Modules
Production-grade streaming data sketching algorithms for Elixir.
Behaviour defining the computation backend for ExDataSketch.
Pure Elixir backend for ExDataSketch.
Rust NIF-accelerated backend for ExDataSketch.
Public facade for the EXSK binary frame.
CRC32C (Castagnoli polynomial, reflected, init 0xFFFFFFFF, xor-out 0xFFFFFFFF) — the checksum used by EXSK v2 frames.
EXSK v2 binary frame header.
Structured validation primitives for EXSK frames.
Bloom filter for probabilistic membership testing.
Count-Min Sketch (CMS) for frequency estimation.
Counting Quotient Filter (CQF) for multiset membership with approximate counting.
ExDataSketch-native binary serialization codec (EXSK format).
Cuckoo filter for probabilistic membership testing with deletion support.
DDSketch quantiles sketch for value-relative-accuracy quantile estimation.
Apache DataSketches CompactSketch binary codec for Theta sketches.
Minimal MurmurHash3_x64_128 implementation for DataSketches seed hash computation.
Error types for ExDataSketch.
Returned when binary data cannot be deserialized.
Returned when attempting to merge sketches with incompatible parameters.
Raised when a FilterChain stage composition is invalid.
Returned when an option value is invalid.
Raised when a stubbed function is called before its implementation is available.
Raised when an operation is not supported by a structure.
Capability-aware composition framework for chaining membership filters.
FrequentItems sketch for approximate heavy-hitter detection using the SpaceSaving algorithm.
HyperLogLog (HLL) sketch for cardinality estimation.
Stable 64-bit hash interface for ExDataSketch.
Shared hash + sketch metadata block.
MurmurHash3_x64_128 (64-bit output).
Compatibility checks for hash configurations across merging sketches.
XXHash3 (64-bit) hash algorithm.
Invertible Bloom Lookup Table (IBLT) for set reconciliation.
KLL (Karnin-Lang-Liberty) quantiles sketch for rank and quantile estimation.
MisraGries sketch for deterministic heavy hitter detection.
Facade for quantile sketch algorithms.
Quotient filter for probabilistic membership testing with safe deletion and merge.
REQ (Relative Error Quantiles) sketch for tail-accuracy quantile estimation.
Theta Sketch for set operations on cardinalities.
UltraLogLog (ULL) sketch for cardinality estimation.
Xor filter for static, immutable probabilistic membership testing.