From ccacdf212c5ca3d952ef92f4556f3415b0b0b907 Mon Sep 17 00:00:00 2001
From: Tamo <tamo@meilisearch.com>
Date: Mon, 20 Jan 2025 15:54:43 +0100
Subject: [PATCH] fix the field distribution when upgrading from the v1_12

---
 Cargo.lock                               |  1 +
 crates/milli/Cargo.toml                  |  1 +
 crates/milli/src/progress.rs             | 14 ++++++++----
 crates/milli/src/update/upgrade/mod.rs   | 27 ++++++++++++++--------
 crates/milli/src/update/upgrade/v1_12.rs | 29 ++++++++++++++++++++++++
 5 files changed, 59 insertions(+), 13 deletions(-)
 create mode 100644 crates/milli/src/update/upgrade/v1_12.rs

diff --git a/Cargo.lock b/Cargo.lock
index 3200d8333..32ae9bc58 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -3776,6 +3776,7 @@ dependencies = [
  "candle-transformers",
  "charabia",
  "concat-arrays",
+ "convert_case 0.6.0",
  "crossbeam-channel",
  "csv",
  "deserr",
diff --git a/crates/milli/Cargo.toml b/crates/milli/Cargo.toml
index d22829045..5eb89ea53 100644
--- a/crates/milli/Cargo.toml
+++ b/crates/milli/Cargo.toml
@@ -20,6 +20,7 @@ bytemuck = { version = "1.21.0", features = ["extern_crate_alloc"] }
 byteorder = "1.5.0"
 charabia = { version = "0.9.2", default-features = false }
 concat-arrays = "0.1.2"
+convert_case = "0.6.0"
 crossbeam-channel = "0.5.14"
 deserr = "0.6.3"
 either = { version = "1.13.0", features = ["serde"] }
diff --git a/crates/milli/src/progress.rs b/crates/milli/src/progress.rs
index 870277bad..3837e173a 100644
--- a/crates/milli/src/progress.rs
+++ b/crates/milli/src/progress.rs
@@ -89,19 +89,24 @@ impl<Name: NamedStep> Step for AtomicSubStep<Name> {
     }
 }
 
+#[doc(hidden)]
+pub use convert_case as _private_convert_case;
+#[doc(hidden)]
+pub use enum_iterator as _private_enum_iterator;
+
 #[macro_export]
 macro_rules! make_enum_progress {
     ($visibility:vis enum $name:ident { $($variant:ident,)+ }) => {
         #[repr(u8)]
-        #[derive(Debug, Clone, Copy, PartialEq, Eq, Sequence)]
+        #[derive(Debug, Clone, Copy, PartialEq, Eq, $crate::progress::_private_enum_iterator::Sequence)]
         #[allow(clippy::enum_variant_names)]
         $visibility enum $name {
             $($variant),+
         }
 
-        impl Step for $name {
-            fn name(&self) -> Cow<'static, str> {
-                use convert_case::Casing;
+        impl $crate::progress::Step for $name {
+            fn name(&self) -> std::borrow::Cow<'static, str> {
+                use $crate::progress::_private_convert_case::Casing;
 
                 match self {
                     $(
@@ -115,6 +120,7 @@ macro_rules! make_enum_progress {
             }
 
             fn total(&self) -> u32 {
+                use $crate::progress::_private_enum_iterator::Sequence;
                 Self::CARDINALITY as u32
             }
         }
diff --git a/crates/milli/src/update/upgrade/mod.rs b/crates/milli/src/update/upgrade/mod.rs
index 6e533177a..4be55e942 100644
--- a/crates/milli/src/update/upgrade/mod.rs
+++ b/crates/milli/src/update/upgrade/mod.rs
@@ -1,4 +1,7 @@
+mod v1_12;
+
 use heed::RwTxn;
+use v1_12::{v1_12_3_to_v1_13, v1_12_to_v1_12_3};
 
 use crate::progress::{Progress, VariableNameStep};
 use crate::{Index, InternalError, Result};
@@ -6,12 +9,21 @@ use crate::{Index, InternalError, Result};
 /// Return true if the cached stats of the index must be regenerated
 pub fn upgrade(wtxn: &mut RwTxn, index: &Index, progress: Progress) -> Result<bool> {
     let from = index.get_version(wtxn)?;
-    let upgrade_functions =
-        [(v1_12_to_v1_13 as fn(&Index, Progress) -> Result<()>, "Upgrading from v1.12 to v1.13")];
+    let upgrade_functions = [
+        (
+            v1_12_to_v1_12_3 as fn(&mut RwTxn, &Index, Progress) -> Result<bool>,
+            "Upgrading from v1.12.(0/1/2) to v1.12.3",
+        ),
+        (
+            v1_12_3_to_v1_13 as fn(&mut RwTxn, &Index, Progress) -> Result<bool>,
+            "Upgrading from v1.12.3+ to v1.13",
+        ),
+    ];
 
-    let (start, regenerate_stats) = match from {
+    let start = match from {
         // If there was no version it means we're coming from the v1.12
-        None | Some((1, 12, _)) => (0, false),
+        None | Some((1, 12, 0..=2)) => 0,
+        Some((1, 12, 3..)) => 1,
         // We must handle the current version in the match because in case of a failure some index may have been upgraded but not other.
         Some((1, 13, _)) => return Ok(false),
         Some((major, minor, patch)) => {
@@ -22,18 +34,15 @@ pub fn upgrade(wtxn: &mut RwTxn, index: &Index, progress: Progress) -> Result<bo
     enum UpgradeVersion {}
     let upgrade_path = &upgrade_functions[start..];
 
+    let mut regenerate_stats = false;
     for (i, (upgrade_function, upgrade_msg)) in upgrade_path.iter().enumerate() {
         progress.update_progress(VariableNameStep::<UpgradeVersion>::new(
             upgrade_msg.to_string(),
             i as u32,
             upgrade_path.len() as u32,
         ));
-        (upgrade_function)(index, progress.clone())?;
+        regenerate_stats |= (upgrade_function)(wtxn, index, progress.clone())?;
     }
 
     Ok(regenerate_stats)
 }
-
-fn v1_12_to_v1_13(_index: &Index, _progress: Progress) -> Result<()> {
-    Ok(())
-}
diff --git a/crates/milli/src/update/upgrade/v1_12.rs b/crates/milli/src/update/upgrade/v1_12.rs
new file mode 100644
index 000000000..4f76752d1
--- /dev/null
+++ b/crates/milli/src/update/upgrade/v1_12.rs
@@ -0,0 +1,29 @@
+use heed::RwTxn;
+
+use crate::{make_enum_progress, Result};
+
+use crate::{progress::Progress, Index};
+
+// The field distribution was not computed correctly in the v1.12 until the v1.12.3
+pub(super) fn v1_12_to_v1_12_3(
+    wtxn: &mut RwTxn,
+    index: &Index,
+    progress: Progress,
+) -> Result<bool> {
+    make_enum_progress! {
+        enum FieldDistribution {
+            RebuildingFieldDistribution,
+        }
+    };
+    progress.update_progress(FieldDistribution::RebuildingFieldDistribution);
+    crate::update::new::reindex::field_distribution(index, wtxn, &progress)?;
+    Ok(true)
+}
+
+pub(super) fn v1_12_3_to_v1_13(
+    _wtxn: &mut RwTxn,
+    _index: &Index,
+    _progress: Progress,
+) -> Result<bool> {
+    Ok(false)
+}