From 04150b3b826c54d6469abaa6bc07cde324058e30 Mon Sep 17 00:00:00 2001 From: Colin Patrick McCabe Date: Wed, 16 Sep 2026 15:03:06 -0700 Subject: [PATCH] feat: index builders for NGram, BloomFilter, RTree (#4205) Add index builders in the lancedb library for NGram, BloomFilter, and RTree indexes. --- Cargo.lock | 297 +++++++++++++++++++++++++ rust/lancedb/Cargo.toml | 2 + rust/lancedb/src/index.rs | 57 ++++- rust/lancedb/src/index/scalar.rs | 170 ++++++++++++++ rust/lancedb/src/remote/table.rs | 30 +++ rust/lancedb/src/table/create_index.rs | 209 +++++++++++++++++ 6 files changed, 763 insertions(+), 2 deletions(-) diff --git a/Cargo.lock b/Cargo.lock index 598f8e8a2..7037168c6 100644 --- a/Cargo.lock +++ b/Cargo.lock @@ -3151,6 +3151,16 @@ version = "0.1.3" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "e1d926b4d407d372f141f93bb444696142c29d32962ccbd3531117cf3aa0bfa9" +[[package]] +name = "earcutr" +version = "0.4.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "79127ed59a85d7687c409e9978547cffb7dc79675355ed22da6b66fd5f6ead01" +dependencies = [ + "itertools 0.11.0", + "num-traits", +] + [[package]] name = "ecdsa" version = "0.14.8" @@ -3450,6 +3460,12 @@ dependencies = [ "rand_distr 0.5.1", ] +[[package]] +name = "float_next_after" +version = "1.0.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "8bf7cc16383c4b8d58b9905a8509f02926ce3058053c056376248d958c9df1e8" + [[package]] name = "fnv" version = "1.0.7" @@ -3776,6 +3792,129 @@ dependencies = [ "version_check", ] +[[package]] +name = "geo" +version = "0.31.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "2fc1a1678e54befc9b4bcab6cd43b8e7f834ae8ea121118b0fd8c42747675b4a" +dependencies = [ + "earcutr", + "float_next_after", + "geo-types", + "geographiclib-rs", + "i_overlay", + "log", + "num-traits", + "robust", + "rstar", + "spade", +] + +[[package]] +name = "geo-traits" +version = "0.3.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "2e7c353d12a704ccfab1ba8bfb1a7fe6cb18b665bf89d37f4f7890edcd260206" +dependencies = [ + "geo-types", +] + +[[package]] +name = "geo-types" +version = "0.7.19" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "94776032c45f950d30a13af6113c2ad5625316c9abfbccee4dd5a6695f8fe0f5" +dependencies = [ + "approx", + "num-traits", + "rayon", + "rstar", + "serde", +] + +[[package]] +name = "geoarrow-array" +version = "0.8.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "dafe7b7de3fab1a8b7099fd6a6434ca955fa65065f9c19f0f8a133693f3c2b0e" +dependencies = [ + "arrow-array", + "arrow-buffer", + "arrow-schema", + "geo-traits", + "geoarrow-schema", + "num-traits", + "wkb", + "wkt", +] + +[[package]] +name = "geoarrow-expr-geo" +version = "0.8.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "8e4a62ac19c86827c6ec81ea584594b3ee96db5a8119b9774d3466c6b373c434" +dependencies = [ + "arrow-array", + "arrow-buffer", + "geo", + "geo-traits", + "geoarrow-array", + "geoarrow-schema", +] + +[[package]] +name = "geoarrow-schema" +version = "0.8.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "4d4a7edb2a1d87024a93805332a9c8184a0354836271d42c0d18cf628a5e3cd0" +dependencies = [ + "arrow-schema", + "geo-traits", + "serde", + "serde_json", + "thiserror 1.0.69", +] + +[[package]] +name = "geodatafusion" +version = "0.5.