A compact binary representation for InChI Keys.
This crate provides a space-efficient binary encoding for International Chemical Identifier (InChI) keys, reducing their size from the standard 27-byte ASCII representation to either 9 or 14 bytes. The implementation is based on the work by John Mayfield (NextMove Software): Data Compression of InChI Keys and 2D Coordinates.
Note: This is a personal project created for fun and to explore Rust. While it implements a real compression algorithm, it's primarily a learning exercise rather than a production-critical library.
Add this to your Cargo.toml:
[dependencies]
zinchi = "0.1"use zinchi::InChIKey;
// Parse an InChI key from a string
let key: InChIKey = "ZZJLMZYUGLJBSO-UHFFFAOYSA-N".parse().expect("Failed to parse InChIKey")
// Convert back to string
println!("{}", key);
// Access individual components
println!("Standard: {}", key.is_standard());
println!("Version: {}", key.version());
println!("Protonation: {}", key.get_protonation());use zinchi::InChIKey;
let key: InChIKey = "ZZJLMZYUGLJBSO-UHFFFAOYSA-N".parse()?;
// Pack to binary (9 or 14 bytes)
let packed = key.packed_bytes();
println!("Packed size: {} bytes", packed.len());
// Unpack from binary
let unpacked = InChIKey::unpack_from(&packed)?;
assert_eq!(key, unpacked);use zinchi::InChIKey;
let key: InChIKey = "ZZJLMZYUGLJBSO-UHFFFAOYSA-N".parse()?;
// Pack into an existing buffer
let mut buffer = [0u8; 14];
let size = key.pack_into(&mut buffer);
// Use only the relevant bytes
let packed_data = &buffer[..size];An InChI key has the format: AAAAAAAAAAAAAA-BBBBBBBBFV-P
- First block (14 chars): Encodes core molecular constitution (65 bits → 9 bytes)
- Second block (8 chars): Encodes stereochemistry and isotopes (37 bits → 5 bytes)
- Flag (1 char):
Sfor standard,Nfor non-standard - Version (1 char): Currently always
A - Protonation (1 char):
Nfor neutral, orA-Mfor protonated states
Standard InChI keys with the common second block UHFFFAOYSA (empty stereochemistry hash) are packed into just 9 bytes. All other InChI keys require 14 bytes.
This represents a 48-66% reduction in size compared to the ASCII representation.
The first block (14 characters) is decoded into four 14-bit triples and one 9-bit pair, then packed into 9 bytes. The second block (8 characters) is decoded into two 14-bit triples and one 9-bit pair, then packed into 5 bytes. Additional metadata (standard flag, version, protonation) is encoded into spare bits.
When the serde feature is enabled, InChIKey implements Serialize and Deserialize:
use zinchi::InChIKey;
let key: InChIKey = "ZZJLMZYUGLJBSO-UHFFFAOYSA-N".parse()?;
// JSON serialization (human-readable)
let json = serde_json::to_string(&key)?;
assert_eq!(json, "\"ZZJLMZYUGLJBSO-UHFFFAOYSA-N\"");
// Binary serialization with bincode (compact)
let bytes = bincode::serde::encode_to_vec(&key, bincode::config::standard())?;
// Uses the 9 or 14 byte packed representationThe serialization format automatically adapts:
- Human-readable formats (JSON, YAML, etc.): Serializes as an InChIKey string
- Binary formats (bincode, etc.): Serializes using the compact 9 or 14 byte representation
Contributions are welcome! Please feel free to submit a Pull Request.
This project is licensed under the MIT License - see the LICENSE file for details.
- John Mayfield and NextMove Software for the original compression algorithm
- The InChI Trust for the InChI specification