chore: Simplify the update application

This commit is contained in:
Clément Renault 2019-08-22 14:18:50 +02:00 committed by Clément Renault
parent 4deee93a55
commit b0be06540a
No known key found for this signature in database
GPG Key ID: 92ADA4E935E71FA4
6 changed files with 336 additions and 411 deletions

View File

@ -1,15 +1,15 @@
use std::collections::HashSet; use std::collections::HashSet;
use std::sync::Arc; use std::sync::Arc;
use meilidb_core::DocumentId;
use fst::{SetBuilder, set::OpBuilder}; use fst::{SetBuilder, set::OpBuilder};
use sdset::{SetOperation, duo::Union}; use sdset::{SetOperation, duo::Union};
use serde::Serialize;
use crate::indexer::Indexer; use crate::indexer::Indexer;
use crate::serde::{extract_document_id, Serializer, RamDocumentStore}; use crate::serde::{extract_document_id, Serializer, RamDocumentStore};
use crate::RankedMap; use crate::RankedMap;
use super::{Error, Index, FinalDocumentsDeletion}; use super::{Error, Index, apply_documents_deletion};
use super::index::Cache; use super::index::Cache;
pub struct DocumentsAddition<'a, D> { pub struct DocumentsAddition<'a, D> {
@ -33,123 +33,106 @@ impl<'a, D> DocumentsAddition<'a, D> {
} }
} }
pub struct FinalDocumentsAddition<'a> { pub fn apply_documents_addition(
inner: &'a Index, index: &Index,
document_ids: HashSet<DocumentId>, mut ranked_map: RankedMap,
document_store: RamDocumentStore, addition: Vec<serde_json::Value>,
indexer: Indexer, ) -> Result<(), Error>
ranked_map: RankedMap, {
} let mut document_ids = HashSet::new();
let mut document_store = RamDocumentStore::new();
let mut indexer = Indexer::new();
impl<'a> FinalDocumentsAddition<'a> { let schema = &index.schema();
pub fn new(inner: &'a Index, ranked_map: RankedMap) -> FinalDocumentsAddition<'a> { let identifier = schema.identifier_name();
FinalDocumentsAddition {
inner,
document_ids: HashSet::new(),
document_store: RamDocumentStore::new(),
indexer: Indexer::new(),
ranked_map,
}
}
pub fn update_document<D>(&mut self, document: D) -> Result<(), Error>
where D: serde::Serialize,
{
let schema = &self.inner.schema();
let identifier = schema.identifier_name();
for document in addition {
let document_id = match extract_document_id(identifier, &document)? { let document_id = match extract_document_id(identifier, &document)? {
Some(id) => id, Some(id) => id,
None => return Err(Error::MissingDocumentId), None => return Err(Error::MissingDocumentId),
}; };
// 1. store the document id for future deletion // 1. store the document id for future deletion
self.document_ids.insert(document_id); document_ids.insert(document_id);
// 2. index the document fields in ram stores // 2. index the document fields in ram stores
let serializer = Serializer { let serializer = Serializer {
schema, schema,
document_store: &mut self.document_store, document_store: &mut document_store,
indexer: &mut self.indexer, indexer: &mut indexer,
ranked_map: &mut self.ranked_map, ranked_map: &mut ranked_map,
document_id, document_id,
}; };
document.serialize(serializer)?; document.serialize(serializer)?;
Ok(())
} }
pub fn finalize(self) -> Result<(), Error> { let ref_index = index.as_ref();
let ref_index = self.inner.as_ref(); let docs_words = ref_index.docs_words_index;
let docs_words = ref_index.docs_words_index; let documents = ref_index.documents_index;
let documents = ref_index.documents_index; let main = ref_index.main_index;
let main = ref_index.main_index; let words = ref_index.words_index;
let words = ref_index.words_index;
