snowykami/meilisearch
1
0
Fork 0
mirror of https://github.com/meilisearch/meilisearch.git synced 2024-12-04 02:55:20 +08:00
Code Issues Actions 11 Packages Projects Releases Wiki Activity

Add document deletion snapshot tests and tests for hard-deletion

Browse Source
This commit is contained in:
Loïc Lecrenier 2022-08-25 14:51:50 +02:00 committed by Loïc Lecrenier
parent d885de1600
commit ab5e56fd16
1 changed files with 150 additions and 29 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

179
milli/src/update/delete_documents.rs
View File

@ -27,7 +27,7 @@ pub struct DeleteDocuments<'t, 'u, 'i> {
external_documents_ids: ExternalDocumentsIds<'static>, external_documents_ids: ExternalDocumentsIds<'static>,
to_delete_docids: RoaringBitmap, to_delete_docids: RoaringBitmap,
#[cfg(test)] #[cfg(test)]
disable_soft_delete: bool, disable_soft_deletion: bool,
} }
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)] #[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
@ -49,13 +49,13 @@ impl<'t, 'u, 'i> DeleteDocuments<'t, 'u, 'i> {
external_documents_ids, external_documents_ids,
to_delete_docids: RoaringBitmap::new(), to_delete_docids: RoaringBitmap::new(),
#[cfg(test)] #[cfg(test)]
disable_soft_delete: false, disable_soft_deletion: false,
}) })
} }
#[cfg(test)] #[cfg(test)]
fn disable_soft_delete(&mut self, disable: bool) { fn disable_soft_deletion(&mut self, disable: bool) {
self.disable_soft_delete = disable; self.disable_soft_deletion = disable;
} }
pub fn delete_document(&mut self, docid: u32) { pub fn delete_document(&mut self, docid: u32) {
@ -156,17 +156,17 @@ impl<'t, 'u, 'i> DeleteDocuments<'t, 'u, 'i> {
// We run the deletion. // We run the deletion.
// - With 100Go of disk and 50Go used including 15Go of soft-deleted documents // - With 100Go of disk and 50Go used including 15Go of soft-deleted documents
// We run the deletion. // We run the deletion.
let disable_soft_delete = { let disable_soft_deletion = {
#[cfg(not(test))] #[cfg(not(test))]
{ {
false false
} }
#[cfg(test)] #[cfg(test)]
{ {
self.disable_soft_delete self.disable_soft_deletion
} }
}; };
if !disable_soft_delete if !disable_soft_deletion
&& percentage_available > 10 && percentage_available > 10
&& percentage_used_by_soft_deleted_documents < 10 && percentage_used_by_soft_deleted_documents < 10
{ {
@ -634,7 +634,7 @@ mod tests {
use super::*; use super::*;
use crate::index::tests::TempIndex; use crate::index::tests::TempIndex;
use crate::Filter; use crate::{db_snap, Filter};
fn delete_documents<'t>( fn delete_documents<'t>(
wtxn: &mut RwTxn<'t, '_>, wtxn: &mut RwTxn<'t, '_>,
@ -680,6 +680,10 @@ mod tests {
wtxn.commit().unwrap(); wtxn.commit().unwrap();
db_snap!(index, documents_ids, @"[]");
db_snap!(index, word_docids, @"");
db_snap!(index, soft_deleted_documents_ids, @"[]");
let rtxn = index.read_txn().unwrap(); let rtxn = index.read_txn().unwrap();
assert!(index.field_distribution(&rtxn).unwrap().is_empty()); assert!(index.field_distribution(&rtxn).unwrap().is_empty());
@ -689,6 +693,10 @@ mod tests {
fn delete_documents_with_strange_primary_key() { fn delete_documents_with_strange_primary_key() {
let index = TempIndex::new(); let index = TempIndex::new();
index
.update_settings(|settings| settings.set_searchable_fields(vec!["name".to_string()]))
.unwrap();
let mut wtxn = index.write_txn().unwrap(); let mut wtxn = index.write_txn().unwrap();
index index
.add_documents_using_wtxn( .add_documents_using_wtxn(
@ -700,14 +708,32 @@ mod tests {
]), ]),
) )
.unwrap(); .unwrap();
wtxn.commit().unwrap();
db_snap!(index, documents_ids, @"[0, 1, 2, ]");
db_snap!(index, word_docids, @r###"
benoit [2, ]
kevin [0, ]
kevina [1, ]
"###);
db_snap!(index, soft_deleted_documents_ids, @"[]");
let mut wtxn = index.write_txn().unwrap();
// Delete not all of the documents but some of them. // Delete not all of the documents but some of them.