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "fecbdd00d0fff2b04635c1b1e4129c217908f0c2d17539e0a2275308afce2552" +dependencies = [ + "arrow-arith", + "arrow-array", + "arrow-buffer", + "arrow-schema", + "datafusion", + "geo", + "geo-traits", + "geoarrow-array", + "geoarrow-expr-geo", + "geoarrow-schema", + "geohash", + "thiserror 1.0.69", + "wkt", +] + +[[package]] +name = "geographiclib-rs" +version = "0.2.7" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "c5a7f08910fd98737a6eda7568e7c5e645093e073328eeef49758cfe8b0489c7" +dependencies = [ + "libm", +] + +[[package]] +name = "geohash" +version = "0.13.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "7f58890382f70caccc5fa388981f7ac80c913795042afce9f3e065695d8f7464" +dependencies = [ + "geo-types", + "libm", +] + [[package]] name = "getrandom" version = "0.2.17" @@ -3954,6 +4093,15 @@ dependencies = [ "zerocopy", ] +[[package]] +name = "hash32" +version = "0.3.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "47d60b12902ba28e2730cd37e95b8c9223af2808df9e902d4df49588d1470606" +dependencies = [ + "byteorder", +] + [[package]] name = "hashbrown" version = "0.12.3" @@ -4010,6 +4158,16 @@ version = "0.2.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "0049b265b7f201ca9ab25475b22b47fe444060126a51abe00f77d986fc5cc52e" +[[package]] +name = "heapless" +version = "0.8.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0bfb9eb618601c89945a70e254898da93b13be0388091d42117462b265bb3fad" +dependencies = [ + "hash32", + "stable_deref_trait", +] + [[package]] name = "heck" version = "0.4.1" @@ -4327,6 +4485,49 @@ dependencies = [ "serde", ] +[[package]] +name = "i_float" +version = "1.15.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "010025c2c532c8d82e42d0b8bb5184afa449fa6f06c709ea9adcb16c49ae405b" +dependencies = [ + "libm", +] + +[[package]] +name = "i_key_sort" +version = "0.6.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9190f86706ca38ac8add223b2aed8b1330002b5cdbbce28fb58b10914d38fc27" + +[[package]] +name = "i_overlay" +version = "4.0.7" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "413183068e6e0289e18d7d0a1f661b81546e6918d5453a44570b9ab30cbed1b3" +dependencies = [ + "i_float", + "i_key_sort", + "i_shape", + "i_tree", + "rayon", +] + +[[package]] +name = "i_shape" +version = "1.14.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "1ea154b742f7d43dae2897fcd5ead86bc7b5eefcedd305a7ebf9f69d44d61082" +dependencies = [ + "i_float", +] + +[[package]] +name = "i_tree" +version = "0.16.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "35e6d558e6d4c7b82bc51d9c771e7a927862a161a7d87bf2b0541450e0e20915" + [[package]] name = "iana-time-zone" version = "0.1.65" @@ -5061,6 +5262,7 @@ dependencies = [ "jsonb", "lance-arrow", "lance-core", + "lance-geo", "log", "pin-project", "prost", @@ -5164,6 +5366,21 @@ dependencies = [ "tracing", ] +[[package]] +name = "lance-geo" +version = "13.0.0-beta.3" +source = "git+https://github.com/lance-format/lance.git?tag=v13.0.0-beta.3#565808c992c3d2adf1909b9753b50a28b6bd8a19" +dependencies = [ + "datafusion", + "geo-traits", + "geo-types", + "geoarrow-array", + "geoarrow-schema", + "geodatafusion", + "lance-core", + "serde", +] + [[package]] name = "lance-index" version = "13.0.0-beta.3" @@ -5190,6 +5407,9 @@ dependencies = [ "dirs", "fst", "futures", + "geo-types", + "geoarrow-array", + "geoarrow-schema", "half", "itertools 0.14.0", "jieba-rs", @@ -5201,6 +5421,7 @@ dependencies = [ "lance-datafusion", "lance-encoding", "lance-file", + "lance-geo", "lance-index-core", "lance-io", "lance-linalg", @@ -6414,6 +6635,28 @@ dependencies = [ "libc", ] +[[package]] +name = "num_enum" +version = "0.7.