From 7be275b6925d36e3a49dfaf8dcb98ab7a789af15 Mon Sep 17 00:00:00 2001 From: Kerollmops Date: Tue, 26 Jan 2021 14:14:37 +0100 Subject: [PATCH] Add the count to the facet distribution --- http-ui/Cargo.lock | 1 + http-ui/public/script.js | 2 +- http-ui/src/main.rs | 5 ++-- src/facet/facet_value.rs | 22 +++++++++++++-- src/search/facet/facet_distribution.rs | 38 ++++++++++++++------------ 5 files changed, 46 insertions(+), 22 deletions(-) diff --git a/http-ui/Cargo.lock b/http-ui/Cargo.lock index 1ef0ccfbd..cb5103003 100644 --- a/http-ui/Cargo.lock +++ b/http-ui/Cargo.lock @@ -1335,6 +1335,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "9fe9037165d7023b1228bc4ae9a2fa1a2b0095eca6c2998c624723dfd01314a5" dependencies = [ "num-traits", + "serde", ] [[package]] diff --git a/http-ui/public/script.js b/http-ui/public/script.js index f935cc07d..9887f06ce 100644 --- a/http-ui/public/script.js +++ b/http-ui/public/script.js @@ -57,7 +57,7 @@ $('#query, #filters').on('input', function () { } // Create the newly discovered facets - let diff = diffArray(data.facets[facet_name], selected_values); + let diff = diffArray(Object.keys(data.facets[facet_name]), selected_values); for (value of diff) { let option = $('') .text(value) diff --git a/http-ui/src/main.rs b/http-ui/src/main.rs index 6e4f42a4e..54f87c3e6 100644 --- a/http-ui/src/main.rs +++ b/http-ui/src/main.rs @@ -1,4 +1,4 @@ -use std::collections::{HashMap, HashSet}; +use std::collections::{BTreeMap, HashMap, HashSet}; use std::fmt::Display; use std::fs::{File, create_dir_all}; use std::net::SocketAddr; @@ -29,6 +29,7 @@ use warp::filters::ws::Message; use warp::{Filter, http::Response}; use meilisearch_tokenizer::{Analyzer, AnalyzerConfig}; +use milli::facet::FacetValue; use milli::update::UpdateIndexingStep::*; use milli::update::{UpdateBuilder, IndexDocumentsMethod, UpdateFormat}; use milli::{obkv_to_json, Index, UpdateStore, SearchResult, FacetCondition}; @@ -652,7 +653,7 @@ async fn main() -> anyhow::Result<()> { struct Answer { documents: Vec>, number_of_candidates: u64, - facets: HashMap>, + facets: BTreeMap>, } let disable_highlighting = opt.disable_highlighting; diff --git a/src/facet/facet_value.rs b/src/facet/facet_value.rs index 823eddcee..9630c6634 100644 --- a/src/facet/facet_value.rs +++ b/src/facet/facet_value.rs @@ -1,8 +1,7 @@ use ordered_float::OrderedFloat; -use serde::{Serialize, Deserialize}; +use serde::{Serialize, Serializer}; #[derive(Debug, Clone, PartialOrd, Ord, PartialEq, Eq, Hash)] -#[derive(Serialize, Deserialize)] pub enum FacetValue { String(String), Float(OrderedFloat), @@ -38,3 +37,22 @@ impl From for FacetValue { FacetValue::Integer(integer) } } + +impl Serialize for FacetValue { + fn serialize(&self, serializer: S) -> Result + where + S: Serializer, + { + match self { + FacetValue::String(string) => serializer.serialize_str(string), + FacetValue::Float(float) => { + let string = float.to_string(); + serializer.serialize_str(&string) + }, + FacetValue::Integer(integer) => { + let string = integer.to_string(); + serializer.serialize_str(&string) + }, + } + } +} diff --git a/src/search/facet/facet_distribution.rs b/src/search/facet/facet_distribution.rs index 244169fc2..55b529308 100644 --- a/src/search/facet/facet_distribution.rs +++ b/src/search/facet/facet_distribution.rs @@ -1,4 +1,4 @@ -use std::collections::{HashSet, BTreeSet, BTreeMap}; +use std::collections::{HashSet, BTreeMap}; use std::ops::Bound::Unbounded; use std::{cmp, fmt}; @@ -38,13 +38,18 @@ impl<'a> FacetDistribution<'a> { self } - fn facet_values(&self, field_id: FieldId, facet_type: FacetType) -> heed::Result> { + fn facet_values( + &self, + field_id: FieldId, + facet_type: FacetType, + ) -> heed::Result> + { if let Some(candidates) = self.candidates.as_ref() { if candidates.len() <= 1000 { let mut key_buffer = vec![field_id]; match facet_type { FacetType::Float => { - let mut facet_values = BTreeSet::new(); + let mut facet_values = BTreeMap::new(); for docid in candidates { key_buffer.truncate(1); key_buffer.extend_from_slice(&docid.to_be_bytes()); @@ -53,13 +58,13 @@ impl<'a> FacetDistribution<'a> { .remap_key_type::(); for result in iter { let ((_, _, value), ()) = result?; - facet_values.insert(FacetValue::from(value)); + *facet_values.entry(FacetValue::from(value)).or_insert(0) += 1; } } Ok(facet_values) }, FacetType::Integer => { - let mut facet_values = BTreeSet::new(); + let mut facet_values = BTreeMap::new(); for docid in candidates { key_buffer.truncate(1); key_buffer.extend_from_slice(&docid.to_be_bytes()); @@ -68,13 +73,13 @@ impl<'a> FacetDistribution<'a> { .remap_key_type::(); for result in iter { let ((_, _, value), ()) = result?; - facet_values.insert(FacetValue::from(value)); + *facet_values.entry(FacetValue::from(value)).or_insert(0) += 1; } } Ok(facet_values) }, FacetType::String => { - let mut facet_values = BTreeSet::new(); + let mut facet_values = BTreeMap::new(); for docid in candidates { key_buffer.truncate(1); key_buffer.extend_from_slice(&docid.to_be_bytes()); @@ -83,7 +88,7 @@ impl<'a> FacetDistribution<'a> { .remap_key_type::(); for result in iter { let ((_, _, value), ()) = result?; - facet_values.insert(FacetValue::from(value)); + *facet_values.entry(FacetValue::from(value)).or_insert(0) += 1; } } Ok(facet_values) @@ -113,11 +118,12 @@ impl<'a> FacetDistribution<'a> { }, }; - let mut facet_values = BTreeSet::new(); + let mut facet_values = BTreeMap::new(); for result in iter { - let (value, docids) = result?; - if self.candidates.as_ref().map_or(true, |c| docids.is_disjoint(c)) { - facet_values.insert(value); + let (value, mut docids) = result?; + docids.intersect_with(candidates); + if !docids.is_empty() { + facet_values.insert(value, docids.len()); } if facet_values.len() == self.max_values_by_facet { break; @@ -152,12 +158,10 @@ impl<'a> FacetDistribution<'a> { }, }; - let mut facet_values = BTreeSet::new(); + let mut facet_values = BTreeMap::new(); for result in iter { let (value, docids) = result?; - if self.candidates.as_ref().map_or(true, |c| docids.is_disjoint(c)) { - facet_values.insert(value); - } + facet_values.insert(value, docids.len()); if facet_values.len() == self.max_values_by_facet { break; } @@ -167,7 +171,7 @@ impl<'a> FacetDistribution<'a> { } } - pub fn execute(&self) -> heed::Result>> { + pub fn execute(&self) -> heed::Result>> { let fields_ids_map = self.index.fields_ids_map(self.rtxn)?; let faceted_fields = self.index.faceted_fields(self.rtxn)?; let fields_ids: Vec<_> = match &self.facets {