Compare commits

...

5 Commits

9 changed files with 112 additions and 45 deletions

29
Cargo.lock generated
View File

@ -644,9 +644,9 @@ dependencies = [
[[package]]
name = "charabia"
version = "0.5.0"
version = "0.5.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "4a26a3df4d9c9231eb1e757fe6b1c66c471e0c2cd5410265e7c3109a726663c4"
checksum = "2ed19edcd98f5bf6572f48d6f5982d595cb8718e47c6f0066d942b280575ff02"
dependencies = [
"character_converter",
"cow-utils",
@ -1124,7 +1124,7 @@ dependencies = [
[[package]]
name = "filter-parser"
version = "0.31.1"
source = "git+https://github.com/meilisearch/milli.git?tag=v0.31.1#83ad1aaf0552db9f63fc21ae9fe3976e61577dc8"
source = "git+https://github.com/meilisearch/milli.git?branch=ease-search-results-pagination#ff97a6017c9e409b46df68930b0d2774159b29ad"
dependencies = [
"nom",
"nom_locate",
@ -1149,7 +1149,7 @@ dependencies = [
[[package]]
name = "flatten-serde-json"
version = "0.31.1"
source = "git+https://github.com/meilisearch/milli.git?tag=v0.31.1#83ad1aaf0552db9f63fc21ae9fe3976e61577dc8"
source = "git+https://github.com/meilisearch/milli.git?branch=ease-search-results-pagination#ff97a6017c9e409b46df68930b0d2774159b29ad"
dependencies = [
"serde_json",
]
@ -1662,7 +1662,7 @@ dependencies = [
[[package]]
name = "json-depth-checker"
version = "0.31.1"
source = "git+https://github.com/meilisearch/milli.git?tag=v0.31.1#83ad1aaf0552db9f63fc21ae9fe3976e61577dc8"
source = "git+https://github.com/meilisearch/milli.git?branch=ease-search-results-pagination#ff97a6017c9e409b46df68930b0d2774159b29ad"
dependencies = [
"serde_json",
]
@ -2013,7 +2013,7 @@ dependencies = [
"sha2",
"thiserror",
"time 0.3.9",
"uuid 1.1.2",
"uuid",
]
[[package]]
@ -2082,7 +2082,7 @@ dependencies = [
"tokio",
"tokio-stream",
"urlencoding",
"uuid 1.1.2",
"uuid",
"vergen",
"walkdir",
"yaup",
@ -2146,7 +2146,7 @@ dependencies = [
"thiserror",
"time 0.3.9",
"tokio",
"uuid 1.1.2",
"uuid",
"walkdir",
"whoami",
]
@ -2189,7 +2189,7 @@ dependencies = [
[[package]]
name = "milli"
version = "0.31.1"
source = "git+https://github.com/meilisearch/milli.git?tag=v0.31.1#83ad1aaf0552db9f63fc21ae9fe3976e61577dc8"
source = "git+https://github.com/meilisearch/milli.git?branch=ease-search-results-pagination#ff97a6017c9e409b46df68930b0d2774159b29ad"
dependencies = [
"bimap",
"bincode",
@ -2228,7 +2228,7 @@ dependencies = [
"tempfile",
"thiserror",
"time 0.3.9",
"uuid 0.8.2",
"uuid",
]
[[package]]
@ -3670,15 +3670,6 @@ version = "0.1.6"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "5190c9442dcdaf0ddd50f37420417d219ae5261bbf5db120d0f9bab996c9cba1"
[[package]]
name = "uuid"
version = "0.8.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "bc5cf98d8186244414c848017f0e2676b3fcb46807f6668a97dfe67359a3c4b7"
dependencies = [
"getrandom",
]
[[package]]
name = "uuid"
version = "1.1.2"

