searchengines-web/MeiliSearch
searchengines-web/MeiliSearch
1
0
Fork 0
mirror of https://github.com/meilisearch/MeiliSearch synced 2025-07-04 20:37:15 +02:00
Code Issues Releases Wiki Activity

Use the Error enum everywhere in the project

Browse source
This commit is contained in:
Kerollmops 2021-06-14 16:46:19 +02:00
parent ca78cb5aca
commit 312c2d1d8e
No known key found for this signature in database
GPG key ID: 92ADA4E935E71FA4
35 changed files with 385 additions and 300 deletions
Download patch file Download diff file Expand all files Collapse all files

21
milli/src/search/criteria/asc_desc.rs
View file

@ -1,15 +1,15 @@
use std::mem::take;
use anyhow::Context;
use itertools::Itertools;
use log::debug;
use ordered_float::OrderedFloat;
use roaring::RoaringBitmap;
use crate::error::FieldIdMapMissingEntry;
use crate::search::criteria::{resolve_query_tree, CriteriaBuilder};
use crate::search::facet::FacetIter;
use crate::search::query_tree::Operation;
use crate::{FieldId, Index};
use crate::{FieldId, Index, Result};
use super::{Criterion, CriterionParameters, CriterionResult};
/// Threshold on the number of candidates that will make
@ -36,7 +36,7 @@ impl<'t> AscDesc<'t> {
rtxn: &'t heed::RoTxn,
parent: Box<dyn Criterion + 't>,
field_name: String,
) -> anyhow::Result<Self> {
) -> Result<Self> {
Self::new(index, rtxn, parent, field_name, true)
}
@ -45,7 +45,7 @@ impl<'t> AscDesc<'t> {
rtxn: &'t heed::RoTxn,
parent: Box<dyn Criterion + 't>,
field_name: String,
) -> anyhow::Result<Self> {
) -> Result<Self> {
Self::new(index, rtxn, parent, field_name, false)
}
@ -55,11 +55,14 @@ impl<'t> AscDesc<'t> {
parent: Box<dyn Criterion + 't>,
field_name: String,
ascending: bool,
) -> anyhow::Result<Self> {
) -> Result<Self> {
let fields_ids_map = index.fields_ids_map(rtxn)?;
let field_id = fields_ids_map
.id(&field_name)
.with_context(|| format!("field {:?} isn't registered", field_name))?;
.ok_or_else(|| FieldIdMapMissingEntry::FieldName {
field_name: field_name.clone(),
from_db_name: "asc-desc",
})?;
Ok(AscDesc {
index,
@ -79,7 +82,7 @@ impl<'t> AscDesc<'t> {
impl<'t> Criterion for AscDesc<'t> {
#[logging_timer::time("AscDesc::{}")]
fn next(&mut self, params: &mut CriterionParameters) -> anyhow::Result<Option<CriterionResult>> {
fn next(&mut self, params: &mut CriterionParameters) -> Result<Option<CriterionResult>> {
// remove excluded candidates when next is called, instead of doing it in the loop.
self.allowed_candidates -= params.excluded_candidates;
@ -162,7 +165,7 @@ fn facet_ordered<'t>(
field_id: FieldId,
ascending: bool,
candidates: RoaringBitmap,
) -> anyhow::Result<Box<dyn Iterator<Item = heed::Result<RoaringBitmap>> + 't>> {
) -> Result<Box<dyn Iterator<Item = heed::Result<RoaringBitmap>> + 't>> {
if candidates.len() <= CANDIDATES_THRESHOLD {
let iter = iterative_facet_ordered_iter(index, rtxn, field_id, ascending, candidates)?;
Ok(Box::new(iter.map(Ok)) as Box<dyn Iterator<Item = _>>)
@ -186,7 +189,7 @@ fn iterative_facet_ordered_iter<'t>(
field_id: FieldId,
ascending: bool,
candidates: RoaringBitmap,
) -> anyhow::Result<impl Iterator<Item = RoaringBitmap> + 't> {
) -> Result<impl Iterator<Item = RoaringBitmap> + 't> {
let mut docids_values = Vec::with_capacity(candidates.len() as usize);
for docid in candidates.iter() {
let left = (field_id, docid, f64::MIN);