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "5d0bca838442ec211fa11de3a8b0e0e8f3a4522575b5c4c06ed722e005036f26" +dependencies = [ + "num_enum_derive", + "rustversion", +] + +[[package]] +name = "num_enum_derive" +version = "0.7.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "680998035259dcfcafe653688bf2aa6d3e2dc05e98be6ab46afb089dc84f1df8" +dependencies = [ + "proc-macro-crate", + "proc-macro2", + "quote", + "syn 2.0.117", +] + [[package]] name = "number_prefix" version = "0.4.0" @@ -8771,6 +9014,12 @@ dependencies = [ "byteorder", ] +[[package]] +name = "robust" +version = "1.2.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "4e27ee8bb91ca0adcf0ecb116293afa12d393f9c2b9b9cd54d33e8078fe19839" + [[package]] name = "rsa" version = "0.9.10" @@ -8792,6 +9041,17 @@ dependencies = [ "zeroize", ] +[[package]] +name = "rstar" +version = "0.12.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "421400d13ccfd26dfa5858199c30a5d76f9c54e0dba7575273025b43c5175dbb" +dependencies = [ + "heapless", + "num-traits", + "smallvec", +] + [[package]] name = "rstest" version = "0.23.0" @@ -9530,6 +9790,18 @@ dependencies = [ "winapi", ] +[[package]] +name = "spade" +version = "2.15.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9699399fd9349b00b184f5635b074f9ec93afffef30c853f8c875b32c0f8c7fa" +dependencies = [ + "hashbrown 0.16.1", + "num-traits", + "robust", + "smallvec", +] + [[package]] name = "spin" version = "0.9.9" @@ -11353,6 +11625,31 @@ dependencies = [ "wasmparser", ] +[[package]] +name = "wkb" +version = "0.9.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "a120b336c7ad17749026d50427c23d838ecb50cd64aaea6254b5030152f890a9" +dependencies = [ + "byteorder", + "geo-traits", + "num_enum", + "thiserror 1.0.69", +] + +[[package]] +name = "wkt" +version = "0.14.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "efb2b923ccc882312e559ffaa832a055ba9d1ac0cc8e86b3e25453247e4b81d7" +dependencies = [ + "geo-traits", + "geo-types", + "log", + "num-traits", + "thiserror 1.0.69", +] + [[package]] name = "writeable" version = "0.6.3" diff --git a/rust/lancedb/Cargo.toml b/rust/lancedb/Cargo.toml index 899600bf9..bde33283d 100644 --- a/rust/lancedb/Cargo.toml +++ b/rust/lancedb/Cargo.toml @@ -128,6 +128,8 @@ pprof = { version = "0.14", features = ["flamegraph"] } [features] default = [] +# Native spatial indexes and queries. Remote RTree requests do not require this feature. +geo = ["lance/geo"] aws = [ "dep:aws-smithy-types", "lance/aws", diff --git a/rust/lancedb/src/index.rs b/rust/lancedb/src/index.rs index 0611f95b1..cdc973044 100644 --- a/rust/lancedb/src/index.rs +++ b/rust/lancedb/src/index.rs @@ -14,8 +14,8 @@ use crate::{DistanceType, Error, Result, job::Job, table::BaseTable}; use self::{ scalar::{ - BTreeIndexBuilder, BitmapIndexBuilder, FmIndexBuilder, LabelListIndexBuilder, - ZoneMapIndexBuilder, + BTreeIndexBuilder, BitmapIndexBuilder, BloomFilterIndexBuilder, FmIndexBuilder, + LabelListIndexBuilder, NGramIndexBuilder, RTreeIndexBuilder, ZoneMapIndexBuilder, }, vector::{ IvfHnswFlatIndexBuilder, IvfHnswPqIndexBuilder, IvfHnswSqIndexBuilder, IvfPqIndexBuilder, @@ -63,6 +63,16 @@ pub enum Index { /// prove they cannot match the predicate. ZoneMap(ZoneMapIndexBuilder), + /// An NGram index accelerates substring and pattern filters on UTF-8 strings. + NGram(NGramIndexBuilder), + + /// A Bloom filter index skips groups of scalar values that cannot match a filter. + BloomFilter(BloomFilterIndexBuilder), + + /// An R-tree index accelerates spatial intersection filters on GeoArrow geometries. + /// Native creation requires the `geo` feature. + RTree(RTreeIndexBuilder), + /// Full text search index using BM25. /// /// The posting block size defaults to 128. Supported values are 128 and 256; @@ -352,6 +362,12 @@ pub enum IndexType { Fm, #[serde(alias = "ZONEMAP", alias = "ZONE_MAP")] ZoneMap, + #[serde(alias = "NGRAM", alias = "N_GRAM")] + NGram, + #[serde(alias = "BLOOM_FILTER", alias = "BLOOMFILTER")] + BloomFilter, + #[serde(alias = "RTREE", alias = "R_TREE")] + RTree, // FTS #[serde(alias = "INVERTED", alias = "Inverted")] FTS, @@ -374,6 +390,9 @@ impl std::fmt::Display for IndexType { Self::LabelList => write!(f, "LABEL_LIST"), Self::Fm => write!(f, "FM"), Self::ZoneMap => write!(f, "ZONEMAP"), + Self::NGram => write!(f, "NGRAM"), + Self::BloomFilter => write!(f, "BLOOM_FILTER"), + Self::RTree => write!(f, "RTREE"), Self::FTS => write!(f, "FTS"), Self::Unknown => write!(f, "UNKNOWN"), } @@ -390,6 +409,9 @@ impl std::str::FromStr for IndexType { "LABEL_LIST" | "LABELLIST" => Ok(Self::LabelList), "FM" | "FMINDEX" => Ok(Self::Fm), "ZONEMAP" | "ZONE_MAP" => Ok(Self::ZoneMap), + "NGRAM" | "N_GRAM" => Ok(Self::NGram), + "BLOOM_FILTER" | "BLOOMFILTER" => Ok(Self::BloomFilter), + "RTREE" | "R_TREE" => Ok(Self::RTree), "FTS" | "INVERTED" => Ok(Self::FTS), "IVF_FLAT" => Ok(Self::IvfFlat), "IVF_SQ" => Ok(Self::IvfSq), @@ -495,3 +517,34 @@ pub struct IndexStatistics { /// The number of parts this index is split into. pub num_indices: Option, } + +#[cfg(test)] +mod tests { + use super::IndexType; + + #[test] + fn builtin_scalar_index_type_names() { + for (index_type, canonical, aliases) in [ + (IndexType::NGram, "NGRAM", ["NGram", "NGRAM", "N_GRAM"]), + ( + IndexType::BloomFilter, + "BLOOM_FILTER", + ["BloomFilter", "BLOOMFILTER", "BLOOM_FILTER"], + ), + (IndexType::RTree, "RTREE", ["RTree", "RTREE", "R_TREE"]), + ] { + assert_eq!(index_type.to_string(), canonical); + for alias in aliases { + assert_eq!(alias.parse::().unwrap(), index_type); + assert_eq!( + alias.to_lowercase().parse::().unwrap(), + index_type + ); + assert_eq!( + serde_json::from_value::(serde_json::json!(alias)).unwrap(), + index_type + ); + } + } + } +} diff --git a/rust/lancedb/src/index/scalar.rs b/rust/lancedb/src/index/scalar.rs index af7407a21..5bfc2b4e6 100644 --- a/rust/lancedb/src/index/scalar.rs +++ b/rust/lancedb/src/index/scalar.rs @@ -82,8 +82,178 @@ pub struct FmIndexBuilder {} #[derive(Debug, Clone, Default, serde::Serialize)] pub struct ZoneMapIndexBuilder {} +/// Builder for an NGram index over UTF-8 strings. +/// +/// This index accelerates certain substring, `LIKE`, and regular-expression filters. +/// It uses Lance's default trigram parameters. +/// +/// ``` +/// use lancedb::index::{Index, scalar::NGramIndexBuilder}; +/// # async