Add DeletionStrategy

Author: dureuill

milli/src/update/delete_documents.rs

@@ -26,7 +26,7 @@ pub struct DeleteDocuments<'t, 'u, 'i> {
     index: &'i Index,
     external_documents_ids: ExternalDocumentsIds<'static>,
     to_delete_docids: RoaringBitmap,
-    disable_soft_deletion: bool,
+    strategy: DeletionStrategy,
 }
 
 /// Result of a [`DeleteDocuments`] operation.
@@ -36,6 +36,36 @@ pub struct DocumentDeletionResult {
     pub remaining_documents: u64,
 }
 
+/// Strategy for deleting documents.
+///
+/// - Soft-deleted documents are simply marked as deleted without being actually removed from DB.
+/// - Hard-deleted documents are definitely suppressed from the DB.
+///
+/// Soft-deleted documents trade disk space for runtime performance.
+///
+/// Note that any of these variants can be used at any given moment for any indexation in a database.
+/// For instance, you can use an [`AlwaysSoft`] followed by an [`AlwaysHard`] option without issue.
+#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash, Default)]
+pub enum DeletionStrategy {
+    #[default]
+    /// Definitely suppress documents according to the number of size of soft-deleted documents
+    Dynamic,
+    /// Never definitely suppress documents
+    AlwaysSoft,
+    /// Always definitely suppress documents
+    AlwaysHard,
+}
+
+impl std::fmt::Display for DeletionStrategy {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        match self {
+            DeletionStrategy::Dynamic => write!(f, "dynamic"),
+            DeletionStrategy::AlwaysSoft => write!(f, "always_soft"),
+            DeletionStrategy::AlwaysHard => write!(f, "always_hard"),
+        }
+    }
+}
+
 /// Result of a [`DeleteDocuments`] operation, used for internal purposes.
 ///
 /// It is a superset of the [`DocumentDeletionResult`] structure, giving
@@ -59,12 +89,12 @@ impl<'t, 'u, 'i> DeleteDocuments<'t, 'u, 'i> {
             index,
             external_documents_ids,
             to_delete_docids: RoaringBitmap::new(),
-            disable_soft_deletion: false,
+            strategy: Default::default(),
         })
     }
 
-    pub fn disable_soft_deletion(&mut self, disable: bool) {
-        self.disable_soft_deletion = disable;
+    pub fn strategy(&mut self, strategy: DeletionStrategy) {
+        self.strategy = strategy;
     }
 
     pub fn delete_document(&mut self, docid: u32) {

milli/src/update/index_documents/mod.rs

@@ -35,8 +35,8 @@ use crate::documents::{obkv_to_object, DocumentsBatchReader};
 use crate::error::{Error, InternalError, UserError};
 pub use crate::update::index_documents::helpers::CursorClonableMmap;
 use crate::update::{
-    self, IndexerConfig, PrefixWordPairsProximityDocids, UpdateIndexingStep, WordPrefixDocids,
-    WordPrefixPositionDocids, WordsPrefixesFst,
+    self, DeletionStrategy, IndexerConfig, PrefixWordPairsProximityDocids, UpdateIndexingStep,
+    WordPrefixDocids, WordPrefixPositionDocids, WordsPrefixesFst,
 };
 use crate::{Index, Result, RoaringBitmapCodec};
 
@@ -88,7 +88,7 @@ pub struct IndexDocumentsConfig {
     pub words_positions_level_group_size: Option<NonZeroU32>,
     pub words_positions_min_level_size: Option<NonZeroU32>,
     pub update_method: IndexDocumentsMethod,
-    pub disable_soft_deletion: bool,
+    pub deletion_strategy: DeletionStrategy,
     pub autogenerate_docids: bool,
 }
 
@@ -332,7 +332,7 @@ where
         // able to simply insert all the documents even if they already exist in the database.
         if !replaced_documents_ids.is_empty() {
             let mut deletion_builder = update::DeleteDocuments::new(self.wtxn, self.index)?;
-            deletion_builder.disable_soft_deletion(self.config.disable_soft_deletion);
+            deletion_builder.strategy(self.config.deletion_strategy);
             debug!("documents to delete {:?}", replaced_documents_ids);
             deletion_builder.delete_documents(&replaced_documents_ids);
             let deleted_documents_result = deletion_builder.execute_inner()?;

milli/src/update/mod.rs

@@ -1,6 +1,6 @@
 pub use self::available_documents_ids::AvailableDocumentsIds;
 pub use self::clear_documents::ClearDocuments;
-pub use self::delete_documents::{DeleteDocuments, DocumentDeletionResult};
+pub use self::delete_documents::{DeleteDocuments, DeletionStrategy, DocumentDeletionResult};
 pub use self::facet::bulk::FacetsUpdateBulk;
 pub use self::facet::incremental::FacetsUpdateIncrementalInner;
 pub use self::index_documents::{