17
milli/src/search/criteria/attribute.rs
View file

@ -5,7 +5,7 @@ use std::mem::take;
use roaring::RoaringBitmap;
use crate::{TreeLevel, search::build_dfa};
use crate::{TreeLevel, Result, search::build_dfa};
use crate::search::criteria::Query;
use crate::search::query_tree::{Operation, QueryKind};
use crate::search::{word_derivations, WordDerivationsCache};
@ -48,7 +48,7 @@ impl<'t> Attribute<'t> {
impl<'t> Criterion for Attribute<'t> {
#[logging_timer::time("Attribute::{}")]
fn next(&mut self, params: &mut CriterionParameters) -> anyhow::Result<Option<CriterionResult>> {
fn next(&mut self, params: &mut CriterionParameters) -> Result<Option<CriterionResult>> {
// remove excluded candidates when next is called, instead of doing it in the loop.
if let Some((_, _, allowed_candidates)) = self.state.as_mut() {
*allowed_candidates -= params.excluded_candidates;
@ -224,7 +224,12 @@ struct QueryLevelIterator<'t, 'q> {
}
impl<'t, 'q> QueryLevelIterator<'t, 'q> {
fn new(ctx: &'t dyn Context<'t>, queries: &'q [Query], wdcache: &mut WordDerivationsCache) -> anyhow::Result<Option<Self>> {
fn new(
ctx: &'t dyn Context<'t>,
queries: &'q [Query],
wdcache: &mut WordDerivationsCache,
) -> Result<Option<Self>>
{
let mut inner = Vec::with_capacity(queries.len());
for query in queries {
match &query.kind {
@ -471,7 +476,7 @@ fn initialize_query_level_iterators<'t, 'q>(
branches: &'q FlattenedQueryTree,
allowed_candidates: &RoaringBitmap,
wdcache: &mut WordDerivationsCache,
) -> anyhow::Result<BinaryHeap<Branch<'t, 'q>>> {
) -> Result<BinaryHeap<Branch<'t, 'q>>> {
let mut positions = BinaryHeap::with_capacity(branches.len());
for branch in branches {
@ -521,7 +526,7 @@ fn set_compute_candidates<'t>(
branches: &FlattenedQueryTree,
allowed_candidates: &RoaringBitmap,
wdcache: &mut WordDerivationsCache,
) -> anyhow::Result<Option<RoaringBitmap>>
) -> Result<Option<RoaringBitmap>>
{
let mut branches_heap = initialize_query_level_iterators(ctx, branches, allowed_candidates, wdcache)?;
let lowest_level = TreeLevel::min_value();
@ -573,7 +578,7 @@ fn linear_compute_candidates(
ctx: &dyn Context,
branches: &FlattenedQueryTree,
allowed_candidates: &RoaringBitmap,
) -> anyhow::Result<BTreeMap<u64, RoaringBitmap>>
) -> Result<BTreeMap<u64, RoaringBitmap>>
{
fn compute_candidate_rank(branches: &FlattenedQueryTree, words_positions: HashMap<String, RoaringBitmap>) -> u64 {
let mut min_rank = u64::max_value();

6
milli/src/search/criteria/exactness.rs
View file

@ -14,7 +14,7 @@ use crate::search::criteria::{
CriterionResult,
resolve_query_tree,
};
use crate::TreeLevel;
use crate::{TreeLevel, Result};
pub struct Exactness<'t> {
ctx: &'t dyn Context<'t>,
@ -45,7 +45,7 @@ impl<'t> Exactness<'t> {
impl<'t> Criterion for Exactness<'t> {
#[logging_timer::time("Exactness::{}")]
fn next(&mut self, params: &mut CriterionParameters) -> anyhow::Result<Option<CriterionResult>> {
fn next(&mut self, params: &mut CriterionParameters) -> Result<Option<CriterionResult>> {
// remove excluded candidates when next is called, instead of doing it in the loop.
if let Some(state) = self.state.as_mut() {
state.difference_with(params.excluded_candidates);
@ -158,7 +158,7 @@ fn resolve_state(
ctx: &dyn Context,
state: State,
query: &[ExactQueryPart],
) -> anyhow::Result<(RoaringBitmap, Option<State>)>
) -> Result<(RoaringBitmap, Option<State>)>
{
use State::*;
match state {

3
milli/src/search/criteria/final.rs
View file

@ -1,6 +1,7 @@
use log::debug;
use roaring::RoaringBitmap;
use crate::Result;
use crate::search::query_tree::Operation;
use crate::search::WordDerivationsCache;
use super::{resolve_query_tree, Criterion, CriterionResult, CriterionParameters, Context};
@ -29,7 +30,7 @@ impl<'t> Final<'t> {
}
#[logging_timer::time("Final::{}")]
pub fn next(&mut self, excluded_candidates: &RoaringBitmap) -> anyhow::Result<Option<FinalResult>> {
pub fn next(&mut self, excluded_candidates: &RoaringBitmap) -> Result<Option<FinalResult>> {
debug!("Final iteration");
let excluded_candidates = &self.returned_candidates | excluded_candidates;
let mut criterion_parameters = CriterionParameters {

4
milli/src/search/criteria/initial.rs
View file

@ -1,7 +1,7 @@
use roaring::RoaringBitmap;
use crate::Result;
use crate::search::query_tree::Operation;
use super::{Criterion, CriterionResult, CriterionParameters};
pub struct Initial {
@ -22,7 +22,7 @@ impl Initial {
impl Criterion for Initial {
#[logging_timer::time("Initial::{}")]
fn next(&mut self, _: &mut CriterionParameters) -> anyhow::Result<Option<CriterionResult>> {
fn next(&mut self, _: &mut CriterionParameters) -> Result<Option<CriterionResult>> {
Ok(self.answer.take())
}
}