View File

@ -7,7 +7,7 @@ edition = "2021"
enum-iterator = "0.7.0"
hmac = "0.12.1"
meilisearch-types = { path = "../meilisearch-types" }
milli = { git = "https://github.com/meilisearch/milli.git", tag = "v0.31.1" }
milli = { git = "https://github.com/meilisearch/milli.git", branch = "ease-search-results-pagination" }
rand = "0.8.4"
serde = { version = "1.0.136", features = ["derive"] }
serde_json = { version = "1.0.79", features = ["preserve_order"] }

View File

@ -10,7 +10,7 @@ use http::header::CONTENT_TYPE;
use meilisearch_auth::SearchRules;
use meilisearch_lib::index::{
SearchQuery, SearchResult, DEFAULT_CROP_LENGTH, DEFAULT_CROP_MARKER,
DEFAULT_HIGHLIGHT_POST_TAG, DEFAULT_HIGHLIGHT_PRE_TAG,
DEFAULT_HIGHLIGHT_POST_TAG, DEFAULT_HIGHLIGHT_PRE_TAG, DEFAULT_SEARCH_LIMIT,
};
use meilisearch_lib::index_controller::Stats;
use meilisearch_lib::MeiliSearch;
@ -369,6 +369,7 @@ pub struct SearchAggregator {
// pagination
max_limit: usize,
max_offset: usize,
finite_pagination: bool,
// formatting
highlight_pre_tag: bool,
@ -423,8 +424,15 @@ impl SearchAggregator {
ret.max_terms_number = q.split_whitespace().count();
}
ret.max_limit = query.limit;
ret.max_offset = query.offset.unwrap_or_default();
if query.limit.is_none() && query.offset.is_none() {
ret.max_limit = query.hits_per_page;
ret.max_offset = query.page.saturating_sub(1) * query.hits_per_page;
ret.finite_pagination = true;
} else {
ret.max_limit = query.limit.unwrap_or_else(DEFAULT_SEARCH_LIMIT);
ret.max_offset = query.offset.unwrap_or_default();
ret.finite_pagination = false;
}
ret.highlight_pre_tag = query.highlight_pre_tag != DEFAULT_HIGHLIGHT_PRE_TAG();
ret.highlight_post_tag = query.highlight_post_tag != DEFAULT_HIGHLIGHT_POST_TAG();

View File

@ -3,7 +3,7 @@ use log::debug;
use meilisearch_auth::IndexSearchRules;
use meilisearch_lib::index::{
SearchQuery, DEFAULT_CROP_LENGTH, DEFAULT_CROP_MARKER, DEFAULT_HIGHLIGHT_POST_TAG,
DEFAULT_HIGHLIGHT_PRE_TAG, DEFAULT_SEARCH_LIMIT,
DEFAULT_HIGHLIGHT_PRE_TAG, DEFAULT_HIT_PER_PAGE, DEFAULT_PAGE,
};
use meilisearch_lib::MeiliSearch;
use meilisearch_types::error::ResponseError;
@ -29,6 +29,10 @@ pub struct SearchQueryGet {
q: Option<String>,
offset: Option<usize>,
limit: Option<usize>,
#[serde(default = "DEFAULT_PAGE")]
page: usize,
#[serde(default = "DEFAULT_HIT_PER_PAGE")]
hits_per_page: usize,
attributes_to_retrieve: Option<CS<String>>,
attributes_to_crop: Option<CS<String>>,
#[serde(default = "DEFAULT_CROP_LENGTH")]
@ -60,7 +64,9 @@ impl From<SearchQueryGet> for SearchQuery {
Self {
q: other.q,
offset: other.offset,
limit: other.limit.unwrap_or_else(DEFAULT_SEARCH_LIMIT),
limit: other.limit,
page: other.page,
hits_per_page: other.hits_per_page,
attributes_to_retrieve: other
.attributes_to_retrieve
.map(|o| o.into_iter().collect()),