// 1. remove the previous documents match indexes // 1. remove the previous documents match indexes
let mut documents_deletion = FinalDocumentsDeletion::new(self.inner, self.ranked_map.clone()); let document_ids = document_ids.into_iter().collect();
documents_deletion.extend(self.document_ids); apply_documents_deletion(index, ranked_map.clone(), document_ids)?;
documents_deletion.finalize()?;
// 2. insert new document attributes in the database // 2. insert new document attributes in the database
for ((id, attr), value) in self.document_store.into_inner() { for ((id, attr), value) in document_store.into_inner() {
documents.set_document_field(id, attr, value)?; documents.set_document_field(id, attr, value)?;
} }
let indexed = self.indexer.build(); let indexed = indexer.build();
let mut delta_words_builder = SetBuilder::memory(); let mut delta_words_builder = SetBuilder::memory();
for (word, delta_set) in indexed.words_doc_indexes { for (word, delta_set) in indexed.words_doc_indexes {
delta_words_builder.insert(&word).unwrap(); delta_words_builder.insert(&word).unwrap();
let set = match words.doc_indexes(&word)? { let set = match words.doc_indexes(&word)? {
Some(set) => Union::new(&set, &delta_set).into_set_buf(), Some(set) => Union::new(&set, &delta_set).into_set_buf(),
None => delta_set, None => delta_set,
};
words.set_doc_indexes(&word, &set)?;
}
for (id, words) in indexed.docs_words {
docs_words.set_doc_words(id, &words)?;
}
let delta_words = delta_words_builder
.into_inner()
.and_then(fst::Set::from_bytes)
.unwrap();
let words = match main.words_set()? {
Some(words) => {
let op = OpBuilder::new()
.add(words.stream())
.add(delta_words.stream())
.r#union();
let mut words_builder = SetBuilder::memory();
words_builder.extend_stream(op).unwrap();
words_builder
.into_inner()
.and_then(fst::Set::from_bytes)
.unwrap()
},
None => delta_words,
}; };
main.set_words_set(&words)?; words.set_doc_indexes(&word, &set)?;
main.set_ranked_map(&self.ranked_map)?;
// update the "consistent" view of the Index
let cache = ref_index.cache;
let words = Arc::new(words);
let ranked_map = self.ranked_map;
let synonyms = cache.synonyms.clone();
let schema = cache.schema.clone();
let cache = Cache { words, synonyms, schema, ranked_map };
self.inner.cache.store(Arc::new(cache));
Ok(())
} }
for (id, words) in indexed.docs_words {
docs_words.set_doc_words(id, &words)?;
}
let delta_words = delta_words_builder
.into_inner()
.and_then(fst::Set::from_bytes)
.unwrap();
let words = match main.words_set()? {
Some(words) => {
let op = OpBuilder::new()
.add(words.stream())
.add(delta_words.stream())
.r#union();
let mut words_builder = SetBuilder::memory();
words_builder.extend_stream(op).unwrap();
words_builder
.into_inner()
.and_then(fst::Set::from_bytes)
.unwrap()
},
None => delta_words,
};
main.set_words_set(&words)?;
main.set_ranked_map(&ranked_map)?;
// update the "consistent" view of the Index
let cache = ref_index.cache;
let words = Arc::new(words);
let synonyms = cache.synonyms.clone();
let schema = cache.schema.clone();
let cache = Cache { words, synonyms, schema, ranked_map };
index.cache.store(Arc::new(cache));
Ok(())
} }

View File

@ -21,10 +21,25 @@ impl<'a> DocumentsDeletion<'a> {
DocumentsDeletion { index, documents: Vec::new() } DocumentsDeletion { index, documents: Vec::new() }
} }
pub fn delete_document(&mut self, document_id: DocumentId) { pub fn delete_document_by_id(&mut self, document_id: DocumentId) {
self.documents.push(document_id); self.documents.push(document_id);
} }
pub fn delete_document<D>(&mut self, document: D) -> Result<(), Error>