27
milli/src/search/criteria/mod.rs
View file

@ -4,7 +4,7 @@ use std::borrow::Cow;
use roaring::RoaringBitmap;
use crate::{FieldId, TreeLevel, search::{word_derivations, WordDerivationsCache}};
use crate::{Index, DocumentId};
use crate::{Index, DocumentId, Result};
use super::query_tree::{Operation, PrimitiveQueryPart, Query, QueryKind};
use self::asc_desc::AscDesc;
@ -26,7 +26,7 @@ mod words;
pub mod r#final;
pub trait Criterion {
fn next(&mut self, params: &mut CriterionParameters) -> anyhow::Result<Option<CriterionResult>>;
fn next(&mut self, params: &mut CriterionParameters) -> Result<Option<CriterionResult>>;
}
/// The result of a call to the parent criterion.
@ -78,8 +78,9 @@ pub trait Context<'c> {
fn synonyms(&self, word: &str) -> heed::Result<Option<Vec<Vec<String>>>>;
fn searchable_fields_ids(&self) -> heed::Result<Vec<FieldId>>;
fn field_id_word_count_docids(&self, field_id: FieldId, word_count: u8) -> heed::Result<Option<RoaringBitmap>>;
fn word_level_position_docids(&self, word: &str, level: TreeLevel, left: u32, right: u32) -> Result<Option<RoaringBitmap>, heed::Error>;
fn word_level_position_docids(&self, word: &str, level: TreeLevel, left: u32, right: u32) -> heed::Result<Option<RoaringBitmap>>;
}
pub struct CriteriaBuilder<'t> {
rtxn: &'t heed::RoTxn<'t>,
index: &'t Index,
@ -185,14 +186,14 @@ impl<'c> Context<'c> for CriteriaBuilder<'c> {
self.index.field_id_word_count_docids.get(self.rtxn, &key)
}
fn word_level_position_docids(&self, word: &str, level: TreeLevel, left: u32, right: u32) -> Result<Option<RoaringBitmap>, heed::Error> {
fn word_level_position_docids(&self, word: &str, level: TreeLevel, left: u32, right: u32) -> heed::Result<Option<RoaringBitmap>> {
let key = (word, level, left, right);
self.index.word_level_position_docids.get(self.rtxn, &key)
}
}
impl<'t> CriteriaBuilder<'t> {
pub fn new(rtxn: &'t heed::RoTxn<'t>, index: &'t Index) -> anyhow::Result<Self> {
pub fn new(rtxn: &'t heed::RoTxn<'t>, index: &'t Index) -> Result<Self> {
let words_fst = index.words_fst(rtxn)?;
let words_prefixes_fst = index.words_prefixes_fst(rtxn)?;
Ok(Self { rtxn, index, words_fst, words_prefixes_fst })
@ -203,7 +204,7 @@ impl<'t> CriteriaBuilder<'t> {
query_tree: Option<Operation>,
primitive_query: Option<Vec<PrimitiveQueryPart>>,
filtered_candidates: Option<RoaringBitmap>,
) -> anyhow::Result<Final<'t>>
) -> Result<Final<'t>>
{
use crate::criterion::Criterion as Name;
@ -230,13 +231,13 @@ pub fn resolve_query_tree<'t>(
ctx: &'t dyn Context,
query_tree: &Operation,
wdcache: &mut WordDerivationsCache,
) -> anyhow::Result<RoaringBitmap>
) -> Result<RoaringBitmap>
{
fn resolve_operation<'t>(
ctx: &'t dyn Context,
query_tree: &Operation,
wdcache: &mut WordDerivationsCache,
) -> anyhow::Result<RoaringBitmap>
) -> Result<RoaringBitmap>
{
use Operation::{And, Phrase, Or, Query};
@ -244,7 +245,7 @@ pub fn resolve_query_tree<'t>(
And(ops) => {
let mut ops = ops.iter().map(|op| {
resolve_operation(ctx, op, wdcache)
}).collect::<anyhow::Result<Vec<_>>>()?;
}).collect::<Result<Vec<_>>>()?;
ops.sort_unstable_by_key(|cds| cds.len());
@ -302,7 +303,7 @@ fn all_word_pair_proximity_docids<T: AsRef<str>, U: AsRef<str>>(
left_words: &[(T, u8)],
right_words: &[(U, u8)],
proximity: u8
) -> anyhow::Result<RoaringBitmap>
) -> Result<RoaringBitmap>
{
let mut docids = RoaringBitmap::new();
for (left, _l_typo) in left_words {
@ -318,7 +319,7 @@ fn query_docids(
ctx: &dyn Context,
query: &Query,
wdcache: &mut WordDerivationsCache,
) -> anyhow::Result<RoaringBitmap>
) -> Result<RoaringBitmap>
{
match &query.kind {
QueryKind::Exact { word, .. } => {
@ -354,7 +355,7 @@ fn query_pair_proximity_docids(
right: &Query,
proximity: u8,
wdcache: &mut WordDerivationsCache,
) -> anyhow::Result<RoaringBitmap>
) -> Result<RoaringBitmap>
{
if proximity >= 8 {
let mut candidates = query_docids(ctx, left, wdcache)?;
@ -481,7 +482,7 @@ pub mod test {
todo!()
}
fn word_level_position_docids(&self, _word: &str, _level: TreeLevel, _left: u32, _right: u32) -> Result<Option<RoaringBitmap>, heed::Error> {
fn word_level_position_docids(&self, _word: &str, _level: TreeLevel, _left: u32, _right: u32) -> heed::Result<Option<RoaringBitmap>> {
todo!()
}