View File

@ -28,7 +28,7 @@ lazy_static = "1.4.0"
log = "0.4.14"
meilisearch-auth = { path = "../meilisearch-auth" }
meilisearch-types = { path = "../meilisearch-types" }
milli = { git = "https://github.com/meilisearch/milli.git", tag = "v0.31.1" }
milli = { git = "https://github.com/meilisearch/milli.git", branch = "ease-search-results-pagination" }
mime = "0.3.16"
num_cpus = "1.13.1"
obkv = "0.2.0"

View File

@ -25,6 +25,8 @@ impl ErrorCode for MilliError<'_> {
// TODO: wait for spec for new error codes.
UserError::SerdeJson(_)
| UserError::DocumentLimitReached
// TODO: REMOVE ME
| UserError::AccessingSoftDeletedDocument { .. }
| UserError::UnknownInternalDocumentId { .. } => Code::Internal,
UserError::InvalidStoreFile => Code::InvalidStore,
UserError::NoSpaceLeftOnDevice => Code::NoSpaceLeftOnDevice,

View File

@ -1,6 +1,7 @@
pub use search::{
SearchQuery, SearchResult, DEFAULT_CROP_LENGTH, DEFAULT_CROP_MARKER,
DEFAULT_HIGHLIGHT_POST_TAG, DEFAULT_HIGHLIGHT_PRE_TAG, DEFAULT_SEARCH_LIMIT,
HitsInfo, SearchQuery, SearchResult, DEFAULT_CROP_LENGTH, DEFAULT_CROP_MARKER,
DEFAULT_HIGHLIGHT_POST_TAG, DEFAULT_HIGHLIGHT_PRE_TAG, DEFAULT_HIT_PER_PAGE, DEFAULT_PAGE,
DEFAULT_SEARCH_LIMIT,
};
pub use updates::{apply_settings_to_builder, Checked, Facets, Settings, Unchecked};