where D: serde::Serialize,
{
let schema = self.index.schema();
let identifier = schema.identifier_name();
let document_id = match extract_document_id(identifier, &document)? {
Some(id) => id,
None => return Err(Error::MissingDocumentId),
};
self.delete_document_by_id(document_id);
Ok(())
}
pub fn finalize(self) -> Result<u64, Error> { pub fn finalize(self) -> Result<u64, Error> {
self.index.push_documents_deletion(self.documents) self.index.push_documents_deletion(self.documents)
} }
@ -36,129 +51,95 @@ impl Extend<DocumentId> for DocumentsDeletion<'_> {
} }
} }
pub struct FinalDocumentsDeletion<'a> { pub fn apply_documents_deletion(
inner: &'a Index, index: &Index,
documents: Vec<DocumentId>, mut ranked_map: RankedMap,
ranked_map: RankedMap, deletion: Vec<DocumentId>,
} ) -> Result<(), Error>
{
let ref_index = index.as_ref();
let schema = index.schema();
let docs_words = ref_index.docs_words_index;
let documents = ref_index.documents_index;
let main = ref_index.main_index;
let words = ref_index.words_index;
impl<'a> FinalDocumentsDeletion<'a> { let idset = SetBuf::from_dirty(deletion);
pub fn new(inner: &'a Index, ranked_map: RankedMap) -> FinalDocumentsDeletion {
FinalDocumentsDeletion { inner, documents: Vec::new(), ranked_map }
}
fn delete_document_by_id(&mut self, id: DocumentId) { // collect the ranked attributes according to the schema
self.documents.push(id); let ranked_attrs: Vec<_> = schema.iter()
} .filter_map(|(_, attr, prop)| {
if prop.is_ranked() { Some(attr) } else { None }
})
.collect();
pub fn delete_document<D>(&mut self, document: D) -> Result<(), Error> let mut words_document_ids = HashMap::new();
where D: serde::Serialize, for id in idset {
{ // remove all the ranked attributes from the ranked_map
let schema = &self.inner.schema(); for ranked_attr in &ranked_attrs {
let identifier = schema.identifier_name(); ranked_map.remove(id, *ranked_attr);
}
let document_id = match extract_document_id(identifier, &document)? { if let Some(words) = docs_words.doc_words(id)? {
Some(id) => id, let mut stream = words.stream();
None => return Err(Error::MissingDocumentId), while let Some(word) = stream.next() {
}; let word = word.to_vec();
words_document_ids.entry(word).or_insert_with(Vec::new).push(id);
self.delete_document_by_id(document_id);
Ok(())
}
pub fn finalize(mut self) -> Result<(), Error> {
let ref_index = self.inner.as_ref();
let schema = self.inner.schema();
let docs_words = ref_index.docs_words_index;
let documents = ref_index.documents_index;
let main = ref_index.main_index;
let words = ref_index.words_index;
let idset = SetBuf::from_dirty(self.documents);
// collect the ranked attributes according to the schema
let ranked_attrs: Vec<_> = schema.iter()
.filter_map(|(_, attr, prop)| {
if prop.is_ranked() { Some(attr) } else { None }
})
.collect();
let mut words_document_ids = HashMap::new();
for id in idset {
// remove all the ranked attributes from the ranked_map
for ranked_attr in &ranked_attrs {
self.ranked_map.remove(id, *ranked_attr);
} }
}
}
if let Some(words) = docs_words.doc_words(id)? { let mut removed_words = BTreeSet::new();
let mut stream = words.stream(); for (word, document_ids) in words_document_ids {
while let Some(word) = stream.next() { let document_ids = SetBuf::from_dirty(document_ids);
let word = word.to_vec();
words_document_ids.entry(word).or_insert_with(Vec::new).push(id); if let Some(doc_indexes) = words.doc_indexes(&word)? {
} let op = DifferenceByKey::new(&doc_indexes, &document_ids, |d| d.document_id, |id| *id);