19
milli/src/search/criteria/proximity.rs
View file

@ -5,9 +5,10 @@ use std::mem::take;
use roaring::RoaringBitmap;
use log::debug;
use crate::{DocumentId, Position, search::{query_tree::QueryKind}};
use crate::search::query_tree::{maximum_proximity, Operation, Query};
use crate::search::{build_dfa, WordDerivationsCache};
use crate::search::{query_tree::QueryKind};
use crate::{DocumentId, Position, Result};
use super::{
Context,
Criterion,
@ -55,7 +56,7 @@ impl<'t> Proximity<'t> {
impl<'t> Criterion for Proximity<'t> {
#[logging_timer::time("Proximity::{}")]
fn next(&mut self, params: &mut CriterionParameters) -> anyhow::Result<Option<CriterionResult>> {
fn next(&mut self, params: &mut CriterionParameters) -> Result<Option<CriterionResult>> {
// remove excluded candidates when next is called, instead of doing it in the loop.
if let Some((_, _, allowed_candidates)) = self.state.as_mut() {
*allowed_candidates -= params.excluded_candidates;
@ -161,7 +162,7 @@ fn resolve_candidates<'t>(
proximity: u8,
cache: &mut Cache,
wdcache: &mut WordDerivationsCache,
) -> anyhow::Result<RoaringBitmap>
) -> Result<RoaringBitmap>
{
fn resolve_operation<'t>(
ctx: &'t dyn Context,
@ -169,7 +170,7 @@ fn resolve_candidates<'t>(
proximity: u8,
cache: &mut Cache,
wdcache: &mut WordDerivationsCache,
) -> anyhow::Result<Vec<(Query, Query, RoaringBitmap)>>
) -> Result<Vec<(Query, Query, RoaringBitmap)>>
{
use Operation::{And, Phrase, Or};
@ -227,7 +228,7 @@ fn resolve_candidates<'t>(
proximity: u8,
cache: &mut Cache,
wdcache: &mut WordDerivationsCache,
) -> anyhow::Result<Vec<(Query, Query, RoaringBitmap)>>
) -> Result<Vec<(Query, Query, RoaringBitmap)>>
{
fn pair_combinations(mana: u8, left_max: u8) -> impl Iterator<Item = (u8, u8)> {
(0..=mana.min(left_max)).map(move |m| (m, mana - m))
@ -281,7 +282,7 @@ fn resolve_candidates<'t>(
proximity: u8,
cache: &mut Cache,
wdcache: &mut WordDerivationsCache,
) -> anyhow::Result<Vec<(Query, Query, RoaringBitmap)>>
) -> Result<Vec<(Query, Query, RoaringBitmap)>>
{
// Extract the first two elements but gives the tail
// that is just after the first element.
@ -324,13 +325,13 @@ fn resolve_plane_sweep_candidates(
query_tree: &Operation,
allowed_candidates: &RoaringBitmap,
wdcache: &mut WordDerivationsCache,
) -> anyhow::Result<BTreeMap<u8, RoaringBitmap>>
) -> Result<BTreeMap<u8, RoaringBitmap>>
{
/// FIXME may be buggy with query like "new new york"
fn plane_sweep(
groups_positions: Vec<Vec<(Position, u8, Position)>>,
consecutive: bool,
) -> anyhow::Result<Vec<(Position, u8, Position)>>
) -> Result<Vec<(Position, u8, Position)>>
{
fn compute_groups_proximity(
groups: &[(usize, (Position, u8, Position))],
@ -451,7 +452,7 @@ fn resolve_plane_sweep_candidates(
rocache: &mut HashMap<&'a Operation, Vec<(Position, u8, Position)>>,
words_positions: &HashMap<String, RoaringBitmap>,
wdcache: &mut WordDerivationsCache,
) -> anyhow::Result<Vec<(Position, u8, Position)>>
) -> Result<Vec<(Position, u8, Position)>>
{
use Operation::{And, Phrase, Or};

13
milli/src/search/criteria/typo.rs
View file

@ -5,6 +5,7 @@ use roaring::RoaringBitmap;
use crate::search::query_tree::{maximum_typo, Operation, Query, QueryKind};
use crate::search::{word_derivations, WordDerivationsCache};
use crate::Result;
use super::{
Candidates,
Context,
@ -43,7 +44,7 @@ impl<'t> Typo<'t> {
impl<'t> Criterion for Typo<'t> {
#[logging_timer::time("Typo::{}")]
fn next(&mut self, params: &mut CriterionParameters) -> anyhow::Result<Option<CriterionResult>> {
fn next(&mut self, params: &mut CriterionParameters) -> Result<Option<CriterionResult>> {
use Candidates::{Allowed, Forbidden};
// remove excluded candidates when next is called, instead of doing it in the loop.
match self.state.as_mut() {
@ -163,14 +164,14 @@ fn alterate_query_tree(
mut query_tree: Operation,
number_typos: u8,
wdcache: &mut WordDerivationsCache,
) -> anyhow::Result<Operation>
) -> Result<Operation>
{
fn recurse(
words_fst: &fst::Set<Cow<[u8]>>,
operation: &mut Operation,
number_typos: u8,
wdcache: &mut WordDerivationsCache,
) -> anyhow::Result<()>
) -> Result<()>
{
use Operation::{And, Phrase, Or};
@ -218,7 +219,7 @@ fn resolve_candidates<'t>(
number_typos: u8,
cache: &mut HashMap<(Operation, u8), RoaringBitmap>,
wdcache: &mut WordDerivationsCache,
) -> anyhow::Result<RoaringBitmap>
) -> Result<RoaringBitmap>
{
fn resolve_operation<'t>(
ctx: &'t dyn Context,
@ -226,7 +227,7 @@ fn resolve_candidates<'t>(
number_typos: u8,
cache: &mut HashMap<(Operation, u8), RoaringBitmap>,
wdcache: &mut WordDerivationsCache,
) -> anyhow::Result<RoaringBitmap>
) -> Result<RoaringBitmap>
{
use Operation::{And, Phrase, Or, Query};
@ -277,7 +278,7 @@ fn resolve_candidates<'t>(
mana: u8,
cache: &mut HashMap<(Operation, u8), RoaringBitmap>,
wdcache: &mut WordDerivationsCache,
) -> anyhow::Result<RoaringBitmap>
) -> Result<RoaringBitmap>
{
match branches.split_first() {
Some((head, [])) => {

