format the whole project

This commit is contained in:
Tamo
2021-06-16 18:33:33 +02:00
parent ba30cef987
commit 9716fb3b36
68 changed files with 3327 additions and 2336 deletions

View File

@ -1,29 +1,28 @@
use std::collections::HashMap;
use std::borrow::Cow;
use std::collections::HashMap;
use roaring::RoaringBitmap;
use crate::{FieldId, TreeLevel, search::{word_derivations, WordDerivationsCache}};
use crate::{Index, DocumentId, Result};
use super::query_tree::{Operation, PrimitiveQueryPart, Query, QueryKind};
use self::asc_desc::AscDesc;
use self::attribute::Attribute;
use self::exactness::Exactness;
use self::r#final::Final;
use self::initial::Initial;
use self::proximity::Proximity;
use self::r#final::Final;
use self::typo::Typo;
use self::words::Words;
use super::query_tree::{Operation, PrimitiveQueryPart, Query, QueryKind};
use crate::search::{word_derivations, WordDerivationsCache};
use crate::{DocumentId, FieldId, Index, Result, TreeLevel};
mod asc_desc;
mod attribute;
mod exactness;
pub mod r#final;
mod initial;
mod proximity;
mod typo;
mod words;
pub mod r#final;
pub trait Criterion {
fn next(&mut self, params: &mut CriterionParameters) -> Result<Option<CriterionResult>>;
@ -55,7 +54,7 @@ pub struct CriterionParameters<'a> {
#[derive(Debug)]
enum Candidates {
Allowed(RoaringBitmap),
Forbidden(RoaringBitmap)
Forbidden(RoaringBitmap),
}
impl Default for Candidates {
@ -68,17 +67,55 @@ pub trait Context<'c> {
fn documents_ids(&self) -> heed::Result<RoaringBitmap>;
fn word_docids(&self, word: &str) -> heed::Result<Option<RoaringBitmap>>;
fn word_prefix_docids(&self, word: &str) -> heed::Result<Option<RoaringBitmap>>;
fn word_pair_proximity_docids(&self, left: &str, right: &str, proximity: u8) -> heed::Result<Option<RoaringBitmap>>;
fn word_prefix_pair_proximity_docids(&self, left: &str, right: &str, proximity: u8) -> heed::Result<Option<RoaringBitmap>>;
fn word_pair_proximity_docids(
&self,
left: &str,
right: &str,
proximity: u8,
) -> heed::Result<Option<RoaringBitmap>>;
fn word_prefix_pair_proximity_docids(
&self,
left: &str,
right: &str,
proximity: u8,
) -> heed::Result<Option<RoaringBitmap>>;
fn words_fst<'t>(&self) -> &'t fst::Set<Cow<[u8]>>;
fn in_prefix_cache(&self, word: &str) -> bool;
fn docid_words_positions(&self, docid: DocumentId) -> heed::Result<HashMap<String, RoaringBitmap>>;
fn word_position_iterator(&self, word: &str, level: TreeLevel, in_prefix_cache: bool, left: Option<u32>, right: Option<u32>) -> heed::Result<Box<dyn Iterator<Item =heed::Result<((&'c str, TreeLevel, u32, u32), RoaringBitmap)>> + 'c>>;
fn word_position_last_level(&self, word: &str, in_prefix_cache: bool) -> heed::Result<Option<TreeLevel>>;
fn docid_words_positions(
&self,
docid: DocumentId,
) -> heed::Result<HashMap<String, RoaringBitmap>>;
fn word_position_iterator(
&self,
word: &str,
level: TreeLevel,
in_prefix_cache: bool,
left: Option<u32>,
right: Option<u32>,
) -> heed::Result<
Box<
dyn Iterator<Item = heed::Result<((&'c str, TreeLevel, u32, u32), RoaringBitmap)>> + 'c,