let mut builder = DeleteDocuments::new(&mut wtxn, &index).unwrap(); let mut builder = DeleteDocuments::new(&mut wtxn, &index).unwrap();
builder.delete_external_id("0"); builder.delete_external_id("0");
builder.delete_external_id("1"); builder.delete_external_id("1");
builder.execute().unwrap(); builder.execute().unwrap();
wtxn.commit().unwrap(); wtxn.commit().unwrap();
db_snap!(index, documents_ids, @"[2, ]");
db_snap!(index, word_docids, @r###"
benoit [2, ]
kevin [0, ]
kevina [1, ]
"###);
db_snap!(index, soft_deleted_documents_ids, @"[0, 1, ]");
} }
#[test] #[test]
@ -727,26 +753,29 @@ mod tests {
.add_documents_using_wtxn( .add_documents_using_wtxn(
&mut wtxn, &mut wtxn,
documents!([ documents!([
{ "docid": "1_4", "label": "sign" }, { "docid": "1_4", "label": ["sign"] },
{ "docid": "1_5", "label": "letter" }, { "docid": "1_5", "label": ["letter"] },
{ "docid": "1_7", "label": "abstract,cartoon,design,pattern" }, { "docid": "1_7", "label": ["abstract","cartoon","design","pattern"] },
{ "docid": "1_36", "label": "drawing,painting,pattern" }, { "docid": "1_36", "label": ["drawing","painting","pattern"] },
{ "docid": "1_37", "label": "art,drawing,outdoor" }, { "docid": "1_37", "label": ["art","drawing","outdoor"] },
{ "docid": "1_38", "label": "aquarium,art,drawing" }, { "docid": "1_38", "label": ["aquarium","art","drawing"] },
{ "docid": "1_39", "label": "abstract" }, { "docid": "1_39", "label": ["abstract"] },
{ "docid": "1_40", "label": "cartoon" }, { "docid": "1_40", "label": ["cartoon"] },
{ "docid": "1_41", "label": "art,drawing" }, { "docid": "1_41", "label": ["art","drawing"] },
{ "docid": "1_42", "label": "art,pattern" }, { "docid": "1_42", "label": ["art","pattern"] },
{ "docid": "1_43", "label": "abstract,art,drawing,pattern" }, { "docid": "1_43", "label": ["abstract","art","drawing","pattern"] },
{ "docid": "1_44", "label": "drawing" }, { "docid": "1_44", "label": ["drawing"] },
{ "docid": "1_45", "label": "art" }, { "docid": "1_45", "label": ["art"] },
{ "docid": "1_46", "label": "abstract,colorfulness,pattern" }, { "docid": "1_46", "label": ["abstract","colorfulness","pattern"] },
{ "docid": "1_47", "label": "abstract,pattern" }, { "docid": "1_47", "label": ["abstract","pattern"] },
{ "docid": "1_52", "label": "abstract,cartoon" }, { "docid": "1_52", "label": ["abstract","cartoon"] },
{ "docid": "1_57", "label": "abstract,drawing,pattern" }, { "docid": "1_57", "label": ["abstract","drawing","pattern"] },
{ "docid": "1_58", "label": "abstract,art,cartoon" }, { "docid": "1_58", "label": ["abstract","art","cartoon"] },
{ "docid": "1_68", "label": "design" }, { "docid": "1_68", "label": ["design"] },
{ "docid": "1_69", "label": "geometry" } { "docid": "1_69", "label": ["geometry"] },
{ "docid": "1_70", "label2": ["geometry", 1.2] },
{ "docid": "1_71", "label2": ["design", 2.2] },
{ "docid": "1_72", "label2": ["geometry", 1.2] }
]), ]),
) )
.unwrap(); .unwrap();
@ -759,6 +788,86 @@ mod tests {
assert!(results.documents_ids.is_empty()); assert!(results.documents_ids.is_empty());
wtxn.commit().unwrap(); wtxn.commit().unwrap();
db_snap!(index, soft_deleted_documents_ids, @"[0, ]");
db_snap!(index, word_docids, @"e89cd44832e960519823e12b1e7e28af");
db_snap!(index, facet_id_f64_docids, @"");
db_snap!(index, facet_id_string_docids, @"720ee1ba8c18342f3714c5863bc6c1f5");
}
#[test]
fn facet_hard_deletion() {
let index = TempIndex::new();
let mut wtxn = index.write_txn().unwrap();
index
.update_settings_using_wtxn(&mut wtxn, |settings| {
settings.set_primary_key(S("docid"));
