MobX RESTful migrator

Data Migration framework based on MobX-RESTful

Overview

MobX-RESTful-migrator is a TypeScript library that provides a flexible data migration framework built on top of MobX-RESTful's ListModel abstraction. It allows you to migrate data from various sources through MobX-RESTful models with customizable field mappings and relationships.

Features

• Flexible Field Mappings: Support for four different mapping types
• Async Generator Pattern: Control migration flow at your own pace
• Cross-table Relationships: Handle complex data relationships
• Event-Driven Architecture: Built-in console logging with customizable event bus
• TypeScript Support: Full TypeScript support with type safety

Installation

npm install mobx-restful mobx-restful-migrator

Usage Example: Article migration

The typical use case is migrating Article data with the following schema:

• Source: Article table with Title, Keywords, Content, Author, Email fields
• Target: Keywords field splits into Category string & Tags array, Author & Email fields map to User table

Source Data Schema

interface SourceArticle {
  id: number;
  title: string;
  subtitle: string;
  keywords: string; // comma-separated keywords to split into category & tags
  content: string;
  author: string; // maps to User table "name" field
  email: string; // maps to User table "email" field
}

Target Models

import { HTTPClient } from 'koajax';
import { ListModel, DataObject, Filter, IDType, toggle } from 'mobx-restful';
import { buildURLData } from 'web-utility';

export abstract class TableModel<
  D extends DataObject,
  F extends Filter<D> = Filter<D>,
> extends ListModel<D, F> {
  client = new HTTPClient({ baseURI: 'http://localhost:8080', responseType: 'json' });

  @toggle('uploading')
  async updateOne(data: Filter<D>, id?: IDType) {
    const { body } = await (id
      ? this.client.put<D>(`${this.baseURI}/${id}`, data)
      : this.client.post<D>(this.baseURI, data));

    return (this.currentOne = body!);
  }

  async loadPage(pageIndex: number, pageSize: number, filter: F) {
    const { body } = await this.client.get<{ list: D[]; count: number }>(
      `${this.baseURI}?${buildURLData({ ...filter, pageIndex, pageSize })}`,
    );
    return { pageData: body!.list, totalCount: body!.count };
  }
}

export interface User {
  id: number;
  name: string;
  email?: string;
}

export interface Article {
  id: number;
  title: string;
  category: string;
  tags: string[];
  content: string;
  author: User;
}

export class UserModel extends TableModel<User> {
  baseURI = '/users';
}

export class ArticleModel extends TableModel<Article> {
  baseURI = '/articles';
}

Migration Configuration

First, export your CSV data file articles.csv from an Excel file or Old database:

title,subtitle,keywords,content,author,email
Introduction to TypeScript,A Comprehensive Guide,"typescript,javascript,programming","TypeScript is a typed superset of JavaScript...",John Doe,[email protected]
MobX State Management,Made Simple,"mobx,react,state-management","MobX makes state management simple...",Jane Smith,[email protected]

Then implement the migration:

#! /usr/bin/env tsx

import { RestMigrator, MigrationSchema, ConsoleLogger } from 'mobx-restful-migrator';
import { FileHandle, open } from 'fs/promises';
import { readTextTable } from 'web-utility';

import { SourceArticle, Article, ArticleModel, UserModel } from './source';

// Load and parse CSV data using async streaming for large files
async function* readCSV<T extends object>(path: string) {
  let fileHandle: FileHandle | undefined;

  try {
    fileHandle = await open(path);

    const stream = fileHandle.createReadStream({ encoding: 'utf-8' });

    yield* readTextTable<T>(stream, true) as AsyncGenerator<T>;
  } finally {
    await fileHandle?.close();
  }
}

const loadSourceArticles = () => readCSV<SourceArticle>('article.csv');

// Complete migration configuration demonstrating all 4 mapping types
const mapping: MigrationSchema<SourceArticle, Article> = {
  // 1. Many-to-One mapping: Title + Subtitle → combined title
  title: ({ title, subtitle }) => ({
    title: { value: `${title}: ${subtitle}` },
  }),
  content: 'content',

  // 2. One-to-Many mapping: Keywords string → category string & tags array
  keywords: ({ keywords }) => {
    const [category, ...tags] = keywords.split(',').map(tag => tag.trim());

    return { category: { value: category }, tags: { value: tags } };
  },
  // 3. Cross-table relationship: Author & Email → User table
  author: ({ author, email }) => ({
    author: {
      value: { name: author, email },
      model: UserModel, // Maps to User table via ListModel
    },
  }),
};