>,
>;
fn word_position_last_level(
&self,
word: &str,
in_prefix_cache: bool,
) -> heed::Result<Option<TreeLevel>>;
fn synonyms(&self, word: &str) -> heed::Result<Option<Vec<Vec<String>>>>;
fn searchable_fields_ids(&self) -> 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) -> heed::Result<Option<RoaringBitmap>>;
fn searchable_fields_ids(&self) -> 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,
) -> heed::Result<Option<RoaringBitmap>>;
}
pub struct CriteriaBuilder<'t> {
@ -101,12 +138,22 @@ impl<'c> Context<'c> for CriteriaBuilder<'c> {
self.index.word_prefix_docids.get(self.rtxn, &word)
}
fn word_pair_proximity_docids(&self, left: &str, right: &str, proximity: u8) -> heed::Result<Option<RoaringBitmap>> {
fn word_pair_proximity_docids(
&self,
left: &str,
right: &str,
proximity: u8,
) -> heed::Result<Option<RoaringBitmap>> {
let key = (left, right, proximity);
self.index.word_pair_proximity_docids.get(self.rtxn, &key)
}
fn word_prefix_pair_proximity_docids(&self, left: &str, right: &str, proximity: u8) -> heed::Result<Option<RoaringBitmap>> {
fn word_prefix_pair_proximity_docids(
&self,
left: &str,
right: &str,
proximity: u8,
) -> heed::Result<Option<RoaringBitmap>> {
let key = (left, right, proximity);
self.index.word_prefix_pair_proximity_docids.get(self.rtxn, &key)
}
@ -119,7 +166,10 @@ impl<'c> Context<'c> for CriteriaBuilder<'c> {
self.words_prefixes_fst.contains(word)
}
fn docid_words_positions(&self, docid: DocumentId) -> heed::Result<HashMap<String, RoaringBitmap>> {
fn docid_words_positions(
&self,
docid: DocumentId,
) -> heed::Result<HashMap<String, RoaringBitmap>> {
let mut words_positions = HashMap::new();
for result in self.index.docid_word_positions.prefix_iter(self.rtxn, &(docid, ""))? {
let ((_, word), positions) = result?;
@ -134,9 +184,12 @@ impl<'c> Context<'c> for CriteriaBuilder<'c> {
level: TreeLevel,
in_prefix_cache: bool,
left: Option<u32>,
right: Option<u32>
) -> heed::Result<Box<dyn Iterator<Item = heed::Result<((&'c str, TreeLevel, u32, u32), RoaringBitmap)>> + 'c>>
{
right: Option<u32>,
) -> heed::Result<
Box<
dyn Iterator<Item = heed::Result<((&'c str, TreeLevel, u32, u32), RoaringBitmap)>> + 'c,
>,
> {
let range = {
let left = left.unwrap_or(u32::min_value());
let right = right.unwrap_or(u32::max_value());
@ -152,7 +205,11 @@ impl<'c> Context<'c> for CriteriaBuilder<'c> {
Ok(Box::new(db.range(self.rtxn, &range)?))
}
fn word_position_last_level(&self, word: &str, in_prefix_cache: bool) -> heed::Result<Option<TreeLevel>> {
fn word_position_last_level(
&self,
word: &str,
in_prefix_cache: bool,
) -> heed::Result<Option<TreeLevel>> {
let range = {
let left = (word, TreeLevel::min_value(), u32::min_value(), u32::min_value());
let right = (word, TreeLevel::max_value(), u32::max_value(), u32::max_value());
@ -164,7 +221,9 @@ impl<'c> Context<'c> for CriteriaBuilder<'c> {
};
let last_level = db
.remap_data_type::<heed::types::DecodeIgnore>()
.range(self.rtxn, &range)?.last().transpose()?
.range(self.rtxn, &range)?
.last()
.transpose()?
