ArticleProcessor

ArticleProcessor is the core processing engine of ArtiPub, providing article preprocessing capabilities including image compression, image upload, and custom middleware processing. It processes Markdown files through a pipeline architecture, allowing developers to extend functionality through middleware.

Overview

The ArticleProcessor class handles the transformation of Markdown content before publishing. It automatically:

• Compresses images to reduce file size
• Uploads images to cloud storage (GitHub or custom)
• Injects unique article IDs for tracking
• Processes content through custom middleware

Constructor

class ArticleProcessor {
  constructor(option: ArticleProcessorOption);
}

ArticleProcessorOption

interface ArticleProcessorOption {
  /**
   * Image compression settings
   * @default { quality: 80, compressed: true }
   */
  compressedOptions?: {
    /**
     * Whether to compress images
     * @default true
     */
    compressed?: boolean;
    /**
     * Compression quality (1-100)
     * @default 80
     */
    quality?: number;
  };

  /**
   * Image upload configuration
   * Can be GitHub configuration or custom upload function
   */
  uploadImgOption: UploadImgOption;
}

UploadImgOption Types

/**
 * GitHub Picture Bed Configuration
 */
interface GithubPicBedOption {
  owner: string; // GitHub username or organization
  repo: string; // Repository name
  dir: string; // Directory path in repository
  branch: string; // Target branch (usually 'main' or 'master')
  token: string; // GitHub Personal Access Token
  commit_author: string; // Commit author name
  commit_email: string; // Commit author email
}

/**
 * Custom Upload Function
 * @param imgFilePath - Local image file path
 * @returns Promise resolving to the uploaded image URL
 */
type UploadImg = (imgFilePath: string) => Promise<string>;

/**
 * Upload option can be either GitHub config or custom function
 */
type UploadImgOption = GithubPicBedOption | UploadImg;

Methods

processMarkdown

Processes a Markdown file through the configured pipeline.

processMarkdown(filePath: string): Promise<ArticleProcessResult>

Parameters:

• filePath (string): Absolute path to the Markdown file

Returns:

• Promise<ArticleProcessResult>: Processed content

interface ArticleProcessResult {
  content: string; // Processed Markdown content
}

use

Adds custom middleware to the processing pipeline.

use(middleware: Middleware): ArticleProcessor

Parameters:

• middleware (Middleware): Custom middleware function

Returns:

• ArticleProcessor: Returns itself for method chaining

Middleware

Middleware allows you to customize the processing pipeline by manipulating the Markdown AST (Abstract Syntax Tree).

Middleware Type Definition

type Middleware = (context: ProcessorContext, visitor: TVisitor, next: Next) => Promise<void>;

interface ProcessorContext {
  option: ArticleProcessorOption;
  filePath: string;
}

type Next = () => void;
type TVisitor = (
  testOrVisitor: Visitor | Test,
  visitorOrReverse: Visitor | boolean | null | undefined,
  maybeReverse?: boolean | null | undefined
) => void;

Writing Custom Middleware

const customMiddleware: Middleware = async (context, visit, next) => {
  // Visit specific node types in the AST
  visit("heading", (node, index, parent) => {
    // Modify heading nodes
    if (node.depth === 1) {
      // Process h1 headings
      console.log("Found title:", node.children[0].value);
    }
  });

  // Visit image nodes
  visit("image", (node) => {
    // Process image URLs
    console.log("Found image:", node.url);
  });

  // IMPORTANT: Always call next() to continue the pipeline
  next();
};

Complete Examples

Basic Usage with GitHub Image Hosting

import { ArticleProcessor } from "@artipub/core";
import path from "path";

// Configure with environment variables
const processor = new ArticleProcessor({
  compressedOptions: {
    quality: 85,
    compressed: true,
  },
  uploadImgOption: {
    owner: process.env.GITHUB_OWNER!,
    repo: process.env.GITHUB_REPO!,
    dir: "blog/images",
    branch: "main",
    token: process.env.GITHUB_TOKEN!,
    commit_author: "ArtiPub Bot",
    commit_email: "bot@example.com",
  },
});

// Process a Markdown file
const result = await processor.processMarkdown(path.resolve(__dirname, "./articles/my-post.md"));

console.log("Processed content:", result.content);

Custom Image Upload Function

import { ArticleProcessor } from "@artipub/core";
import { uploadToS3 } from "./utils/s3-upload";

const processor = new ArticleProcessor({
  uploadImgOption: async (imgFilePath: string) => {
    // Custom upload logic (e.g., to AWS S3, Cloudinary, etc.)
    const url = await uploadToS3(imgFilePath);
    return url;
  },
});

Advanced: Using Multiple Middleware

import { ArticleProcessor } from "@artipub/core";
import { Heading, Text, Image } from "mdast";

const processor = new ArticleProcessor({
  uploadImgOption: {
    /* ... */
  },
});

// Middleware 1: Add reading time estimate
processor.use(async (context, visit, next) => {
  let wordCount = 0;

  visit("text", (node: Text) => {
    wordCount += node.value.split(/\s+/).length;
  });

  const readingTime = Math.ceil(wordCount / 200); // 200 words per minute

  // Add reading time to the beginning of the document
  visit("root", (node) => {
    const readingTimeNode = {
      type: "paragraph",
      children: [
        {
          type: "text",
          value: `⏱️ Reading time: ${readingTime} min`,
        },
      ],
    };
    node.children.unshift(readingTimeNode);
  });

  next();
});

// Middleware 2: Add watermark to images
processor.use(async (context, visit, next) => {
  visit("image", (node: Image) => {
    // Add watermark parameter to image URLs
    if (!node.url.includes("?")) {
      node.url += "?watermark=artipub";
    }
  });

  next();
});

// Middleware 3: Convert relative links to absolute
processor.use(async (context, visit, next) => {
  const baseUrl = "https://myblog.com";

  visit("link", (node) => {
    if (node.url.startsWith("./") || node.url.startsWith("../")) {
      node.url = new URL(node.url, baseUrl).href;
    }
  });

  next();
});

Node Types Reference

Common node types you can visit in middleware:

Node Type	Description	Properties
root	Document root	children[]
heading	Headers (h1-h6)	depth, children[]
paragraph	Paragraphs	children[]
text	Text content	value
image	Images	url, alt, title
link	Links	url, children[]
code	Code blocks	lang, value
inlineCode	Inline code	value
blockquote	Quotes	children[]
list	Lists	ordered, children[]
listItem	List items	children[]
table	Tables	children[]
emphasis	Italic text	children[]
strong	Bold text	children[]

Best Practices

1. Always call next(): Forgetting to call next() in middleware will stop the pipeline.

2. Order matters: Middleware executes in the order they are added. Image processing middleware should come before publishing.

3. Error handling: Wrap async operations in try-catch blocks:

   processor.use(async (context, visit, next) => {
     try {
       // Your async operations
       await someAsyncOperation();
     } catch (error) {
       console.error("Middleware error:", error);
       // Decide whether to continue or throw
     }
     next();
   });

4. Performance: Avoid heavy operations in middleware. Consider caching results when possible.

5. Testing: Test middleware independently before adding to the processor:

   // Test middleware separately
   const testMiddleware = async () => {
     const mockContext = {
       /* ... */
     };
     const mockVisit = jest.fn();
     const mockNext = jest.fn();
   
     await myMiddleware(mockContext, mockVisit, mockNext);
   
     expect(mockNext).toHaveBeenCalled();
   };

API Reference

For more details on the visitor pattern and AST manipulation:

• unist-util-visit - Visitor utility documentation
• mdast - Markdown AST specification
• unified - Content processing ecosystem