Introduce the facet field id values engine database

This commit is contained in:
Clément Renault
2020-11-11 16:04:04 +01:00
parent e0058c1125
commit 92ec908303
3 changed files with 39 additions and 19 deletions

View File

@ -18,33 +18,25 @@ impl<'t, 'u, 'i> ClearDocuments<'t, 'u, 'i> {
word_docids,
docid_word_positions,
word_pair_proximity_docids,
facet_field_id_value_docids,
documents,
} = self.index;
// We clear the word fst.
// We retrieve the number of documents ids that we are deleting.
let number_of_documents = self.index.number_of_documents(self.wtxn)?;
// We clean some of the main engine datastructures.
self.index.put_words_fst(self.wtxn, &fst::Set::default())?;
// We clear the users ids documents ids.
self.index.put_users_ids_documents_ids(self.wtxn, &fst::Map::default())?;
// We retrieve the documents ids.
let documents_ids = self.index.documents_ids(self.wtxn)?;
// We clear the internal documents ids.
self.index.put_documents_ids(self.wtxn, &RoaringBitmap::default())?;
// We clear the word docids.
// Clear the other databases.
word_docids.clear(self.wtxn)?;
// We clear the docid word positions.
docid_word_positions.clear(self.wtxn)?;
// We clear the word pair proximity docids.
word_pair_proximity_docids.clear(self.wtxn)?;
// We clear the documents themselves.
facet_field_id_value_docids.clear(self.wtxn)?;
documents.clear(self.wtxn)?;
Ok(documents_ids.len() as usize)
Ok(number_of_documents)
}
}

View File

@ -76,6 +76,7 @@ impl<'t, 'u, 'i> DeleteDocuments<'t, 'u, 'i> {
word_docids,
docid_word_positions,
word_pair_proximity_docids,
facet_field_id_value_docids,
documents,
} = self.index;
@ -158,7 +159,9 @@ impl<'t, 'u, 'i> DeleteDocuments<'t, 'u, 'i> {
}
// We construct an FST set that contains the words to delete from the words FST.
let words_to_delete = words.iter().filter_map(|(w, d)| if *d { Some(w.as_ref()) } else { None });
let words_to_delete = words.iter().filter_map(|(word, must_remove)| {
if *must_remove { Some(word.as_ref()) } else { None }
});
let words_to_delete = fst::Set::from_iter(words_to_delete)?;
let new_words_fst = {
@ -191,6 +194,20 @@ impl<'t, 'u, 'i> DeleteDocuments<'t, 'u, 'i> {
}
}
drop(iter);
// We delete the documents ids that are under the facet field id values.
let mut iter = facet_field_id_value_docids.iter_mut(self.wtxn)?;
while let Some(result) = iter.next() {
let (bytes, mut docids) = result?;
docids.difference_with(&self.documents_ids);
if docids.is_empty() {
iter.del_current()?;
} else {
iter.put_current(bytes, &docids)?;
}
}
Ok(self.documents_ids.len() as usize)
}
}