Crate rustledger_ops 

Pure operations on beancount directives.

This crate provides reusable functions for transforming and analyzing collections of beancount directives. All operations are pure — they take directives in and return results out, with no I/O or framework coupling.

Analogous to Python beancount’s ops/ module.

Modules

• fingerprint — structural hashing and stable fingerprinting of transactions
• dedup — duplicate detection (structural, fuzzy, and fingerprint-based)
• categorize — rules engine for transaction categorization (substring, regex, exact match)
• merchants — built-in merchant dictionary of common patterns
• enrichment — shared types for operation results (confidence, method, alternatives)
• reconcile — balance reconciliation against statement ending balances
• ml — ML-based categorization (TF-IDF + an in-house Multinomial Naive Bayes classifier, pure std, available on every target)
• transfer — inter-account transfer detection and linking

Modules

categorize

Rules-based transaction categorization.

clamp

Clamp directives to a date range, summarizing pre-range balances.

dedup

Duplicate transaction detection.

enrichment

Shared types for operation results.

fingerprint

Transaction fingerprinting and structural hashing.

merchants

Built-in merchant dictionary for transaction categorization.

ml

ML-based transaction categorization.

reconcile

Balance reconciliation.

transfer

Transfer matching across accounts.