Skip to main content

Crate vole_document

Crate vole_document 

Source
Expand description

§VOLE-Document

Byte-exact procedural document storage.

The governing invariant of the exact profile is:

materialize(descriptor) == original_bytes

Parsing, semantic equivalence, canonicalization, or a successful round trip through a document writer are not substitutes for this invariant.

§Layered representation

A persisted .voldoc descriptor is a bounded, deterministic reconstruction description:

source bytes X
    -> deterministic format analysis (advisory; never destructive)
    -> bounded reconstruction hypothesis
    -> Document Reconstruction Algebra program + parameters + typed residuals
    -> (later) typed entropy channels / rANS entropy-seed capsules
    -> canonical .voldoc descriptor
    -> bounded deterministic materializer
    -> X exactly

rANS, when it eventually arrives, is the entropy substrate beneath the representation. It is not the procedural model, and an integer state is not a magic seed: a complete decoder-entry capsule carries model, state, renormalization payload, counts, and integrity.

§Phase status

This crate currently implements the exact container core (paper Phase A / brief Phase 1): framing, typed errors, resource limits, integrity, the literal Document Reconstruction Algebra, the coverage certificate, RAW exact representation, the CLI, and decode-before-commit. No compression claim is made yet.

See PROJECT_STATE.md for the mechanism ledger and SPEC.md for the provisional wire format.

Re-exports§

pub use error::Error;
pub use error::ErrorClass;
pub use error::Result;
pub use limits::Limits;

Modules§

accounting
Complete-cost accounting.
adapter
Format adapters.
container
The .voldoc container: header, framing, and descriptor model.
dra
The Document Reconstruction Algebra.
encode
Encoding: candidate portfolio, complete-cost court, and decode-before-commit.
entropy
Entropy substrate: canonical models, channel descriptors, and (later) rANS.
error
Typed errors with stable classes and documented CLI exit codes.
integrity
Integrity primitives.
limits
Centralized resource bounds.
materialize
Deterministic materialization.

Constants§

EXACTNESS_PROFILE_EXACT_BYTES
The exactness profile mandated for the archival path.
SOURCE_FORMAT_OPAQUE
Identifier for the opaque source-format class.
SOURCE_FORMAT_PDF
Identifier for the PDF source-format class.