3
milli/src/search/criteria/words.rs
View file

@ -4,6 +4,7 @@ use log::debug;
use roaring::RoaringBitmap;
use crate::search::query_tree::Operation;
use crate::Result;
use super::{Context, Criterion, CriterionParameters, CriterionResult, resolve_query_tree};
pub struct Words<'t> {
@ -30,7 +31,7 @@ impl<'t> Words<'t> {
impl<'t> Criterion for Words<'t> {
#[logging_timer::time("Words::{}")]
fn next(&mut self, params: &mut CriterionParameters) -> anyhow::Result<Option<CriterionResult>> {
fn next(&mut self, params: &mut CriterionParameters) -> Result<Option<CriterionResult>> {
// remove excluded candidates when next is called, instead of doing it in the loop.
if let Some(candidates) = self.candidates.as_mut() {
*candidates -= params.excluded_candidates;

14
milli/src/search/distinct/facet_distinct.rs
View file

@ -5,7 +5,7 @@ use roaring::RoaringBitmap;
use super::{Distinct, DocIter};
use crate::heed_codec::facet::*;
use crate::{DocumentId, FieldId, Index};
use crate::{DocumentId, FieldId, Index, Result};
const FID_SIZE: usize = size_of::<FieldId>();
const DOCID_SIZE: usize = size_of::<DocumentId>();
@ -57,7 +57,7 @@ impl<'a> FacetDistinctIter<'a> {
.get(self.txn, &(self.distinct, 0, key, key))
}
fn distinct_string(&mut self, id: DocumentId) -> anyhow::Result<()> {
fn distinct_string(&mut self, id: DocumentId) -> Result<()> {
let iter = facet_string_values(id, self.distinct, self.index, self.txn)?;
for item in iter {
@ -73,7 +73,7 @@ impl<'a> FacetDistinctIter<'a> {
Ok(())
}
fn distinct_number(&mut self, id: DocumentId) -> anyhow::Result<()> {
fn distinct_number(&mut self, id: DocumentId) -> Result<()> {
let iter = facet_number_values(id, self.distinct, self.index, self.txn)?;
for item in iter {
@ -92,7 +92,7 @@ impl<'a> FacetDistinctIter<'a> {
/// Performs the next iteration of the facet distinct. This is a convenience method that is
/// called by the Iterator::next implementation that transposes the result. It makes error
/// handling easier.
fn next_inner(&mut self) -> anyhow::Result<Option<DocumentId>> {
fn next_inner(&mut self) -> Result<Option<DocumentId>> {
// The first step is to remove all the excluded documents from our candidates
self.candidates.difference_with(&self.excluded);
@ -129,7 +129,7 @@ fn facet_number_values<'a>(
distinct: FieldId,
index: &Index,
txn: &'a heed::RoTxn,
) -> anyhow::Result<heed::RoPrefix<'a, FieldDocIdFacetF64Codec, heed::types::Unit>> {
) -> Result<heed::RoPrefix<'a, FieldDocIdFacetF64Codec, heed::types::Unit>> {
let key = facet_values_prefix_key(distinct, id);
let iter = index
@ -146,7 +146,7 @@ fn facet_string_values<'a>(
distinct: FieldId,
index: &Index,
txn: &'a heed::RoTxn,
) -> anyhow::Result<heed::RoPrefix<'a, FieldDocIdFacetStringCodec, heed::types::Unit>> {
) -> Result<heed::RoPrefix<'a, FieldDocIdFacetStringCodec, heed::types::Unit>> {
let key = facet_values_prefix_key(distinct, id);
let iter = index
@ -159,7 +159,7 @@ fn facet_string_values<'a>(
}
impl Iterator for FacetDistinctIter<'_> {
type Item = anyhow::Result<DocumentId>;
type Item = Result<DocumentId>;
fn next(&mut self) -> Option<Self::Item> {
self.next_inner().transpose()

6
milli/src/search/distinct/mod.rs
View file

@ -3,13 +3,13 @@ mod noop_distinct;
use roaring::RoaringBitmap;
use crate::DocumentId;
use crate::{DocumentId, Result};
pub use facet_distinct::FacetDistinct;
pub use noop_distinct::NoopDistinct;
/// A trait implemented by document interators that are returned by calls to `Distinct::distinct`.
/// It provides a way to get back the ownership to the excluded set.
pub trait DocIter: Iterator<Item = anyhow::Result<DocumentId>> {
pub trait DocIter: Iterator<Item = Result<DocumentId>> {
/// Returns ownership on the internal exluded set.
fn into_excluded(self) -> RoaringBitmap;
}
@ -106,7 +106,7 @@ mod test {
/// Checks that all the candidates are distinct, and returns the candidates number.
pub(crate) fn validate_distinct_candidates(
candidates: impl Iterator<Item=anyhow::Result<DocumentId>>,
candidates: impl Iterator<Item = crate::Result<DocumentId>>,
distinct: FieldId,
index: &Index,
) -> usize {