.map(|((_, level, _, _), _)| level);
Ok(last_level)
@ -181,12 +240,22 @@ impl<'c> Context<'c> for CriteriaBuilder<'c> {
}
}
fn field_id_word_count_docids(&self, field_id: FieldId, word_count: u8) -> heed::Result<Option<RoaringBitmap>> {
fn field_id_word_count_docids(
&self,
field_id: FieldId,
word_count: u8,
) -> heed::Result<Option<RoaringBitmap>> {
let key = (field_id, word_count);
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) -> heed::Result<Option<RoaringBitmap>> {
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)
}
@ -204,13 +273,13 @@ impl<'t> CriteriaBuilder<'t> {
query_tree: Option<Operation>,
primitive_query: Option<Vec<PrimitiveQueryPart>>,
filtered_candidates: Option<RoaringBitmap>,
) -> Result<Final<'t>>
{
) -> Result<Final<'t>> {
use crate::criterion::Criterion as Name;
let primitive_query = primitive_query.unwrap_or_default();
let mut criterion = Box::new(Initial::new(query_tree, filtered_candidates)) as Box<dyn Criterion>;
let mut criterion =
Box::new(Initial::new(query_tree, filtered_candidates)) as Box<dyn Criterion>;
for name in self.index.criteria(&self.rtxn)? {
criterion = match name {
Name::Typo => Box::new(Typo::new(self, criterion)),
@ -218,8 +287,12 @@ impl<'t> CriteriaBuilder<'t> {
Name::Proximity => Box::new(Proximity::new(self, criterion)),
Name::Attribute => Box::new(Attribute::new(self, criterion)),
Name::Exactness => Box::new(Exactness::new(self, criterion, &primitive_query)?),
Name::Asc(field) => Box::new(AscDesc::asc(&self.index, &self.rtxn, criterion, field)?),
Name::Desc(field) => Box::new(AscDesc::desc(&self.index, &self.rtxn, criterion, field)?),
Name::Asc(field) => {
Box::new(AscDesc::asc(&self.index, &self.rtxn, criterion, field)?)
}
Name::Desc(field) => {
Box::new(AscDesc::desc(&self.index, &self.rtxn, criterion, field)?)
}
};
}
@ -231,21 +304,20 @@ pub fn resolve_query_tree<'t>(
ctx: &'t dyn Context,
query_tree: &Operation,
wdcache: &mut WordDerivationsCache,
) -> Result<RoaringBitmap>
{
) -> Result<RoaringBitmap> {
fn resolve_operation<'t>(
ctx: &'t dyn Context,
query_tree: &Operation,
wdcache: &mut WordDerivationsCache,
) -> Result<RoaringBitmap>
{
use Operation::{And, Phrase, Or, Query};
) -> Result<RoaringBitmap> {
use Operation::{And, Or, Phrase, Query};
match query_tree {
And(ops) => {
let mut ops = ops.iter().map(|op| {
resolve_operation(ctx, op, wdcache)
}).collect::<Result<Vec<_>>>()?;
let mut ops = ops
.iter()
.map(|op| resolve_operation(ctx, op, wdcache))
.collect::<Result<Vec<_>>>()?;
ops.sort_unstable_by_key(|cds| cds.len());
@ -260,7 +332,7 @@ pub fn resolve_query_tree<'t>(
}
}
Ok(candidates)
},
}
Phrase(words) => {
let mut candidates = RoaringBitmap::new();
let mut first_loop = true;
@ -276,12 +348,12 @@ pub fn resolve_query_tree<'t>(
} else {
candidates &= pair_docids;
}
},
None => return Ok(RoaringBitmap::new())
}
None => return Ok(RoaringBitmap::new()),
}
}
Ok(candidates)
},
}
Or(_, ops) => {
let mut candidates = RoaringBitmap::new();
for op in ops {
@ -289,7 +361,7 @@ pub fn resolve_query_tree<'t>(
candidates.union_with(&docids);
}
Ok(candidates)
},
}
Query(q) => Ok(query_docids(ctx, q, wdcache)?),
}
}
@ -297,18 +369,18 @@ pub fn resolve_query_tree<'t>(
resolve_operation(ctx, query_tree, wdcache)
}
fn all_word_pair_proximity_docids<T: AsRef<str>, U: AsRef<str>>(
ctx: &dyn Context,
left_words: &[(T, u8)],
right_words: &[(U, u8)],
proximity: u8
) -> Result<RoaringBitmap>
{
proximity: u8,
) -> Result<RoaringBitmap> {
let mut docids = RoaringBitmap::new();
for (left, _l_typo) in left_words {
for (right, _r_typo) in right_words {
let current_docids = ctx.word_pair_proximity_docids(left.as_ref(), right.as_ref(), proximity)?.unwrap_or_default();
let current_docids = ctx
.word_pair_proximity_docids(left.as_ref(), right.as_ref(), proximity)?