let doc_indexes = op.into_set_buf();
if !doc_indexes.is_empty() {
words.set_doc_indexes(&word, &doc_indexes)?;
} else {
words.del_doc_indexes(&word)?;
removed_words.insert(word);
} }
} }
let mut removed_words = BTreeSet::new(); for id in document_ids {
for (word, document_ids) in words_document_ids { documents.del_all_document_fields(id)?;
let document_ids = SetBuf::from_dirty(document_ids); docs_words.del_doc_words(id)?;
if let Some(doc_indexes) = words.doc_indexes(&word)? {
let op = DifferenceByKey::new(&doc_indexes, &document_ids, |d| d.document_id, |id| *id);
let doc_indexes = op.into_set_buf();
if !doc_indexes.is_empty() {
words.set_doc_indexes(&word, &doc_indexes)?;
} else {
words.del_doc_indexes(&word)?;
removed_words.insert(word);
}
}
for id in document_ids {
documents.del_all_document_fields(id)?;
docs_words.del_doc_words(id)?;
}
} }
let removed_words = fst::Set::from_iter(removed_words).unwrap();
let words = match main.words_set()? {
Some(words_set) => {
let op = fst::set::OpBuilder::new()
.add(words_set.stream())
.add(removed_words.stream())
.difference();
let mut words_builder = SetBuilder::memory();
words_builder.extend_stream(op).unwrap();
words_builder
.into_inner()
.and_then(fst::Set::from_bytes)
.unwrap()
},
None => fst::Set::default(),
};
main.set_words_set(&words)?;
main.set_ranked_map(&self.ranked_map)?;
// update the "consistent" view of the Index
let cache = ref_index.cache;
let words = Arc::new(words);
let ranked_map = self.ranked_map;
let synonyms = cache.synonyms.clone();
let schema = cache.schema.clone();
let cache = Cache { words, synonyms, schema, ranked_map };
self.inner.cache.store(Arc::new(cache));
Ok(())
} }
}
impl Extend<DocumentId> for FinalDocumentsDeletion<'_> { let removed_words = fst::Set::from_iter(removed_words).unwrap();
fn extend<T: IntoIterator<Item=DocumentId>>(&mut self, iter: T) { let words = match main.words_set()? {
self.documents.extend(iter) Some(words_set) => {
} let op = fst::set::OpBuilder::new()
.add(words_set.stream())
.add(removed_words.stream())
.difference();
let mut words_builder = SetBuilder::memory();
words_builder.extend_stream(op).unwrap();
words_builder
.into_inner()
.and_then(fst::Set::from_bytes)
.unwrap()
},
None => fst::Set::default(),
};
main.set_words_set(&words)?;
main.set_ranked_map(&ranked_map)?;
// update the "consistent" view of the Index
let cache = ref_index.cache;
let words = Arc::new(words);
let synonyms = cache.synonyms.clone();
let schema = cache.schema.clone();
let cache = Cache { words, synonyms, schema, ranked_map };
index.cache.store(Arc::new(cache));
Ok(())
} }

View File

@ -1,5 +1,4 @@
use std::collections::{HashSet, BTreeMap}; use std::collections::{HashSet, BTreeMap};
use std::convert::TryInto;
use std::sync::Arc; use std::sync::Arc;
use std::thread; use std::thread;
@ -24,10 +23,10 @@ use self::synonyms_index::SynonymsIndex;
use self::words_index::WordsIndex; use self::words_index::WordsIndex;
use super::{ use super::{
DocumentsAddition, FinalDocumentsAddition, DocumentsAddition, DocumentsDeletion,
DocumentsDeletion, FinalDocumentsDeletion, SynonymsAddition, SynonymsDeletion,
SynonymsAddition, FinalSynonymsAddition, apply_documents_addition, apply_documents_deletion,
SynonymsDeletion, FinalSynonymsDeletion, apply_synonyms_addition, apply_synonyms_deletion,
}; };
mod custom_settings_index; mod custom_settings_index;
@ -71,33 +70,23 @@ fn spawn_update_system(index: Index) -> thread::JoinHandle<()> {
let results = &index.updates_results_index; let results = &index.updates_results_index;