4
milli/src/search/distinct/noop_distinct.rs
View file

@ -1,6 +1,6 @@
use roaring::{RoaringBitmap, bitmap::IntoIter};
use crate::DocumentId;
use crate::{DocumentId, Result};
use super::{DocIter, Distinct};
/// A distinct implementer that does not perform any distinct,
@ -13,7 +13,7 @@ pub struct NoopDistinctIter {
}
impl Iterator for NoopDistinctIter {
type Item = anyhow::Result<DocumentId>;
type Item = Result<DocumentId>;
fn next(&mut self) -> Option<Self::Item> {
self.candidates.next().map(Ok)

11
milli/src/search/facet/facet_distribution.rs
View file

@ -2,15 +2,15 @@ use std::collections::{HashSet, BTreeMap};
use std::ops::Bound::Unbounded;
use std::{cmp, fmt};
use anyhow::Context;
use heed::{Database, BytesDecode};
use heed::types::{ByteSlice, Unit};
use roaring::RoaringBitmap;
use crate::error::FieldIdMapMissingEntry;
use crate::facet::FacetType;
use crate::heed_codec::facet::FacetValueStringCodec;
use crate::search::facet::{FacetIter, FacetRange};
use crate::{Index, FieldId, DocumentId};
use crate::{Index, FieldId, DocumentId, Result};
/// The default number of values by facets that will
/// be fetched from the key-value store.
@ -195,14 +195,15 @@ impl<'a> FacetDistribution<'a> {
}
}
pub fn execute(&self) -> anyhow::Result<BTreeMap<String, BTreeMap<String, u64>>> {
pub fn execute(&self) -> Result<BTreeMap<String, BTreeMap<String, u64>>> {
let fields_ids_map = self.index.fields_ids_map(self.rtxn)?;
let filterable_fields = self.index.filterable_fields(self.rtxn)?;
let mut distribution = BTreeMap::new();
for name in filterable_fields {
let fid = fields_ids_map.id(&name).with_context(|| {
format!("missing field name {:?} from the fields id map", name)
let fid = fields_ids_map.id(&name).ok_or_else(|| FieldIdMapMissingEntry::FieldName {
field_name: name.clone(),
from_db_name: "filterable-fields",
})?;
let values = self.facet_values(fid)?;
distribution.insert(name, values);

