wzc570738205/smartParsePro

What it solves

It solves the problem of parsing unstructured Chinese address strings into structured data. It handles the complexity of Chinese administrative divisions, including cases where multiple locations share the same name (ambiguity), and extracts additional entities like names, phone numbers, postal codes, and ID cards.

How it works

The project uses a rule-based offline engine that matches input text against administrative division data. Instead of guessing a single result, it returns a ranked array of candidates based on confidence scores. This allows the calling application to resolve ambiguities using external context (like user location or order history). It supports various deployment modes, including a client-side SDK for browsers, a Node.js module, and a standalone HTTP API server.

Who it’s for

• Developers building e-commerce or logistics applications that need to structure shipping addresses.
• Systems requiring offline address parsing to ensure data privacy.
• Applications that need to extract contact information (names, phones) alongside location data from a single string.

Highlights

• Offline Processing: Runs entirely locally without uploading data to external servers.
• Ambiguity Handling: Returns multiple candidates instead of a single potentially incorrect guess.
• Flexible Integration: Available as an NPM package, a Docker container, or a browser script.
• ** quantization-free extraction**: Extracts names, phone numbers, postal codes, and ID cards in addition to province, city, district, and street.
• Customizable: Supports loading custom administrative division data and aliases via createParser.