(updates, results).transaction(|(updates, results)| { (updates, results).transaction(|(updates, results)| {
let update = updates.remove(&key)?.unwrap(); let update = updates.remove(&key)?.unwrap();
let array_id = key.as_ref().try_into().unwrap();
let id = u64::from_be_bytes(array_id);
// this is an emulation of the try block (#31436) // this is an emulation of the try block (#31436)
let result: Result<(), Error> = (|| { let result: Result<(), Error> = (|| {
match bincode::deserialize(&update)? { match bincode::deserialize(&update)? {
UpdateOwned::DocumentsAddition(documents) => { UpdateOwned::DocumentsAddition(documents) => {
let ranked_map = index.cache.load().ranked_map.clone(); let ranked_map = index.cache.load().ranked_map.clone();
let mut addition = FinalDocumentsAddition::new(&index, ranked_map); apply_documents_addition(&index, ranked_map, documents)?;
for document in documents {
addition.update_document(document)?;
}
addition.finalize()?;
}, },
UpdateOwned::DocumentsDeletion(documents) => { UpdateOwned::DocumentsDeletion(documents) => {
let ranked_map = index.cache.load().ranked_map.clone(); let ranked_map = index.cache.load().ranked_map.clone();
let mut deletion = FinalDocumentsDeletion::new(&index, ranked_map); apply_documents_deletion(&index, ranked_map, documents)?;
deletion.extend(documents);
deletion.finalize()?;
}, },
UpdateOwned::SynonymsAddition(synonyms) => { UpdateOwned::SynonymsAddition(synonyms) => {
let addition = FinalSynonymsAddition::from_map(&index, synonyms); apply_synonyms_addition(&index, synonyms)?;
addition.finalize()?;
}, },
UpdateOwned::SynonymsDeletion(synonyms) => { UpdateOwned::SynonymsDeletion(synonyms) => {
let deletion = FinalSynonymsDeletion::from_map(&index, synonyms); apply_synonyms_deletion(&index, synonyms)?;
deletion.finalize()?;
}, },
} }
Ok(()) Ok(())
@ -105,7 +94,7 @@ fn spawn_update_system(index: Index) -> thread::JoinHandle<()> {
let result = result.map_err(|e| e.to_string()); let result = result.map_err(|e| e.to_string());
let value = bincode::serialize(&result).unwrap(); let value = bincode::serialize(&result).unwrap();
results.insert(&array_id, value) results.insert(&key, value)
}) })
.unwrap(); .unwrap();
} }
@ -310,7 +299,12 @@ impl Index {
self.raw_push_update(update) self.raw_push_update(update)
} }
pub(crate) fn push_documents_deletion(&self, deletion: Vec<DocumentId>) -> Result<u64, Error> { pub(crate) fn push_documents_deletion(
&self,
deletion: Vec<DocumentId>,
) -> Result<u64, Error>
{
let deletion = Update::<()>::DocumentsDeletion(deletion);
let update = bincode::serialize(&deletion)?; let update = bincode::serialize(&deletion)?;
self.raw_push_update(update) self.raw_push_update(update)
} }
@ -320,6 +314,7 @@ impl Index {
addition: BTreeMap<String, Vec<String>>, addition: BTreeMap<String, Vec<String>>,
) -> Result<u64, Error> ) -> Result<u64, Error>
{ {
let addition = Update::<()>::SynonymsAddition(addition);
let update = bincode::serialize(&addition)?; let update = bincode::serialize(&addition)?;
self.raw_push_update(update) self.raw_push_update(update)
} }
@ -329,6 +324,7 @@ impl Index {
deletion: BTreeMap<String, Option<Vec<String>>>, deletion: BTreeMap<String, Option<Vec<String>>>,
) -> Result<u64, Error> ) -> Result<u64, Error>
{ {
let deletion = Update::<()>::SynonymsDeletion(deletion);
let update = bincode::serialize(&deletion)?; let update = bincode::serialize(&deletion)?;
self.raw_push_update(update) self.raw_push_update(update)
} }

View File

@ -15,10 +15,15 @@ mod synonyms_deletion;