.unwrap_or_default();
docids.union_with(&current_docids);
}
}
@ -319,8 +391,7 @@ fn query_docids(
ctx: &dyn Context,
query: &Query,
wdcache: &mut WordDerivationsCache,
) -> Result<RoaringBitmap>
{
) -> Result<RoaringBitmap> {
match &query.kind {
QueryKind::Exact { word, .. } => {
if query.prefix && ctx.in_prefix_cache(&word) {
@ -336,7 +407,7 @@ fn query_docids(
} else {
Ok(ctx.word_docids(&word)?.unwrap_or_default())
}
},
}
QueryKind::Tolerant { typo, word } => {
let words = word_derivations(&word, query.prefix, *typo, ctx.words_fst(), wdcache)?;
let mut docids = RoaringBitmap::new();
@ -345,7 +416,7 @@ fn query_docids(
docids.union_with(&current_docids);
}
Ok(docids)
},
}
}
}
@ -355,8 +426,7 @@ fn query_pair_proximity_docids(
right: &Query,
proximity: u8,
wdcache: &mut WordDerivationsCache,
) -> Result<RoaringBitmap>
{
) -> Result<RoaringBitmap> {
if proximity >= 8 {
let mut candidates = query_docids(ctx, left, wdcache)?;
let right_candidates = query_docids(ctx, right, wdcache)?;
@ -368,20 +438,31 @@ fn query_pair_proximity_docids(
match (&left.kind, &right.kind) {
(QueryKind::Exact { word: left, .. }, QueryKind::Exact { word: right, .. }) => {
if prefix && ctx.in_prefix_cache(&right) {
Ok(ctx.word_prefix_pair_proximity_docids(left.as_str(), right.as_str(), proximity)?.unwrap_or_default())
Ok(ctx
.word_prefix_pair_proximity_docids(left.as_str(), right.as_str(), proximity)?
.unwrap_or_default())
} else if prefix {
let r_words = word_derivations(&right, true, 0, ctx.words_fst(), wdcache)?;
all_word_pair_proximity_docids(ctx, &[(left, 0)], &r_words, proximity)
} else {
Ok(ctx.word_pair_proximity_docids(left.as_str(), right.as_str(), proximity)?.unwrap_or_default())
Ok(ctx
.word_pair_proximity_docids(left.as_str(), right.as_str(), proximity)?
.unwrap_or_default())
}
},
}
(QueryKind::Tolerant { typo, word: left }, QueryKind::Exact { word: right, .. }) => {
let l_words = word_derivations(&left, false, *typo, ctx.words_fst(), wdcache)?.to_owned();
let l_words =
word_derivations(&left, false, *typo, ctx.words_fst(), wdcache)?.to_owned();
if prefix && ctx.in_prefix_cache(&right) {
let mut docids = RoaringBitmap::new();
for (left, _) in l_words {
let current_docids = ctx.word_prefix_pair_proximity_docids(left.as_ref(), right.as_ref(), proximity)?.unwrap_or_default();
let current_docids = ctx
.word_prefix_pair_proximity_docids(
left.as_ref(),
right.as_ref(),
proximity,
)?