// Run migration with built-in console logging (default)
const migrator = new RestMigrator(loadSourceArticles, ArticleModel, mapping);

// The ConsoleLogger automatically logs each step:
// - saved No.X: successful migrations with source, mapped, and target data
// - skipped No.X: skipped items (duplicate unique fields)
// - error at No.X: migration errors with details

for await (const { title } of migrator.boot()) {
  // Process the migrated target objects
  console.log(`Successfully migrated article: ${title}`);
}

In the end, run your script with a TypeScript runtime:

tsx your-migration.ts 1> saved.log 2> error.log

Optional: Use custom event bus

class CustomEventBus implements MigrationEventBus<SourceArticle, Article> {
  async save({ index, targetItem }) {
    console.info(`✅ Migrated article ${index}: ${targetItem?.title}`);
  }

  async skip({ index, error }) {
    console.warn(`⚠️  Skipped article ${index}: ${error?.message}`);
  }

  async error({ index, error }) {
    console.error(`❌ Error at article ${index}: ${error?.message}`);
  }
}

const migratorWithCustomLogger = new RestMigrator(
  loadSourceArticles,
  ArticleModel,
  mapping,
  new CustomEventBus(),
);

Four Mapping Types

1. Simple 1-to-1 Mapping

Map source field directly to target field using string mapping:

const mapping: MigrationSchema<SourceArticle, Article> = {
  title: 'title',
  content: 'content',
};

2. Many-to-One Mapping

Use resolver function to combine multiple source fields into one target field:

const mapping: MigrationSchema<SourceArticle, Article> = {
  title: ({ title, subtitle }) => ({
    title: { value: `${title}: ${subtitle}` },
  }),
};

3. One-to-Many Mapping

Use resolver function to map one source field to multiple target fields with value property:

const mapping: MigrationSchema<SourceArticle, Article> = {
  keywords: ({ keywords }) => {
    const [category, ...tags] = keywords.split(',').map(tag => tag.trim());

    return { category: { value: category }, tags: { value: tags } };
  },
};

4. Cross-Table Relationships

Use resolver function with model property for related tables:

const mapping: MigrationSchema<SourceArticle, Article> = {
  author: ({ author, email }) => ({
    author: {
      value: { name: author, email },
      model: UserModel, // References User ListModel
    },
  }),
};

Event-Driven Migration Architecture

The migrator includes a built-in Event Bus for monitoring and controlling the migration process:

Built-in Console Logging

By default, RestMigrator uses the ConsoleLogger which provides detailed console output:

import { RestMigrator, ConsoleLogger } from 'mobx-restful-migrator';

import { loadSourceArticles, ArticleModel, mapping } from './source';

// ConsoleLogger is used by default
const migrator = new RestMigrator(loadSourceArticles, ArticleModel, mapping);

for await (const { title } of migrator.boot()) {
  // Console automatically shows:
  // - saved No.X with source, mapped, and target data tables
  // - skipped No.X for duplicate unique fields
  // - error at No.X for migration errors

  // Your processing logic here
  console.log(`✅ Article migrated: ${title}`);
}

Custom Event Handling

Implement your own Event Bus for custom logging and monitoring:

import { MigrationEventBus, MigrationProgress } from 'mobx-restful-migrator';
import { outputJSON } from 'fs-extra';

import { SourceArticle, Article, loadSourceArticles, ArticleModel, mapping } from './source';

class FileLogger implements MigrationEventBus<SourceArticle, Article> {
  bootedAt = new Date().toJSON();

  async save({ index, sourceItem, targetItem }: MigrationProgress<SourceArticle, Article>) {
    // Log to file, send notifications, etc.
    await outputJSON(`logs/save-${this.bootedAt}.json`, {
      type: 'success',
      index,
      sourceId: sourceItem?.id,
      targetId: targetItem?.id,
      savedAt: new Date().toJSON(),
    });
  }

  async skip({ index, sourceItem, error }: MigrationProgress<SourceArticle, Article>) {
    await outputJSON(`logs/skip-${this.bootedAt}.json`, {
      type: 'skipped',
      index,
      sourceId: sourceItem?.id,
      error: error?.message,
      skippedAt: new Date().toJSON(),
    });
  }

  async error({ index, sourceItem, error }: MigrationProgress<SourceArticle, Article>) {
    await outputJSON(`logs/error-${this.bootedAt}.json`, {
      type: 'error',
      index,
      sourceId: sourceItem?.id,
      error: error?.message,
      errorAt: new Date().toJSON(),
    });
  }
}

const migrator = new RestMigrator(loadSourceArticles, ArticleModel, mapping, new FileLogger());