy, suitable for beginners. Its core advantages lie in FastAPI's high performance, automatic API documentation, and concise syntax, combined with SQLAlchemy's ORM to simplify database operations. Prerequisites include installing FastAPI, Uvicorn, SQLAlchemy, and Pydantic. The project adopts a modular structure: database.py configures database connections (using SQLite as an example), models.py defines ORM models for user tables, schemas.py uses Pydantic for data validation (distinguishing requests/responses), crud.py implements CRUD operations, and main.py integrates modules and defines API routes. Core modules include: database engine, session management (automatically creating/closing connections), user table models, data validation, and CRUD logic. The service is started via Uvicorn, and the Swagger UI enables interactive API testing. Advantages include automatic documentation, ORM simplification of SQL, modular design, and dependency injection for session management, making it suitable for quickly building efficient and maintainable web applications.
Read MoreFastAPI + CORS: A Quick Solution to Cross-Origin Resource Sharing Issues
Cross-origin issues occur when a frontend requests backend APIs from different domains, ports, or protocols, which are blocked by the browser's same-origin policy. By default, FastAPI does not handle cross-origin requests and requires the CORS middleware to resolve this. The core solution involves adding `CORSMiddleware` to FastAPI, with key parameters including: `allow_origins` (allowed frontend domains, use `["*"]` for development and specify specific domains in production), `allow_credentials` (whether to allow cookies to be carried across origins), `allow_methods` (allowed HTTP methods), and `allow_headers` (allowed request headers). It is crucial to avoid `allow_origins=["*"]` in production environments, as this should be restricted to specific domains. When allowing credentials, `allow_origins` must be explicitly specified. After configuration, the frontend can request backend APIs normally, such as the example where `fetch("http://localhost:8000/api/hello")` returns data. In summary, configuring the CORS middleware safely resolves cross-origin issues, offering flexibility in development while requiring strict parameter restrictions in production.
Read More