fn example(table: &lancedb::Table) -> lancedb::Result<()> { +/// table.create_index(&["text"], Index::NGram(NGramIndexBuilder::default())) +/// .execute().await?; +/// # Ok(()) +/// # } +/// ``` +#[derive(Debug, Clone, Default, serde::Serialize)] +pub struct NGramIndexBuilder {} + +/// Builder for a Bloom filter index on scalar values. +/// +/// Bloom filters accelerate equality and membership filters by skipping groups of rows that cannot +/// match. Candidate rows are checked to remove false positives. A bloom filter is much smaller than +/// a btree or bitmap index, but not as precise. It is also limited to equality queries. +/// +/// Unset parameters use Lance's defaults. +/// +/// ``` +/// use lancedb::index::{Index, scalar::BloomFilterIndexBuilder}; +/// # async fn example(table: &lancedb::Table) -> lancedb::Result<()> { +/// let params = BloomFilterIndexBuilder::default() +/// .number_of_items(4096)? +/// .probability(0.01)?; +/// table.create_index(&["id"], Index::BloomFilter(params)) +/// .execute().await?; +/// # Ok(()) +/// # } +/// ``` +#[derive(Debug, Clone, Default, serde::Serialize)] +pub struct BloomFilterIndexBuilder { + #[serde(skip_serializing_if = "Option::is_none")] + number_of_items: Option, + #[serde(skip_serializing_if = "Option::is_none")] + probability: Option, +} + +impl BloomFilterIndexBuilder { + /// Set the number of rows covered by each Bloom filter. + /// + /// Must be greater than zero. Defaults to 8192, unless overridden by Lance's + /// `LANCE_BLOOMFILTER_DEFAULT_NUMBER_OF_ITEMS` environment variable. + pub fn number_of_items(mut self, number_of_items: u64) -> crate::Result { + if number_of_items == 0 { + return Err(crate::Error::InvalidInput { + message: "BloomFilter number_of_items must be greater than zero".into(), + }); + } + self.number_of_items = Some(number_of_items); + Ok(self) + } + + /// Set the desired false-positive probability for each Bloom filter. + /// + /// Must be finite and strictly between zero and one. Lower values use more + /// space. Defaults to 0.00057, unless overridden by Lance's + /// `LANCE_BLOOMFILTER_DEFAULT_PROBABILITY` environment variable. + pub fn probability(mut self, probability: f64) -> crate::Result { + if !probability.is_finite() || probability <= 0.0 || probability >= 1.0 { + return Err(crate::Error::InvalidInput { + message: "BloomFilter probability must be finite and strictly between zero and one" + .into(), + }); + } + self.probability = Some(probability); + Ok(self) + } +} + +/// Builder for an R-tree index on GeoArrow geometry columns. +/// +/// This index accelerates spatial intersection filters using geometry bounding +/// boxes. Unset parameters use Lance's defaults. Native creation requires +/// the `geo` feature; remote creation requires server support. +/// +/// ``` +/// use lancedb::index::{Index, scalar::RTreeIndexBuilder}; +/// # async fn example(table: &lancedb::Table) -> lancedb::Result<()> { +/// let params = RTreeIndexBuilder::default().page_size(1024)?; +/// table.create_index(&["geometry"], Index::RTree(params)) +/// .execute().await?; +/// # Ok(()) +/// # } +/// ``` +#[derive(Debug, Clone, Default, serde::Serialize)] +pub struct RTreeIndexBuilder { + #[serde(skip_serializing_if = "Option::is_none")] + page_size: Option, +} + +impl RTreeIndexBuilder { + /// Set