pub use self::error::Error; pub use self::error::Error;
pub use self::index::{Index, CustomSettingsIndex}; pub use self::index::{Index, CustomSettingsIndex};
use self::documents_addition::{DocumentsAddition, FinalDocumentsAddition}; pub use self::documents_addition::DocumentsAddition;
use self::documents_deletion::{DocumentsDeletion, FinalDocumentsDeletion}; pub use self::documents_deletion::DocumentsDeletion;
use self::synonyms_addition::{SynonymsAddition, FinalSynonymsAddition}; pub use self::synonyms_addition::SynonymsAddition;
use self::synonyms_deletion::{SynonymsDeletion, FinalSynonymsDeletion}; pub use self::synonyms_deletion::SynonymsDeletion;
use self::documents_addition::apply_documents_addition;
use self::documents_deletion::apply_documents_deletion;
use self::synonyms_addition::apply_synonyms_addition;
use self::synonyms_deletion::apply_synonyms_deletion;
fn load_indexes(tree: &sled::Tree) -> Result<HashSet<String>, Error> { fn load_indexes(tree: &sled::Tree) -> Result<HashSet<String>, Error> {
match tree.get("indexes")? { match tree.get("indexes")? {

View File

@ -21,10 +21,10 @@ impl<'a> SynonymsAddition<'a> {
pub fn add_synonym<S, T, I>(&mut self, synonym: S, alternatives: I) pub fn add_synonym<S, T, I>(&mut self, synonym: S, alternatives: I)
where S: AsRef<str>, where S: AsRef<str>,
T: AsRef<str>, T: AsRef<str>,
I: Iterator<Item=T>, I: IntoIterator<Item=T>,
{ {
let synonym = normalize_str(synonym.as_ref()); let synonym = normalize_str(synonym.as_ref());
let alternatives = alternatives.map(|s| s.as_ref().to_lowercase()); let alternatives = alternatives.into_iter().map(|s| s.as_ref().to_lowercase());
self.synonyms.entry(synonym).or_insert_with(Vec::new).extend(alternatives); self.synonyms.entry(synonym).or_insert_with(Vec::new).extend(alternatives);
} }
@ -33,87 +33,62 @@ impl<'a> SynonymsAddition<'a> {
} }
} }
pub struct FinalSynonymsAddition<'a> { pub fn apply_synonyms_addition(
inner: &'a Index, index: &Index,
synonyms: BTreeMap<String, Vec<String>>, addition: BTreeMap<String, Vec<String>>,
} ) -> Result<(), Error>
{
let ref_index = index.as_ref();
let synonyms = ref_index.synonyms_index;
let main = ref_index.main_index;
impl<'a> FinalSynonymsAddition<'a> { let mut synonyms_builder = SetBuilder::memory();
pub fn new(inner: &'a Index) -> FinalSynonymsAddition<'a> {
FinalSynonymsAddition { inner, synonyms: BTreeMap::new() }
}
pub fn from_map( for (synonym, alternatives) in addition {
inner: &'a Index, synonyms_builder.insert(&synonym).unwrap();
synonyms: BTreeMap<String, Vec<String>>,
) -> FinalSynonymsAddition<'a>
{
FinalSynonymsAddition { inner, synonyms }
}
pub fn add_synonym<S, T, I>(&mut self, synonym: S, alternatives: I) let alternatives = {
where S: AsRef<str>, let alternatives = SetBuf::from_dirty(alternatives);
T: AsRef<str>, let mut alternatives_builder = SetBuilder::memory();
I: IntoIterator<Item=T>, alternatives_builder.extend_iter(alternatives).unwrap();
{ alternatives_builder.into_inner().unwrap()
let synonym = normalize_str(synonym.as_ref());
let alternatives = alternatives.into_iter().map(|s| s.as_ref().to_lowercase());
self.synonyms.entry(synonym).or_insert_with(Vec::new).extend(alternatives);
}
pub fn finalize(self) -> Result<(), Error> {
let ref_index = self.inner.as_ref();
let synonyms = ref_index.synonyms_index;
let main = ref_index.main_index;
let mut synonyms_builder = SetBuilder::memory();
for (synonym, alternatives) in self.synonyms {
synonyms_builder.insert(&synonym).unwrap();
let alternatives = {
