From 980921e078f05b35c9907a38843d406b4851f95c Mon Sep 17 00:00:00 2001
From: Louis Dureuil <louis@meilisearch.com>
Date: Tue, 12 Nov 2024 16:31:22 +0100
Subject: [PATCH] Vector fixes

---
 crates/milli/src/update/new/extract/vectors/mod.rs | 7 ++-----
 crates/milli/src/update/new/vector_document.rs     | 8 +++++++-
 crates/milli/src/vector/ollama.rs                  | 2 +-
 crates/milli/src/vector/openai.rs                  | 2 +-
 crates/milli/src/vector/rest.rs                    | 2 +-
 5 files changed, 12 insertions(+), 9 deletions(-)

diff --git a/crates/milli/src/update/new/extract/vectors/mod.rs b/crates/milli/src/update/new/extract/vectors/mod.rs
index df8e2ed09..514791a65 100644
--- a/crates/milli/src/update/new/extract/vectors/mod.rs
+++ b/crates/milli/src/update/new/extract/vectors/mod.rs
@@ -115,11 +115,8 @@ impl<'a, 'extractor> Extractor<'extractor> for EmbeddingExtractor<'a> {
                             new_vectors.vectors_for_key(embedder_name).transpose()
                         }) {
                             let new_vectors = new_vectors?;
-                            match (old_vectors.regenerate, new_vectors.regenerate) {
-                                (true, true) | (false, false) => todo!(),
-                                _ => {
-                                    chunks.set_regenerate(update.docid(), new_vectors.regenerate);
-                                }
+                            if old_vectors.regenerate != new_vectors.regenerate {
+                                chunks.set_regenerate(update.docid(), new_vectors.regenerate);
                             }
                             // do we have set embeddings?
                             if let Some(embeddings) = new_vectors.embeddings {
diff --git a/crates/milli/src/update/new/vector_document.rs b/crates/milli/src/update/new/vector_document.rs
index e96e29053..381c4dab6 100644
--- a/crates/milli/src/update/new/vector_document.rs
+++ b/crates/milli/src/update/new/vector_document.rs
@@ -180,7 +180,13 @@ fn entry_from_raw_value(
         },
         RawVectors::ImplicitlyUserProvided(value) => VectorEntry {
             has_configured_embedder,
-            embeddings: value.map(Embeddings::FromJsonImplicityUserProvided),
+            // implicitly user provided always provide embeddings
+            // `None` here means that there are no embeddings
+            embeddings: Some(
+                value
+                    .map(Embeddings::FromJsonImplicityUserProvided)
+                    .unwrap_or(Embeddings::FromDb(Default::default())),
+            ),
             regenerate: false,
             implicit: true,
         },
diff --git a/crates/milli/src/vector/ollama.rs b/crates/milli/src/vector/ollama.rs
index 65fd05416..263d9d3c9 100644
--- a/crates/milli/src/vector/ollama.rs
+++ b/crates/milli/src/vector/ollama.rs
@@ -113,7 +113,7 @@ impl Embedder {
         threads
             .install(move || {
                 let embeddings: Result<Vec<Vec<Embedding>>, _> = texts
-                    .par_chunks(self.chunk_count_hint())
+                    .par_chunks(self.prompt_count_in_chunk_hint())
                     .map(move |chunk| self.embed(chunk))
                     .collect();
 
diff --git a/crates/milli/src/vector/openai.rs b/crates/milli/src/vector/openai.rs
index 466fd1660..375b2878a 100644
--- a/crates/milli/src/vector/openai.rs
+++ b/crates/milli/src/vector/openai.rs
@@ -266,7 +266,7 @@ impl Embedder {
         threads
             .install(move || {
                 let embeddings: Result<Vec<Vec<Embedding>>, _> = texts
-                    .par_chunks(self.chunk_count_hint())
+                    .par_chunks(self.prompt_count_in_chunk_hint())
                     .map(move |chunk| self.embed(chunk))
                     .collect();
 
diff --git a/crates/milli/src/vector/rest.rs b/crates/milli/src/vector/rest.rs
index dc2ab95f9..eeb5b16af 100644
--- a/crates/milli/src/vector/rest.rs
+++ b/crates/milli/src/vector/rest.rs
@@ -193,7 +193,7 @@ impl Embedder {
         threads
             .install(move || {
                 let embeddings: Result<Vec<Vec<Embedding>>, _> = texts
-                    .par_chunks(self.chunk_count_hint())
+                    .par_chunks(self.prompt_count_in_chunk_hint())
                     .map(move |chunk| self.embed_ref(chunk))
                     .collect();