the maximum number of entries in each R-tree page. + /// + /// Must be at least 2. Defaults to 4096. + pub fn page_size(mut self, page_size: u32) -> crate::Result { + if page_size < 2 { + return Err(crate::Error::InvalidInput { + message: "RTree page_size must be at least 2".into(), + }); + } + self.page_size = Some(page_size); + Ok(self) + } +} + pub use lance_index::scalar::FullTextSearchQuery; pub use lance_index::scalar::InvertedIndexParams as FtsIndexBuilder; pub use lance_index::scalar::InvertedIndexParams; pub use lance_index::scalar::inverted::DocumentGranularity; pub use lance_index::scalar::inverted::query::*; + +#[cfg(test)] +mod tests { + use super::*; + + #[test] + fn scalar_index_parameters() { + assert_eq!( + serde_json::to_value(BloomFilterIndexBuilder::default()).unwrap(), + serde_json::json!({}) + ); + assert_eq!( + serde_json::to_value(RTreeIndexBuilder::default()).unwrap(), + serde_json::json!({}) + ); + assert_eq!( + serde_json::to_value( + BloomFilterIndexBuilder::default() + .number_of_items(1) + .unwrap() + ) + .unwrap(), + serde_json::json!({"number_of_items": 1}) + ); + assert_eq!( + serde_json::to_value( + BloomFilterIndexBuilder::default() + .probability(0.01) + .unwrap() + ) + .unwrap(), + serde_json::json!({"probability": 0.01}) + ); + assert!( + BloomFilterIndexBuilder::default() + .number_of_items(0) + .is_err() + ); + for probability in [ + f64::NAN, + f64::INFINITY, + f64::NEG_INFINITY, + -0.1, + 0.0, + 1.0, + 1.1, + ] { + assert!( + BloomFilterIndexBuilder::default() + .probability(probability) + .is_err() + ); + } + for page_size in [0, 1] { + assert!(RTreeIndexBuilder::default().page_size(page_size).is_err()); + } + assert!(RTreeIndexBuilder::default().page_size(2).is_ok()); + } +} diff --git a/rust/lancedb/src/remote/table.rs b/rust/lancedb/src/remote/table.rs index c1182f996..75cf8cd00 100644 --- a/rust/lancedb/src/remote/table.rs +++ b/rust/lancedb/src/remote/table.rs @@ -588,6 +588,9 @@ impl RemoteTable { Index::LabelList(p) => ("LABEL_LIST", Some(to_json(p)?)), Index::Fm(p) => ("FM", Some(to_json(p)?)), Index::ZoneMap(p) => ("ZONEMAP", Some(to_json(p)?)), + Index::NGram(p) => ("NGRAM", Some(to_json(p)?)), + Index::BloomFilter(p) => ("BLOOM_FILTER", Some(to_json(p)?)), + Index::RTree(p) => ("RTREE", Some(to_json(p)?)), Index::FTS(p) => { let mut params = to_json(p)?; if p.get_document_granularity().is_list_element() { @@ -6378,6 +6381,33 @@ mod tests { ("BTREE", json!({}), Index::BTree(Default::default())), ("BITMAP", json!({}), Index::Bitmap(Default::default())), ("ZONEMAP", json!({}), Index::ZoneMap(Default::default())), + ("NGRAM", json!({}), Index::NGram(Default::default())), + ( + "BLOOM_FILTER", + json!({}), + Index::BloomFilter(Default::default()), + ), + ("RTREE", json!({}), Index::RTree(Default::default())), + ( + "BLOOM_FILTER", + json!({"number_of_items": 4096, "probability": 0.01}), + Index::BloomFilter( + crate::index::scalar::BloomFilterIndexBuilder::default() + .number_of_items(4096) + .unwrap() + .probability(0.01) + .unwrap(), + ), + ), + ( + "RTREE", + json!({"page_size": 1024}), + Index::RTree( + crate::index::scalar::RTreeIndexBuilder::default() + .page_size(1024) + .unwrap(), + ), + ), ( "LABEL_LIST", json!