lance_index/
prefilter.rs

1// SPDX-License-Identifier: Apache-2.0
2// SPDX-FileCopyrightText: Copyright The Lance Authors
3
4use std::sync::Arc;
5
6use async_trait::async_trait;
7use lance_core::utils::mask::RowAddrMask;
8use lance_core::Result;
9
10/// A trait to be implemented by anything supplying a prefilter row addr mask
11///
12/// This trait is for internal use only and has no stability guarantees.
13#[async_trait]
14pub trait FilterLoader: Send + 'static {
15    async fn load(self: Box<Self>) -> Result<RowAddrMask>;
16}
17
18/// Filter out row ids that we know are not relevant to the query.
19///
20/// This could be both rows that are deleted or a prefilter
21/// that should be applied to the search
22///
23/// <section class="warning">
24/// Internal use only. No API stability guarantees.
25/// </section>
26#[async_trait]
27pub trait PreFilter: Send + Sync {
28    /// Waits for the prefilter to be fully loaded
29    ///
30    /// The prefilter loads in the background while the rest of the index
31    /// search is running.  When you are ready to use the prefilter you
32    /// must first call this method to ensure it is fully loaded.  This
33    /// allows `filter_row_ids` to be a synchronous method.
34    async fn wait_for_ready(&self) -> Result<()>;
35
36    /// If the filter is empty.
37    fn is_empty(&self) -> bool;
38
39    /// Get the row addr mask for this prefilter
40    ///
41    /// This method must be called after `wait_for_ready`
42    fn mask(&self) -> Arc<RowAddrMask>;
43
44    /// Check whether a slice of row ids should be included in a query.
45    ///
46    /// Returns a vector of indices into the input slice that should be included,
47    /// also known as a selection vector.
48    ///
49    /// This method must be called after `wait_for_ready`
50    fn filter_row_ids<'a>(&self, row_ids: Box<dyn Iterator<Item = &'a u64> + 'a>) -> Vec<u64>;
51}
52
53/// A prefilter that does nothing
54pub struct NoFilter;
55
56#[async_trait]
57impl PreFilter for NoFilter {
58    async fn wait_for_ready(&self) -> Result<()> {
59        Ok(())
60    }
61
62    fn is_empty(&self) -> bool {
63        true
64    }
65
66    fn mask(&self) -> Arc<RowAddrMask> {
67        Arc::new(RowAddrMask::all_rows())
68    }
69
70    fn filter_row_ids<'a>(&self, row_ids: Box<dyn Iterator<Item = &'a u64> + 'a>) -> Vec<u64> {
71        row_ids.enumerate().map(|(i, _)| i as u64).collect()
72    }
73}
lance_index/prefilter.rs

lance_index/
prefilter.rs