70
milli/src/search/facet/filter_condition.rs
View file

@ -1,6 +1,7 @@
use std::collections::HashSet;
use std::fmt::Debug;
use std::ops::Bound::{self, Included, Excluded};
use std::result::Result as StdResult;
use std::str::FromStr;
use either::Either;
@ -11,8 +12,9 @@ use pest::iterators::{Pair, Pairs};
use pest::Parser;
use roaring::RoaringBitmap;
use crate::error::UserError;
use crate::heed_codec::facet::{FacetValueStringCodec, FacetLevelValueF64Codec};
use crate::{Index, FieldId, FieldsIdsMap, CboRoaringBitmapCodec};
use crate::{Index, FieldId, FieldsIdsMap, CboRoaringBitmapCodec, Result};
use super::FacetRange;
use super::parser::Rule;
@ -60,7 +62,7 @@ impl FilterCondition {
rtxn: &heed::RoTxn,
index: &Index,
array: I,
) -> anyhow::Result<Option<FilterCondition>>
) -> Result<Option<FilterCondition>>
where I: IntoIterator<Item=Either<J, B>>,
J: IntoIterator<Item=A>,
A: AsRef<str>,
@ -104,11 +106,11 @@ impl FilterCondition {
rtxn: &heed::RoTxn,
index: &Index,
expression: &str,
) -> anyhow::Result<FilterCondition>
) -> Result<FilterCondition>
{
let fields_ids_map = index.fields_ids_map(rtxn)?;
let filterable_fields = index.filterable_fields_ids(rtxn)?;
let lexed = FilterParser::parse(Rule::prgm, expression)?;
let lexed = FilterParser::parse(Rule::prgm, expression).map_err(UserError::FilterParsing)?;
FilterCondition::from_pairs(&fields_ids_map, &filterable_fields, lexed)
}
@ -116,7 +118,7 @@ impl FilterCondition {
fim: &FieldsIdsMap,
ff: &HashSet<FieldId>,
expression: Pairs<Rule>,
) -> anyhow::Result<Self>
) -> Result<Self>
{
PREC_CLIMBER.climb(
expression,
@ -133,7 +135,7 @@ impl FilterCondition {
Rule::term => Self::from_pairs(fim, ff, pair.into_inner()),
_ => unreachable!(),
},
|lhs: anyhow::Result<Self>, op: Pair<Rule>, rhs: anyhow::Result<Self>| {
|lhs: Result<Self>, op: Pair<Rule>, rhs: Result<Self>| {
match op.as_rule() {
Rule::or => Ok(Or(Box::new(lhs?), Box::new(rhs?))),
Rule::and => Ok(And(Box::new(lhs?), Box::new(rhs?))),
@ -158,16 +160,17 @@ impl FilterCondition {
fields_ids_map: &FieldsIdsMap,
filterable_fields: &HashSet<FieldId>,
item: Pair<Rule>,
) -> anyhow::Result<FilterCondition>
) -> Result<FilterCondition>
{
let mut items = item.into_inner();
let fid = field_id(fields_ids_map, filterable_fields, &mut items)?;
let fid = field_id(fields_ids_map, filterable_fields, &mut items)
.map_err(UserError::FilterParsing)?;
let (lresult, _) = pest_parse(items.next().unwrap());
let (rresult, _) = pest_parse(items.next().unwrap());
let lvalue = lresult?;
let rvalue = rresult?;
let lvalue = lresult.map_err(UserError::FilterParsing)?;
let rvalue = rresult.map_err(UserError::FilterParsing)?;
Ok(Operator(fid, Between(lvalue, rvalue)))
}
@ -176,10 +179,11 @@ impl FilterCondition {
fields_ids_map: &FieldsIdsMap,
filterable_fields: &HashSet<FieldId>,
item: Pair<Rule>,
) -> anyhow::Result<FilterCondition>
) -> Result<FilterCondition>
{
let mut items = item.into_inner();
let fid = field_id(fields_ids_map, filterable_fields, &mut items)?;
let fid = field_id(fields_ids_map, filterable_fields, &mut items)
.map_err(UserError::FilterParsing)?;
let value = items.next().unwrap();
let (result, svalue) = pest_parse(value);
@ -192,60 +196,68 @@ impl FilterCondition {
fields_ids_map: &FieldsIdsMap,
filterable_fields: &HashSet<FieldId>,
item: Pair<Rule>,
) -> anyhow::Result<FilterCondition>
) -> Result<FilterCondition>
{
let mut items = item.into_inner();
let fid = field_id(fields_ids_map, filterable_fields, &mut items)?;
let fid = field_id(fields_ids_map, filterable_fields, &mut items)
.map_err(UserError::FilterParsing)?;
let value = items.next().unwrap();
let (result, _svalue) = pest_parse(value);
let value = result.map_err(UserError::FilterParsing)?;
Ok(Operator(fid, GreaterThan(result?)))
Ok(Operator(fid, GreaterThan(value)))
}
fn greater_than_or_equal(
fields_ids_map: &FieldsIdsMap,
filterable_fields: &HashSet<FieldId>,
item: Pair<Rule>,
) -> anyhow::Result<FilterCondition>
) -> Result<FilterCondition>
{
let mut items = item.into_inner();
let fid = field_id(fields_ids_map, filterable_fields, &mut items)?;
let fid = field_id(fields_ids_map, filterable_fields, &mut items)
.map_err(UserError::FilterParsing)?;
let value = items.next().unwrap();
let (result, _svalue) = pest_parse(value);
let value = result.map_err(UserError::FilterParsing)?;
Ok(Operator(fid, GreaterThanOrEqual(result?)))
Ok(Operator(fid, GreaterThanOrEqual(value)))
}
fn lower_than(
fields_ids_map: &FieldsIdsMap,
filterable_fields: &HashSet<FieldId>,
item: Pair<Rule>,
) -> anyhow::Result<FilterCondition>
) -> Result<FilterCondition>
{
let mut items = item.into_inner();
let fid = field_id(fields_ids_map, filterable_fields, &mut items)?;
let fid = field_id(fields_ids_map, filterable_fields, &mut items)
.map_err(UserError::FilterParsing)?;
let value = items.next().unwrap();
let (result, _svalue) = pest_parse(value);
let value = result.map_err(UserError::FilterParsing)?;
Ok(Operator(fid, LowerThan(result?)))
Ok(Operator(fid, LowerThan(value)))
}
fn lower_than_or_equal(
fields_ids_map: &FieldsIdsMap,
filterable_fields: &HashSet<FieldId>,
item: Pair<Rule>,
) -> anyhow::Result<FilterCondition>
) -> Result<FilterCondition>
{
let mut items = item.into_inner();
let fid = field_id(fields_ids_map, filterable_fields, &mut items)?;
let fid = field_id(fields_ids_map, filterable_fields, &mut items)
.map_err(UserError::FilterParsing)?;
let value = items.next().unwrap();
let (result, _svalue) = pest_parse(value);
let value = result.map_err(UserError::FilterParsing)?;
Ok(Operator(fid, LowerThanOrEqual(result?)))
Ok(Operator(fid, LowerThanOrEqual(value)))
}
}
@ -260,7 +272,7 @@ impl FilterCondition {
left: Bound<f64>,
right: Bound<f64>,
output: &mut RoaringBitmap,
) -> anyhow::Result<()>
) -> Result<()>
{
match (left, right) {
// If the request is an exact value we must go directly to the deepest level.
@ -332,7 +344,7 @@ impl FilterCondition {
strings_db: heed::Database<FacetValueStringCodec, CboRoaringBitmapCodec>,
field_id: FieldId,
operator: &Operator,
) -> anyhow::Result<RoaringBitmap>
) -> Result<RoaringBitmap>
{
// Make sure we always bound the ranges with the field id and the level,
// as the facets values are all in the same database and prefixed by the
@ -390,7 +402,7 @@ impl FilterCondition {
&self,
rtxn: &heed::RoTxn,
index: &Index,
) -> anyhow::Result<RoaringBitmap>
) -> Result<RoaringBitmap>
{
let numbers_db = index.facet_id_f64_docids;
let strings_db = index.facet_id_string_docids;
@ -422,7 +434,7 @@ fn field_id(
fields_ids_map: &FieldsIdsMap,
filterable_fields: &HashSet<FieldId>,
items: &mut Pairs<Rule>,
) -> Result<FieldId, PestError<Rule>>
) -> StdResult<FieldId, PestError<Rule>>
{
// lexing ensures that we at least have a key
let key = items.next().unwrap();
@ -463,7 +475,7 @@ fn field_id(
/// the original string that we tried to parse.
///
/// Returns the parsing error associated with the span if the conversion fails.
fn pest_parse<T>(pair: Pair<Rule>) -> (Result<T, pest::error::Error<Rule>>, String)
fn pest_parse<T>(pair: Pair<Rule>) -> (StdResult<T, pest::error::Error<Rule>>, String)
where T: FromStr,
T::Err: ToString,
{