({}), diff --git a/rust/lancedb/src/table/create_index.rs b/rust/lancedb/src/table/create_index.rs index a5fb59b0b..0a0d00da1 100644 --- a/rust/lancedb/src/table/create_index.rs +++ b/rust/lancedb/src/table/create_index.rs @@ -265,6 +265,26 @@ impl NativeTable { BuiltinIndexType::ZoneMap, ))) } + Index::NGram(_) => Ok(Box::new(ScalarIndexParams::for_builtin( + BuiltinIndexType::NGram, + ))), + Index::BloomFilter(params) => { + let params = serde_json::to_value(params).map_err(|e| Error::InvalidInput { + message: format!("failed to serialize index params: {e}"), + })?; + Ok(Box::new( + ScalarIndexParams::for_builtin(BuiltinIndexType::BloomFilter) + .with_params(¶ms), + )) + } + Index::RTree(params) => { + let params = serde_json::to_value(params).map_err(|e| Error::InvalidInput { + message: format!("failed to serialize index params: {e}"), + })?; + Ok(Box::new( + ScalarIndexParams::for_builtin(BuiltinIndexType::RTree).with_params(¶ms), + )) + } Index::FTS(fts_opts) => { Self::validate_index_type(field, "FTS", supported_fts_data_type)?; Ok(Box::new(fts_opts)) @@ -425,6 +445,9 @@ impl NativeTable { Index::LabelList(_) => IndexType::LabelList, Index::Fm(_) => IndexType::Fm, Index::ZoneMap(_) => IndexType::ZoneMap, + Index::NGram(_) => IndexType::NGram, + Index::BloomFilter(_) => IndexType::BloomFilter, + Index::RTree(_) => IndexType::RTree, Index::FTS(_) => IndexType::Inverted, Index::IvfFlat(_) | Index::IvfSq(_) @@ -439,6 +462,7 @@ impl NativeTable { #[cfg(test)] mod tests { + use lance::index::DatasetIndexExt; use std::sync::Arc; use std::time::Duration; @@ -1321,6 +1345,191 @@ mod tests { assert_eq!(null_count, 1); } + #[tokio::test] + async fn test_create_builtin_scalar_index() { + for (index, index_type, predicate, invalid_column) in [ + ( + Index::NGram(Default::default()), + crate::index::IndexType::NGram, + "contains(text, 'abc')", + "id", + ), + ( + Index::BloomFilter(Default::default()), + crate::index::IndexType::BloomFilter, + "text = 'abc'", + "flag", + ), + ( + Index::BloomFilter( + crate::index::scalar::BloomFilterIndexBuilder::default() + .number_of_items(2) + .unwrap() + .probability(0.01) + .unwrap(), + ), + crate::index::IndexType::BloomFilter, + "text = 'abc'", + "flag", + ), + ] { + let conn = connect("memory://").execute().await.unwrap(); + let batch = record_batch!( + ("id", Int32, [1, 2, 3, 4]), + ("text", Utf8, [Some("abc"), Some("xyz"), None, Some("")]), + ("flag", Boolean, [true, false, true, false]) + ) + .unwrap(); + let table = conn + .create_table("scalar", batch.clone()) + .execute() + .await + .unwrap(); + table.add(batch).execute().await.unwrap(); + + table + .create_index(&["text"], index.clone()) + .name("text_search".into()) + .train(false) + .execute() + .await + .unwrap(); + assert_eq!( + table.list_indices().await.unwrap()[0].index_type, + index_type + ); + assert!( + table + .create_index(&["text"], index.clone()) + .name("text_search".into()) + .replace(false) + .execute() + .await + .is_err() + ); + table + .create_index(&["text"], index.clone()) + .name("text_search".into()) + .execute() + .await + .unwrap(); + + let stats = table.index_stats("text_search").await.unwrap().unwrap(); + assert_eq!(stats.index_type, index_type); + assert_eq!