View File

@ -1,4 +1,4 @@
use std::cmp::min;
use std::cmp::{max, min};
use std::collections::{BTreeMap, BTreeSet, HashSet};
use std::str::FromStr;
use std::time::Instant;
@ -26,6 +26,8 @@ pub const DEFAULT_CROP_LENGTH: fn() -> usize = || 10;
pub const DEFAULT_CROP_MARKER: fn() -> String = || "…".to_string();
pub const DEFAULT_HIGHLIGHT_PRE_TAG: fn() -> String = || "<em>".to_string();
pub const DEFAULT_HIGHLIGHT_POST_TAG: fn() -> String = || "</em>".to_string();
pub const DEFAULT_PAGE: fn() -> usize = || 1;
pub const DEFAULT_HIT_PER_PAGE: fn() -> usize = || 20;
/// The maximimum number of results that the engine
/// will be able to return in one search call.
@ -36,8 +38,11 @@ pub const DEFAULT_PAGINATION_MAX_TOTAL_HITS: usize = 1000;
pub struct SearchQuery {
pub q: Option<String>,
pub offset: Option<usize>,
#[serde(default = "DEFAULT_SEARCH_LIMIT")]
pub limit: usize,
pub limit: Option<usize>,
#[serde(default = "DEFAULT_PAGE")]
pub page: usize,
#[serde(default = "DEFAULT_HIT_PER_PAGE")]
pub hits_per_page: usize,
pub attributes_to_retrieve: Option<BTreeSet<String>>,
pub attributes_to_crop: Option<Vec<String>>,
#[serde(default = "DEFAULT_CROP_LENGTH")]
@ -71,15 +76,31 @@ pub struct SearchHit {
#[serde(rename_all = "camelCase")]
pub struct SearchResult {
pub hits: Vec<SearchHit>,
pub estimated_total_hits: u64,
pub query: String,
pub limit: usize,
pub offset: usize,
pub processing_time_ms: u128,
#[serde(flatten)]
pub hits_info: HitsInfo,
#[serde(skip_serializing_if = "Option::is_none")]
pub facet_distribution: Option<BTreeMap<String, BTreeMap<String, u64>>>,
}
#[derive(Serialize, Debug, Clone, PartialEq)]
#[serde(untagged)]
pub enum HitsInfo {
#[serde(rename_all = "camelCase")]
Pagination {
hits_per_page: usize,
page: usize,
total_pages: usize,
},
#[serde(rename_all = "camelCase")]
OffsetLimit {
limit: usize,
offset: usize,
estimated_total_hits: usize,
},
}
impl Index {
pub fn perform_search(&self, query: SearchQuery) -> Result<SearchResult> {
let before_search = Instant::now();
@ -97,8 +118,28 @@ impl Index {
// Make sure that a user can't get more documents than the hard limit,
// we align that on the offset too.
let offset = min(query.offset.unwrap_or(0), max_total_hits);
let limit = min(query.limit, max_total_hits.saturating_sub(offset));
let is_finite_pagination = query.offset.is_none() && query.limit.is_none();
search.exhaustive_number_hits(is_finite_pagination);
let (offset, limit) = if is_finite_pagination {
// we start at least at page 1.
let page = max(query.page, 1);
// return at least 1 document.
let hits_per_page = max(query.hits_per_page, 1);
let offset = min(hits_per_page * (page - 1), max_total_hits);
let limit = min(hits_per_page, max_total_hits.saturating_sub(offset));
(offset, limit)
} else {
let offset = min(query.offset.unwrap_or(0), max_total_hits);
let limit = min(
query.limit.unwrap_or_else(DEFAULT_SEARCH_LIMIT),
max_total_hits.saturating_sub(offset),
);
(offset, limit)
};
search.offset(offset);
search.limit(limit);
@ -223,7 +264,23 @@ impl Index {
documents.push(hit);
}
let estimated_total_hits = candidates.len();
let number_of_hits = min(candidates.len() as usize, max_total_hits);
let hits_info = if is_finite_pagination {
// return at least 1 document.
let hits_per_page = max(query.hits_per_page, 1);
HitsInfo::Pagination {
hits_per_page,
page: offset / hits_per_page + 1,
// TODO @many: estimation for now but we should ask milli to return an exact value
total_pages: (number_of_hits + hits_per_page - 1) / query.hits_per_page,
}
} else {
HitsInfo::OffsetLimit {
limit: query.limit.unwrap_or_else(DEFAULT_SEARCH_LIMIT),
offset,
estimated_total_hits: number_of_hits,
}
};
let facet_distribution = match query.facets {
Some(ref fields) => {
@ -246,10 +303,8 @@ impl Index {
let result = SearchResult {
hits: documents,
estimated_total_hits,
hits_info,
query: query.q.clone().unwrap_or_default(),
limit: query.limit,
offset: query.offset.unwrap_or_default(),
processing_time_ms: before_search.elapsed().as_millis(),
facet_distribution,
};

View File

@ -652,7 +652,7 @@ mod test {
use nelson::Mocker;
use crate::index::error::Result as IndexResult;
use crate::index::Index;
use crate::index::{HitsInfo, Index};
use crate::index::{
DEFAULT_CROP_MARKER, DEFAULT_HIGHLIGHT_POST_TAG, DEFAULT_HIGHLIGHT_PRE_TAG,
};
@ -685,7 +685,9 @@ mod test {
let query = SearchQuery {
q: Some(String::from("hello world")),
offset: Some(10),
limit: 0,
limit: Some(0),
page: 1,
hits_per_page: 10,
attributes_to_retrieve: Some(vec!["string".to_owned()].into_iter().collect()),
attributes_to_crop: None,
crop_length: 18,
@ -701,10 +703,12 @@ mod test {
let result = SearchResult {
hits: vec![],
estimated_total_hits: 29,
query: "hello world".to_string(),
limit: 24,
offset: 0,
hits_info: HitsInfo::OffsetLimit {
limit: 24,
offset: 0,
estimated_total_hits: 29,
},
processing_time_ms: 50,
facet_distribution: None,
};