From 8d2e3e4aba86d62df0faf1f08d561c17e55951c0 Mon Sep 17 00:00:00 2001
From: Kerollmops <clement@meilisearch.com>
Date: Mon, 21 Feb 2022 17:59:03 +0100
Subject: [PATCH] Add a new wiki benchmark to test multi batch indexing

---
 benchmarks/benches/indexing.rs | 71 ++++++++++++++++++++++++++++++++++
 benchmarks/build.rs            |  6 +++
 2 files changed, 77 insertions(+)

diff --git a/benchmarks/benches/indexing.rs b/benchmarks/benches/indexing.rs
index 8536dabe8..97eee0a34 100644
--- a/benchmarks/benches/indexing.rs
+++ b/benchmarks/benches/indexing.rs
@@ -294,6 +294,76 @@ fn indexing_wiki(c: &mut Criterion) {
     });
 }
 
+fn indexing_wiki_in_three_batches(c: &mut Criterion) {
+    let mut group = c.benchmark_group("indexing");
+    group.sample_size(10);
+    group.bench_function("Indexing wiki in three batches", |b| {
+        b.iter_with_setup(
+            move || {
+                let index = setup_index();
+
+                let config = IndexerConfig::default();
+                let mut wtxn = index.write_txn().unwrap();
+                let mut builder = Settings::new(&mut wtxn, &index, &config);
+
+                builder.set_primary_key("id".to_owned());
+                let displayed_fields =
+                    ["title", "body", "url"].iter().map(|s| s.to_string()).collect();
+                builder.set_displayed_fields(displayed_fields);
+
+                let searchable_fields = ["title", "body"].iter().map(|s| s.to_string()).collect();
+                builder.set_searchable_fields(searchable_fields);
+
+                // there is NO faceted fields at all
+                builder.execute(|_| ()).unwrap();
+
+                // We index only one half of the dataset in the setup part
+                // as we don't care about the time it take.
+                let config = IndexerConfig::default();
+                let indexing_config =
+                    IndexDocumentsConfig { autogenerate_docids: true, ..Default::default() };
+                let mut builder =
+                    IndexDocuments::new(&mut wtxn, &index, &config, indexing_config, |_| ());
+                let documents =
+                    utils::documents_from(datasets_paths::SMOL_WIKI_ARTICLES_1_2, "csv");
+                builder.add_documents(documents).unwrap();
+                builder.execute().unwrap();
+
+                wtxn.commit().unwrap();
+
+                index
+            },
+            move |index| {
+                let config = IndexerConfig::default();
+                let indexing_config =
+                    IndexDocumentsConfig { autogenerate_docids: true, ..Default::default() };
+                let mut wtxn = index.write_txn().unwrap();
+                let mut builder =
+                    IndexDocuments::new(&mut wtxn, &index, &config, indexing_config, |_| ());
+
+                let documents =
+                    utils::documents_from(datasets_paths::SMOL_WIKI_ARTICLES_3_4, "csv");
+                builder.add_documents(documents).unwrap();
+                builder.execute().unwrap();
+
+                let indexing_config =
+                    IndexDocumentsConfig { autogenerate_docids: true, ..Default::default() };
+                let mut builder =
+                    IndexDocuments::new(&mut wtxn, &index, &config, indexing_config, |_| ());
+
+                let documents =
+                    utils::documents_from(datasets_paths::SMOL_WIKI_ARTICLES_4_4, "csv");
+                builder.add_documents(documents).unwrap();
+                builder.execute().unwrap();
+
+                wtxn.commit().unwrap();
+
+                index.prepare_for_closing().wait();
+            },
+        )
+    });
+}
+
 fn indexing_movies_default(c: &mut Criterion) {
     let mut group = c.benchmark_group("indexing");
     group.sample_size(10);
@@ -405,6 +475,7 @@ criterion_group!(
     indexing_songs_without_faceted_fields,
     indexing_songs_in_three_batches_default,
     indexing_wiki,
+    indexing_wiki_in_three_batches,
     indexing_movies_default,
     indexing_geo
 );
diff --git a/benchmarks/build.rs b/benchmarks/build.rs
index 90ebf70af..66a0a841b 100644
--- a/benchmarks/build.rs
+++ b/benchmarks/build.rs
@@ -15,6 +15,9 @@ const DATASET_SONGS_1_2: (&str, &str) = ("smol-songs-1_2", "csv");
 const DATASET_SONGS_3_4: (&str, &str) = ("smol-songs-3_4", "csv");
 const DATASET_SONGS_4_4: (&str, &str) = ("smol-songs-4_4", "csv");
 const DATASET_WIKI: (&str, &str) = ("smol-wiki-articles", "csv");
+const DATASET_WIKI_1_2: (&str, &str) = ("smol-wiki-articles-1_2", "csv");
+const DATASET_WIKI_3_4: (&str, &str) = ("smol-wiki-articles-3_4", "csv");
+const DATASET_WIKI_4_4: (&str, &str) = ("smol-wiki-articles-4_4", "csv");
 const DATASET_MOVIES: (&str, &str) = ("movies", "json");
 const DATASET_GEO: (&str, &str) = ("smol-all-countries", "jsonl");
 
@@ -24,6 +27,9 @@ const ALL_DATASETS: &[(&str, &str)] = &[
     DATASET_SONGS_3_4,
     DATASET_SONGS_4_4,
     DATASET_WIKI,
+    DATASET_WIKI_1_2,
+    DATASET_WIKI_3_4,
+    DATASET_WIKI_4_4,
     DATASET_MOVIES,
     DATASET_GEO,
 ];