Crate triplets_core 

Core types, traits, and algorithms for the triplets data pipeline framework.

Re-exports

pub use chunking::ChunkingAlgorithm;

pub use chunking::SlidingWindowChunker;

pub use config::ChunkingStrategy;

pub use config::DenoiserConfig;

pub use config::NegativeStrategy;

pub use config::SamplerConfig;

pub use config::Selector;

pub use config::TextRecipe;

pub use config::TripletRecipe;

pub use data::DataRecord;

pub use data::PairLabel;

pub use data::QualityScore;

pub use data::RecordChunk;

pub use data::SampleBatch;

pub use data::SamplePair;

pub use data::SampleTriplet;

pub use data::SectionRole;

pub use data::TextBatch;

pub use data::TextSample;

pub use data::TripletBatch;

pub use hash::stable_hash_str;

pub use ingestion::IngestionManager;

pub use ingestion::RecordCache;

pub use kvp::KvpField;

pub use kvp::KvpPrefixSampler;

pub use preprocessor::TextPreprocessor;

pub use preprocessor::backends::denoiser_preprocessor::DenoiserPreprocessor;

pub use sampler::BatchPrefetcher;

pub use sampler::Sampler;

pub use sampler::TripletSampler;

pub use source::InMemorySource;

pub use source::backends::csv_source::CsvSource;

pub use source::backends::csv_source::CsvSourceConfig;

pub use source::DataSource;

pub use source::SourceCursor;

pub use splits::DeterministicSplitStore;

pub use splits::FileSplitStore;

pub use splits::SplitLabel;

pub use splits::SplitRatios;

pub use splits::SplitStore;

pub use types::CategoryId;

pub use types::HashPart;

pub use types::KvpValue;

pub use types::LogMessage;

pub use types::MetaValue;

pub use types::PathString;

pub use types::RecipeKey;

pub use types::RecordId;

pub use types::Sentence;

pub use types::SourceId;

pub use types::TaxonomyValue;

Modules

chunking

Pluggable chunking algorithms and default sliding-window implementation. Pluggable chunking algorithms.

config

Sampling configuration types.

constants

Centralized constants used across sampler, splits, and sources.

data

Data record and sample batch types.

hash

Stable deterministic hashing utilities.

heuristics

Capacity and sampling estimation helpers.

ingestion

Background ingestion and caching infrastructure.

kvp

Key/value prefix sampling helpers.

metadata

Metadata keys and helpers.

metrics

Aggregate metrics helpers.

preprocessor

OCR denoising and markdown-table cleanup for text chunks. Pluggable text preprocessor infrastructure.

sampler

Sampler implementations and public sampling API.

source

Data source traits and built-in sources. Data source interfaces and paging helpers.

splits

Split stores and persistence helpers.

tokenizer

Structural text tokenizer trait and whitespace implementation. Tokenization primitives used across chunking, sampling, and BM25 indexing.

types

Shared type aliases.

utils

Text normalization helpers. Text normalization helpers shared by source implementations.

Enums

SamplerError

Error type for sampler configuration, IO, and persistence failures.