let alternatives = SetBuf::from_dirty(alternatives);
let mut alternatives_builder = SetBuilder::memory();
alternatives_builder.extend_iter(alternatives).unwrap();
alternatives_builder.into_inner().unwrap()
};
synonyms.set_alternatives_to(synonym.as_bytes(), alternatives)?;
}
let delta_synonyms = synonyms_builder
.into_inner()
.and_then(fst::Set::from_bytes)
.unwrap();
let synonyms = match main.synonyms_set()? {
Some(synonyms) => {
let op = OpBuilder::new()
.add(synonyms.stream())
.add(delta_synonyms.stream())
.r#union();
let mut synonyms_builder = SetBuilder::memory();
synonyms_builder.extend_stream(op).unwrap();
synonyms_builder
.into_inner()
.and_then(fst::Set::from_bytes)
.unwrap()
},
None => delta_synonyms,
}; };
synonyms.set_alternatives_to(synonym.as_bytes(), alternatives)?;
main.set_synonyms_set(&synonyms)?;
// update the "consistent" view of the Index
let cache = ref_index.cache;
let words = Arc::new(main.words_set()?.unwrap_or_default());
let ranked_map = cache.ranked_map.clone();
let synonyms = Arc::new(synonyms);
let schema = cache.schema.clone();
let cache = Cache { words, synonyms, schema, ranked_map };
self.inner.cache.store(Arc::new(cache));
Ok(())
} }
let delta_synonyms = synonyms_builder
.into_inner()
.and_then(fst::Set::from_bytes)
.unwrap();
let synonyms = match main.synonyms_set()? {
Some(synonyms) => {
let op = OpBuilder::new()
.add(synonyms.stream())
.add(delta_synonyms.stream())
.r#union();
let mut synonyms_builder = SetBuilder::memory();
synonyms_builder.extend_stream(op).unwrap();
synonyms_builder
.into_inner()
.and_then(fst::Set::from_bytes)
.unwrap()
},
None => delta_synonyms,
};
main.set_synonyms_set(&synonyms)?;
// update the "consistent" view of the Index
let cache = ref_index.cache;
let words = Arc::new(main.words_set()?.unwrap_or_default());
let ranked_map = cache.ranked_map.clone();
let synonyms = Arc::new(synonyms);
let schema = cache.schema.clone();
let cache = Cache { words, synonyms, schema, ranked_map };
index.cache.store(Arc::new(cache));
Ok(())
} }

View File

@ -43,110 +43,95 @@ impl<'a> SynonymsDeletion<'a> {
} }
} }
pub struct FinalSynonymsDeletion<'a> { pub fn apply_synonyms_deletion(
inner: &'a Index, index: &Index,
synonyms: BTreeMap<String, Option<Vec<String>>>, deletion: BTreeMap<String, Option<Vec<String>>>,
} ) -> Result<(), Error>
{
let ref_index = index.as_ref();
let synonyms = ref_index.synonyms_index;
let main = ref_index.main_index;
impl<'a> FinalSynonymsDeletion<'a> { let mut delete_whole_synonym_builder = SetBuilder::memory();
pub fn new(inner: &'a Index) -> FinalSynonymsDeletion<'a> {
FinalSynonymsDeletion { inner, synonyms: BTreeMap::new() }
}
pub fn from_map( for (synonym, alternatives) in deletion {
inner: &'a Index, match alternatives {
synonyms: BTreeMap<String, Option<Vec<String>>>, Some(alternatives) => {
) -> FinalSynonymsDeletion<'a> let prev_alternatives = synonyms.alternatives_to(synonym.as_bytes())?;
{ let prev_alternatives = match prev_alternatives {
FinalSynonymsDeletion { inner, synonyms } Some(alternatives) => alternatives,
} None => continue,
};
pub fn finalize(self) -> Result<(), Error> { let delta_alternatives = {
let ref_index = self.inner.as_ref(); let alternatives = SetBuf::from_dirty(alternatives);
let synonyms = ref_index.synonyms_index; let mut builder = SetBuilder::memory();
let main = ref_index.main_index; builder.extend_iter(alternatives).unwrap();
builder.into_inner()
.and_then(fst::Set::from_bytes)
.unwrap()
};
let mut delete_whole_synonym_builder = SetBuilder::memory();
for (synonym, alternatives) in self.synonyms {
