From 5185aa21b8a5b31bf7b9df61f3b43447e2221faa Mon Sep 17 00:00:00 2001
From: Louis Dureuil <louis@meilisearch.com>
Date: Fri, 8 Nov 2024 00:05:36 +0100
Subject: [PATCH] Know if your vectors are implicit when writing them back in
 documents + don't write empty _vectors

---
 crates/milli/src/update/new/document.rs        | 18 +++++++++++++-----
 crates/milli/src/update/new/vector_document.rs |  4 ++++
 2 files changed, 17 insertions(+), 5 deletions(-)

diff --git a/crates/milli/src/update/new/document.rs b/crates/milli/src/update/new/document.rs
index f43eb63e4..692277597 100644
--- a/crates/milli/src/update/new/document.rs
+++ b/crates/milli/src/update/new/document.rs
@@ -332,14 +332,22 @@ where
             }
             vectors.insert(
                 name,
-                serde_json::json!({
-                    "regenerate": entry.regenerate,
-                    // TODO: consider optimizing the shape of embedders here to store an array of f32 rather than a JSON object
-                    "embeddings": entry.embeddings,
-                }),
+                if entry.implicit {
+                    serde_json::json!(entry.embeddings)
+                } else {
+                    serde_json::json!({
+                        "regenerate": entry.regenerate,
+                        // TODO: consider optimizing the shape of embedders here to store an array of f32 rather than a JSON object
+                        "embeddings": entry.embeddings,
+                    })
+                },
             );
         }
 
+        if vectors.is_empty() {
+            break 'inject_vectors;
+        }
+
         vectors_value = serde_json::value::to_raw_value(&vectors).unwrap();
         unordered_field_buffer.push((vectors_fid, &vectors_value));
     }
diff --git a/crates/milli/src/update/new/vector_document.rs b/crates/milli/src/update/new/vector_document.rs
index 4a27361a9..e96e29053 100644
--- a/crates/milli/src/update/new/vector_document.rs
+++ b/crates/milli/src/update/new/vector_document.rs
@@ -71,6 +71,7 @@ pub struct VectorEntry<'doc> {
     pub has_configured_embedder: bool,
     pub embeddings: Option<Embeddings<'doc>>,
     pub regenerate: bool,
+    pub implicit: bool,
 }
 
 pub trait VectorDocument<'doc> {
@@ -125,6 +126,7 @@ impl<'t> VectorDocumentFromDb<'t> {
             has_configured_embedder: true,
             embeddings: Some(Embeddings::FromDb(vectors)),
             regenerate: !config.user_provided.contains(self.docid),
+            implicit: false,
         })
     }
 }
@@ -174,11 +176,13 @@ fn entry_from_raw_value(
             has_configured_embedder,
             embeddings: raw_explicit_vectors.embeddings.map(Embeddings::FromJsonExplicit),
             regenerate: raw_explicit_vectors.regenerate,
+            implicit: false,
         },
         RawVectors::ImplicitlyUserProvided(value) => VectorEntry {
             has_configured_embedder,
             embeddings: value.map(Embeddings::FromJsonImplicityUserProvided),
             regenerate: false,
+            implicit: true,
         },
     })
 }