ferrous-opencc

ferrous-opencc is a pure-Rust reimplementation of the OpenCC (Open Chinese Convert) project, focused on fast and reliable conversion between Traditional and Simplified Chinese, along with regional variants such as Taiwan and Hong Kong forms. It exposes an OpenCC type you initialize with a built-in configuration name or a custom config file and then use to convert text with a single method call.

Unlike the original C++ OpenCC, this crate has no C++ dependencies and uses finite state transducers (FST) for dictionary lookups, which the project reports as significantly faster than HashMap-based approaches. It also ships a command-line tool for compiling text dictionaries into an efficient binary .ocb format for custom conversion setups.

What You Get

• An OpenCC type initialized from built-in configuration names with no external files needed
• Conversion between Simplified, Traditional, and regional variants (Taiwan, Hong Kong)
• FST-backed dictionary lookups for high-performance conversion
• A CLI tool to compile custom text dictionaries into the binary .ocb format

Common Use Cases

• Converting user-generated content between Simplified and Traditional Chinese
• Localizing Chinese text to regional variants for Taiwan or Hong Kong audiences
• Preprocessing Chinese text corpora with consistent script normalization

Under The Hood

Architecture - The public entry point is the OpenCC type (src/lib.rs) constructed via from_config with either a BuiltinConfig (src/config.rs) or a custom file. A config declares an ordered chain of conversion segments; each segment is a dictionary (src/dictionary.rs) loaded as an FST, and src/conversion.rs applies them as a longest-match pass over the input string. An FFI layer (src/ffi.rs) exposes a C-compatible surface, and a bin target compiles text dictionaries into the binary .ocb format.

Tech Stack - Pure Rust with an fst crate backing dictionary storage and lookup; no C++ or native OpenCC dependency. Built dictionaries ship in an .ocb binary format produced by the crate’s own CLI, and a small amount of C++ in the repo appears to support FFI/interop rather than core logic.

Code Quality - The crate has CI, docs.rs coverage, a dedicated error module (src/error.rs), and both English and Chinese READMEs. Development is active though driven mainly by a single maintainer, and the project is relatively young with a small community.

API Design - Onboarding is minimal: call OpenCC::from_config with a built-in name and then convert(text). Built-in configurations remove the need to manage dictionary files for common cases, while custom config loading and the .ocb compiler keep the door open for advanced setups without complicating the common path.