.unwrap_or_default();
docids.union_with(&current_docids);
}
Ok(docids)
@ -391,28 +472,36 @@ fn query_pair_proximity_docids(
} else {
all_word_pair_proximity_docids(ctx, &l_words, &[(right, 0)], proximity)
}
},
}
(QueryKind::Exact { word: left, .. }, QueryKind::Tolerant { typo, word: right }) => {
let r_words = word_derivations(&right, prefix, *typo, ctx.words_fst(), wdcache)?;
all_word_pair_proximity_docids(ctx, &[(left, 0)], &r_words, proximity)
},
(QueryKind::Tolerant { typo: l_typo, word: left }, QueryKind::Tolerant { typo: r_typo, word: right }) => {
let l_words = word_derivations(&left, false, *l_typo, ctx.words_fst(), wdcache)?.to_owned();
}
(
QueryKind::Tolerant { typo: l_typo, word: left },
QueryKind::Tolerant { typo: r_typo, word: right },
) => {
let l_words =
word_derivations(&left, false, *l_typo, ctx.words_fst(), wdcache)?.to_owned();
let r_words = word_derivations(&right, prefix, *r_typo, ctx.words_fst(), wdcache)?;
all_word_pair_proximity_docids(ctx, &l_words, &r_words, proximity)
},
}
}
}
#[cfg(test)]
pub mod test {
use maplit::hashmap;
use rand::{Rng, SeedableRng, rngs::StdRng};
use super::*;
use std::collections::HashMap;
fn s(s: &str) -> String { s.to_string() }
use maplit::hashmap;
use rand::rngs::StdRng;
use rand::{Rng, SeedableRng};
use super::*;
fn s(s: &str) -> String {
s.to_string()
}
pub struct TestContext<'t> {
words_fst: fst::Set<Cow<'t, [u8]>>,
word_docids: HashMap<String, RoaringBitmap>,
@ -435,12 +524,22 @@ pub mod test {
Ok(self.word_prefix_docids.get(&word.to_string()).cloned())
}
fn word_pair_proximity_docids(&self, left: &str, right: &str, proximity: u8) -> heed::Result<Option<RoaringBitmap>> {
fn word_pair_proximity_docids(
&self,
left: &str,
right: &str,
proximity: u8,
) -> heed::Result<Option<RoaringBitmap>> {
let key = (left.to_string(), right.to_string(), proximity.into());
Ok(self.word_pair_proximity_docids.get(&key).cloned())
}
fn word_prefix_pair_proximity_docids(&self, left: &str, right: &str, proximity: u8) -> heed::Result<Option<RoaringBitmap>> {
fn word_prefix_pair_proximity_docids(
&self,
left: &str,
right: &str,
proximity: u8,
) -> heed::Result<Option<RoaringBitmap>> {
let key = (left.to_string(), right.to_string(), proximity.into());
Ok(self.word_prefix_pair_proximity_docids.get(&key).cloned())
}
@ -453,24 +552,44 @@ pub mod test {
self.word_prefix_docids.contains_key(&word.to_string())
}
fn docid_words_positions(&self, docid: DocumentId) -> heed::Result<HashMap<String, RoaringBitmap>> {
fn docid_words_positions(
&self,
docid: DocumentId,
) -> heed::Result<HashMap<String, RoaringBitmap>> {
if let Some(docid_words) = self.docid_words.get(&docid) {
Ok(docid_words
.iter()
.enumerate()
.map(|(i,w)| (w.clone(), RoaringBitmap::from_sorted_iter(std::iter::once(i as u32))))
.collect()
)
.map(|(i, w)| {
(w.clone(), RoaringBitmap::from_sorted_iter(std::iter::once(i as u32)))
})
.collect())
} else {
Ok(HashMap::new())
}
}
fn word_position_iterator(&self, _word: &str, _level: TreeLevel, _in_prefix_cache: bool, _left: Option<u32>, _right: Option<u32>) -> heed::Result<Box<dyn Iterator<Item =heed::Result<((&'c str, TreeLevel, u32, u32), RoaringBitmap)>> + 'c>> {
fn word_position_iterator(
&self,
_word: &str,
_level: TreeLevel,