(stats.num_indexed_rows, 8); + if let Index::BloomFilter(params) = &index { + let expected = serde_json::to_value(params).unwrap(); + let dataset = table.as_native().unwrap().dataset.get().await.unwrap(); + let stats: serde_json::Value = + serde_json::from_str(&dataset.index_statistics("text_search").await.unwrap()) + .unwrap(); + for (key, value) in expected.as_object().unwrap() { + assert_eq!(&stats["indices"][0][key], value); + } + } + for (filter, expected_rows) in [(predicate, 2), ("text IS NULL", 2), ("text = ''", 2)] { + let query = table.query().only_if(filter); + let plan = query.explain_plan(false).await.unwrap(); + if filter == predicate { + assert!(plan.contains("ScalarIndexQuery"), "{plan}"); + } + let batches = query + .execute() + .await + .unwrap() + .try_collect::>() + .await + .unwrap(); + assert_eq!( + batches.iter().map(RecordBatch::num_rows).sum::(), + expected_rows + ); + } + let error = table + .create_index(&[invalid_column], index) + .execute() + .await + .unwrap_err(); + assert!( + error.to_string().to_lowercase().contains( + if index_type == crate::index::IndexType::NGram { + "ngram" + } else { + "bloom" + } + ), + "{error}" + ); + } + } + + #[cfg(feature = "geo")] + #[tokio::test] + async fn test_create_rtree_index() { + let points = StructArray::new( + vec![ + Field::new("x", DataType::Float64, false), + Field::new("y", DataType::Float64, false), + ] + .into(), + vec![ + Arc::new(arrow_array::Float64Array::from(vec![0.0, 10.0, 0.0])), + Arc::new(arrow_array::Float64Array::from(vec![0.0, 10.0, 0.0])), + ], + Some(arrow_buffer::NullBuffer::from(vec![true, true, false])), + ); + let field = Field::new("geometry", points.data_type().clone(), true).with_metadata( + std::collections::HashMap::from([ + ("ARROW:extension:name".into(), "geoarrow.point".into()), + ("ARROW:extension:metadata".into(), "{}".into()), + ]), + ); + let batch = + RecordBatch::try_new(Arc::new(Schema::new(vec![field])), vec![Arc::new(points)]) + .unwrap(); + let conn = connect("memory://").execute().await.unwrap(); + let table = conn.create_table("spatial", batch).execute().await.unwrap(); + table + .create_index( + &["geometry"], + Index::RTree( + crate::index::scalar::RTreeIndexBuilder::default() + .page_size(2) + .unwrap(), + ), + ) + .name("spatial_idx".into()) + .execute() + .await + .unwrap(); + let indices = table.list_indices().await.unwrap(); + assert_eq!(indices.len(), 1); + assert_eq!(indices[0].index_type, crate::index::IndexType::RTree); + let stats = table.index_stats("spatial_idx").await.unwrap().unwrap(); + assert_eq!(stats.index_type, crate::index::IndexType::RTree); + assert_eq!(stats.num_indexed_rows, 3); + let dataset = table.as_native().unwrap().dataset.get().await.unwrap(); + let stats: serde_json::Value = + serde_json::from_str(&dataset.index_statistics("spatial_idx").await.unwrap()).unwrap(); + assert_eq!(stats["indices"][0]["page_size"], 2); + for predicate in [ + "ST_Intersects(geometry, ST_GeomFromText('POLYGON ((-1 -1, 1 -1, 1 1, -1 1, -1 -1))'))", + "geometry IS NULL", + ] { + let query = table.query().only_if(predicate); + let plan = query.explain_plan(false).await.unwrap(); + assert!(plan.contains("ScalarIndexQuery"), "{plan}"); + let batches = query + .execute() + .await + .unwrap() + .try_collect::>() + .await + .unwrap(); + assert_eq!(batches.iter().map(RecordBatch::num_rows).sum::(), 1); + } + } + #[tokio::test] async fn test_create_index_nested_field_paths() { let tmp_dir = tempdir().unwrap();