3
milli/src/search/facet/mod.rs
View file

@ -9,8 +9,9 @@ use crate::heed_codec::CboRoaringBitmapCodec;
use crate::heed_codec::facet::FacetLevelValueF64Codec;
use crate::{Index, FieldId};
pub use self::filter_condition::{FilterCondition, Operator};
pub use self::facet_distribution::FacetDistribution;
pub use self::filter_condition::{FilterCondition, Operator};
pub(crate) use self::parser::Rule as ParserRule;
mod filter_condition;
mod facet_distribution;

10
milli/src/search/mod.rs
View file

@ -2,6 +2,7 @@ use std::borrow::Cow;
use std::collections::hash_map::{Entry, HashMap};
use std::fmt;
use std::mem::take;
use std::result::Result as StdResult;
use std::str::Utf8Error;
use std::time::Instant;
@ -14,10 +15,11 @@ use roaring::bitmap::RoaringBitmap;
use distinct::{Distinct, DocIter, FacetDistinct, NoopDistinct};
use crate::search::criteria::r#final::{Final, FinalResult};
use crate::{Index, DocumentId};
use crate::{Index, DocumentId, Result};
pub use self::facet::{FilterCondition, FacetDistribution, FacetIter, Operator};
pub use self::matching_words::MatchingWords;
pub(crate) use self::facet::ParserRule;
use self::query_tree::QueryTreeBuilder;
// Building these factories is not free.
@ -93,7 +95,7 @@ impl<'a> Search<'a> {
self
}
pub fn execute(&self) -> anyhow::Result<SearchResult> {
pub fn execute(&self) -> Result<SearchResult> {
// We create the query tree by spliting the query into tokens.
let before = Instant::now();
let (query_tree, primitive_query) = match self.query.as_ref() {
@ -152,7 +154,7 @@ impl<'a> Search<'a> {
mut distinct: D,
matching_words: MatchingWords,
mut criteria: Final,
) -> anyhow::Result<SearchResult>
) -> Result<SearchResult>
{
let mut offset = self.offset;
let mut initial_candidates = RoaringBitmap::new();
@ -225,7 +227,7 @@ pub fn word_derivations<'c>(
max_typo: u8,
fst: &fst::Set<Cow<[u8]>>,
cache: &'c mut WordDerivationsCache,
) -> Result<&'c [(String, u8)], Utf8Error> {
) -> StdResult<&'c [(String, u8)], Utf8Error> {
match cache.entry((word.to_string(), is_prefix, max_typo)) {
Entry::Occupied(entry) => Ok(entry.into_mut()),
Entry::Vacant(entry) => {

14
milli/src/search/query_tree.rs
View file

@ -7,7 +7,7 @@ use meilisearch_tokenizer::TokenKind;
use roaring::RoaringBitmap;
use slice_group_by::GroupBy;
use crate::Index;
use crate::{Index, Result};
type IsOptionalWord = bool;
type IsPrefix = bool;
@ -219,7 +219,7 @@ impl<'a> QueryTreeBuilder<'a> {
/// - if `authorize_typos` is set to `false` the query tree will be generated
/// forcing all query words to match documents without any typo
/// (the criterion `typo` will be ignored)
pub fn build(&self, query: TokenStream) -> anyhow::Result<Option<(Operation, PrimitiveQuery)>> {
pub fn build(&self, query: TokenStream) -> Result<Option<(Operation, PrimitiveQuery)>> {
let stop_words = self.index.stop_words(self.rtxn)?;
let primitive_query = create_primitive_query(query, stop_words, self.words_limit);
if !primitive_query.is_empty() {
@ -291,14 +291,14 @@ fn create_query_tree(
optional_words: bool,
authorize_typos: bool,
query: &[PrimitiveQueryPart],
) -> anyhow::Result<Operation>
) -> Result<Operation>
{
/// Matches on the `PrimitiveQueryPart` and create an operation from it.
fn resolve_primitive_part(
ctx: &impl Context,
authorize_typos: bool,
part: PrimitiveQueryPart,
) -> anyhow::Result<Operation>
) -> Result<Operation>
{
match part {
// 1. try to split word in 2
@ -325,7 +325,7 @@ fn create_query_tree(
ctx: &impl Context,
authorize_typos: bool,
query: &[PrimitiveQueryPart],
) -> anyhow::Result<Operation>
) -> Result<Operation>
{
const MAX_NGRAM: usize = 3;
let mut op_children = Vec::new();
@ -379,7 +379,7 @@ fn create_query_tree(
ctx: &impl Context,
authorize_typos: bool,
query: PrimitiveQuery,
) -> anyhow::Result<Operation>
) -> Result<Operation>
{
let number_phrases = query.iter().filter(|p| p.is_phrase()).count();
let mut operation_children = Vec::new();
@ -532,7 +532,7 @@ mod test {
authorize_typos: bool,
words_limit: Option<usize>,
query: TokenStream,
) -> anyhow::Result<Option<(Operation, PrimitiveQuery)>>
) -> Result<Option<(Operation, PrimitiveQuery)>>
{
let primitive_query = create_primitive_query(query, None, words_limit);
if !primitive_query.is_empty() {