_in_prefix_cache: bool,
_left: Option<u32>,
_right: Option<u32>,
) -> heed::Result<
Box<
dyn Iterator<Item = heed::Result<((&'c str, TreeLevel, u32, u32), RoaringBitmap)>>
+ 'c,
>,
> {
todo!()
}
fn word_position_last_level(&self, _word: &str, _in_prefix_cache: bool) -> heed::Result<Option<TreeLevel>> {
fn word_position_last_level(
&self,
_word: &str,
_in_prefix_cache: bool,
) -> heed::Result<Option<TreeLevel>> {
todo!()
}
@ -478,15 +597,25 @@ pub mod test {
todo!()
}
fn searchable_fields_ids(&self) -> Result<Vec<FieldId>> {
fn searchable_fields_ids(&self) -> Result<Vec<FieldId>> {
todo!()
}
fn word_level_position_docids(&self, _word: &str, _level: TreeLevel, _left: u32, _right: u32) -> heed::Result<Option<RoaringBitmap>> {
fn word_level_position_docids(
&self,
_word: &str,
_level: TreeLevel,
_left: u32,
_right: u32,
) -> heed::Result<Option<RoaringBitmap>> {
todo!()
}
fn field_id_word_count_docids(&self, _field_id: FieldId, _word_count: u8) -> heed::Result<Option<RoaringBitmap>> {
fn field_id_word_count_docids(
&self,
_field_id: FieldId,
_word_count: u8,
) -> heed::Result<Option<RoaringBitmap>> {
todo!()
}
}
@ -506,7 +635,7 @@ pub mod test {
RoaringBitmap::from_sorted_iter(values.into_iter())
}
let word_docids = hashmap!{
let word_docids = hashmap! {
s("hello") => random_postings(rng, 1500),
s("hi") => random_postings(rng, 4000),
s("word") => random_postings(rng, 2500),
@ -530,7 +659,7 @@ pub mod test {
}
}
let word_prefix_docids = hashmap!{
let word_prefix_docids = hashmap! {
s("h") => &word_docids[&s("hello")] | &word_docids[&s("hi")],
s("wor") => &word_docids[&s("word")] | &word_docids[&s("world")],
s("20") => &word_docids[&s("2020")] | &word_docids[&s("2021")],
@ -540,7 +669,9 @@ pub mod test {
let mut word_prefix_pair_proximity_docids = HashMap::new();
for (lword, lcandidates) in &word_docids {
for (rword, rcandidates) in &word_docids {
if lword == rword { continue }
if lword == rword {
continue;
}
let candidates = lcandidates & rcandidates;
for candidate in candidates {
if let Some(docid_words) = docid_words.get(&candidate) {
@ -551,24 +682,31 @@ pub mod test {
} else {
(s(lword), s(rword), (lposition - rposition + 1) as i32)
};
let docids = word_pair_proximity_docids.entry(key).or_insert(RoaringBitmap::new());
let docids = word_pair_proximity_docids
.entry(key)
.or_insert(RoaringBitmap::new());
docids.push(candidate);
}
}
}
for (pword, pcandidates) in &word_prefix_docids {
if lword.starts_with(pword) { continue }
if lword.starts_with(pword) {
continue;
}
let candidates = lcandidates & pcandidates;
for candidate in candidates {
if let Some(docid_words) = docid_words.get(&candidate) {
let lposition = docid_words.iter().position(|w| w == lword).unwrap();
let rposition = docid_words.iter().position(|w| w.starts_with(pword)).unwrap();
let rposition =
docid_words.iter().position(|w| w.starts_with(pword)).unwrap();
let key = if lposition < rposition {
(s(lword), s(pword), (rposition - lposition) as i32)
} else {
(s(lword), s(pword), (lposition - rposition + 1) as i32)
};
let docids = word_prefix_pair_proximity_docids.entry(key).or_insert(RoaringBitmap::new());
let docids = word_prefix_pair_proximity_docids
.entry(key)
.or_insert(RoaringBitmap::new());
docids.push(candidate);
}
}