match alternatives {
Some(alternatives) => {
let prev_alternatives = synonyms.alternatives_to(synonym.as_bytes())?;
let prev_alternatives = match prev_alternatives {
Some(alternatives) => alternatives,
None => continue,
};
let delta_alternatives = {
let alternatives = SetBuf::from_dirty(alternatives);
let mut builder = SetBuilder::memory();
builder.extend_iter(alternatives).unwrap();
builder.into_inner()
.and_then(fst::Set::from_bytes)
.unwrap()
};
let op = OpBuilder::new()
.add(prev_alternatives.stream())
.add(delta_alternatives.stream())
.difference();
let (alternatives, empty_alternatives) = {
let mut builder = SetBuilder::memory();
let len = builder.get_ref().len();
builder.extend_stream(op).unwrap();
let is_empty = len == builder.get_ref().len();
let alternatives = builder.into_inner().unwrap();
(alternatives, is_empty)
};
if empty_alternatives {
delete_whole_synonym_builder.insert(synonym.as_bytes())?;
} else {
synonyms.set_alternatives_to(synonym.as_bytes(), alternatives)?;
}
},
None => {
delete_whole_synonym_builder.insert(&synonym).unwrap();
synonyms.del_alternatives_of(synonym.as_bytes())?;
}
}
}
let delta_synonyms = delete_whole_synonym_builder
.into_inner()
.and_then(fst::Set::from_bytes)
.unwrap();
let synonyms = match main.synonyms_set()? {
Some(synonyms) => {
let op = OpBuilder::new() let op = OpBuilder::new()
.add(synonyms.stream()) .add(prev_alternatives.stream())
.add(delta_synonyms.stream()) .add(delta_alternatives.stream())
.difference(); .difference();
let mut synonyms_builder = SetBuilder::memory(); let (alternatives, empty_alternatives) = {
synonyms_builder.extend_stream(op).unwrap(); let mut builder = SetBuilder::memory();
synonyms_builder let len = builder.get_ref().len();
.into_inner() builder.extend_stream(op).unwrap();
.and_then(fst::Set::from_bytes) let is_empty = len == builder.get_ref().len();
.unwrap() let alternatives = builder.into_inner().unwrap();
(alternatives, is_empty)
};
if empty_alternatives {
delete_whole_synonym_builder.insert(synonym.as_bytes())?;
} else {
synonyms.set_alternatives_to(synonym.as_bytes(), alternatives)?;
}
}, },
None => fst::Set::default(), None => {
}; delete_whole_synonym_builder.insert(&synonym).unwrap();
synonyms.del_alternatives_of(synonym.as_bytes())?;
main.set_synonyms_set(&synonyms)?; }
}
// update the "consistent" view of the Index
let cache = ref_index.cache;
let words = Arc::new(main.words_set()?.unwrap_or_default());
let ranked_map = cache.ranked_map.clone();
let synonyms = Arc::new(synonyms);
let schema = cache.schema.clone();
let cache = Cache { words, synonyms, schema, ranked_map };
self.inner.cache.store(Arc::new(cache));
Ok(())
} }
let delta_synonyms = delete_whole_synonym_builder
.into_inner()
.and_then(fst::Set::from_bytes)
.unwrap();
let synonyms = match main.synonyms_set()? {
Some(synonyms) => {
let op = OpBuilder::new()
.add(synonyms.stream())
.add(delta_synonyms.stream())
.difference();
let mut synonyms_builder = SetBuilder::memory();
synonyms_builder.extend_stream(op).unwrap();
synonyms_builder
.into_inner()
.and_then(fst::Set::from_bytes)
.unwrap()
},
None => fst::Set::default(),
};
main.set_synonyms_set(&synonyms)?;
// update the "consistent" view of the Index
let cache = ref_index.cache;
let words = Arc::new(main.words_set()?.unwrap_or_default());
let ranked_map = cache.ranked_map.clone();
let synonyms = Arc::new(synonyms);
let schema = cache.schema.clone();
let cache = Cache { words, synonyms, schema, ranked_map };
index.cache.store(Arc::new(cache));
Ok(())
} }