settings.set_filterable_fields(hashset! { S("label") });
})
.unwrap();
index
.add_documents_using_wtxn(
&mut wtxn,
documents!([
{ "docid": "1_4", "label": ["sign"] },
{ "docid": "1_5", "label": ["letter"] },
{ "docid": "1_7", "label": ["abstract","cartoon","design","pattern"] },
{ "docid": "1_36", "label": ["drawing","painting","pattern"] },
{ "docid": "1_37", "label": ["art","drawing","outdoor"] },
{ "docid": "1_38", "label": ["aquarium","art","drawing"] },
{ "docid": "1_39", "label": ["abstract"] },
{ "docid": "1_40", "label": ["cartoon"] },
{ "docid": "1_41", "label": ["art","drawing"] },
{ "docid": "1_42", "label": ["art","pattern"] },
{ "docid": "1_43", "label": ["abstract","art","drawing","pattern"] },
{ "docid": "1_44", "label": ["drawing"] },
{ "docid": "1_45", "label": ["art"] },
{ "docid": "1_46", "label": ["abstract","colorfulness","pattern"] },
{ "docid": "1_47", "label": ["abstract","pattern"] },
{ "docid": "1_52", "label": ["abstract","cartoon"] },
{ "docid": "1_57", "label": ["abstract","drawing","pattern"] },
{ "docid": "1_58", "label": ["abstract","art","cartoon"] },
{ "docid": "1_68", "label": ["design"] },
{ "docid": "1_69", "label": ["geometry"] },
{ "docid": "1_70", "label2": ["geometry", 1.2] },
{ "docid": "1_71", "label2": ["design", 2.2] },
{ "docid": "1_72", "label2": ["geometry", 1.2] }
]),
)
.unwrap();
// Delete not all of the documents but some of them.
let mut builder = DeleteDocuments::new(&mut wtxn, &index).unwrap();
builder.disable_soft_deletion(true);
builder.delete_external_id("1_4");
builder.execute().unwrap();
wtxn.commit().unwrap();
db_snap!(index, soft_deleted_documents_ids, 1, @"[]");
db_snap!(index, word_docids, 1, @"999733c2461093d4873321902fc8dcd7");
db_snap!(index, facet_id_f64_docids, 1, @"");
db_snap!(index, facet_id_string_docids, 1, @"a12e80655ed5f0f8e869bb9c32af61e9");
let mut wtxn = index.write_txn().unwrap();
// Delete more than one document
let mut builder = DeleteDocuments::new(&mut wtxn, &index).unwrap();
builder.disable_soft_deletion(true);
builder.delete_external_id("1_5");
builder.delete_external_id("1_7");
builder.delete_external_id("1_70");
builder.delete_external_id("1_72");
builder.execute().unwrap();
wtxn.commit().unwrap();
db_snap!(index, soft_deleted_documents_ids, 2, @"[]");
db_snap!(index, word_docids, 2, @"b892636eaff43c917d5aa8b09c107a02");
db_snap!(index, facet_id_f64_docids, 2, @"");
db_snap!(index, facet_id_string_docids, 2, @"b9946a9cb0ed2df40352e98d6836c8d0");
} }
#[test] #[test]
@ -814,6 +923,8 @@ mod tests {
} }
wtxn.commit().unwrap(); wtxn.commit().unwrap();
db_snap!(index, soft_deleted_documents_ids, @"[0, ]");
} }
#[test] #[test]
@ -869,6 +980,8 @@ mod tests {
} }
wtxn.commit().unwrap(); wtxn.commit().unwrap();
db_snap!(index, soft_deleted_documents_ids, @"[2, 15, ]");
} }
#[test] #[test]
@ -923,6 +1036,10 @@ mod tests {
} }
wtxn.commit().unwrap(); wtxn.commit().unwrap();
db_snap!(index, soft_deleted_documents_ids, @"[4, 5, 6, 11, 16, 18, ]");
db_snap!(index, facet_id_f64_docids, @"20727a38c0b1e1a20a44526b85cf2cbc");
db_snap!(index, facet_id_string_docids, @"");
} }
#[test] #[test]
@ -995,6 +1112,8 @@ mod tests {
} }
wtxn.commit().unwrap(); wtxn.commit().unwrap();
db_snap!(index, soft_deleted_documents_ids, @"[2, 15, ]");
} }
#[test] #[test]
@ -1045,5 +1164,7 @@ mod tests {
assert_eq!(Some(&2), results.get("number")); assert_eq!(Some(&2), results.get("number"));
wtxn.commit().unwrap(); wtxn.commit().unwrap();
db_snap!(index, soft_deleted_documents_ids, @"[2, 15, ]");
} }
} }