From 5ccff4887c57997784a94ca1aa4eedcaf8a67fd4 Mon Sep 17 00:00:00 2001
From: "R. Ryantm" <ryantm-bot@ryantm.com>
Date: Mon, 20 Apr 2026 10:13:43 +0000
Subject: [PATCH 1/2] python3Packages.mlcroissant: 1.0.22 -> 1.1.0

---
 .../python-modules/mlcroissant/default.nix        | 15 ++++++++-------
 1 file changed, 8 insertions(+), 7 deletions(-)

diff --git a/pkgs/development/python-modules/mlcroissant/default.nix b/pkgs/development/python-modules/mlcroissant/default.nix
index b4512f183a16..04f1bdf2cd92 100644
--- a/pkgs/development/python-modules/mlcroissant/default.nix
+++ b/pkgs/development/python-modules/mlcroissant/default.nix
@@ -29,19 +29,20 @@
   writableTmpDirAsHomeHook,
 }:
 
-buildPythonPackage rec {
+buildPythonPackage (finalAttrs: {
   pname = "mlcroissant";
-  version = "1.0.22";
+  version = "1.1.0";
   pyproject = true;
+  __structuredAttrs = true;
 
   src = fetchFromGitHub {
     owner = "mlcommons";
     repo = "croissant";
-    tag = "v${version}";
-    hash = "sha256-uJOxKNrK3eN2wyPFEQr2J4+vZeSK1KPyFDag2jcyWZw=";
+    tag = "v${finalAttrs.version}";
+    hash = "sha256-IaRlmNQjOSIT3/b6AM68eRmweZEI5yjo6I9ievQIIsE=";
   };
 
-  sourceRoot = "${src.name}/python/mlcroissant";
+  sourceRoot = "${finalAttrs.src.name}/python/mlcroissant";
 
   build-system = [
     setuptools
@@ -91,10 +92,10 @@ buildPythonPackage rec {
   meta = {
     description = "High-level format for machine learning datasets that brings together four rich layers";
     homepage = "https://github.com/mlcommons/croissant";
-    changelog = "https://github.com/mlcommons/croissant/releases/tag/${src.tag}";
+    changelog = "https://github.com/mlcommons/croissant/releases/tag/${finalAttrs.src.tag}";
     license = lib.licenses.asl20;
     maintainers = with lib.maintainers; [ GaetanLepage ];
     platforms = lib.platforms.all;
     mainProgram = "mlcroissant";
   };
-}
+})

From 25f11f8eeccdc0c09a42fabe1133f24649c9f1d9 Mon Sep 17 00:00:00 2001
From: Gaetan Lepage <gaetan@glepage.com>
Date: Mon, 20 Apr 2026 12:16:48 +0000
Subject: [PATCH 2/2] python3Packages.tensorflow-datasets: fix mlcroissant
 1.1.0 compatibility

---
 .../python-modules/tensorflow-datasets/default.nix     | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/pkgs/development/python-modules/tensorflow-datasets/default.nix b/pkgs/development/python-modules/tensorflow-datasets/default.nix
index 6de4478928a9..5dde0d7682e1 100644
--- a/pkgs/development/python-modules/tensorflow-datasets/default.nix
+++ b/pkgs/development/python-modules/tensorflow-datasets/default.nix
@@ -93,6 +93,16 @@ buildPythonPackage (finalAttrs: {
           "elif field.label == field.LABEL_REPEATED:" \
           "elif hasattr(field_value, 'extend'):"
     ''
+    # mlcroissant 1.1.0 requires leaf fields to define `source` or `value`
+    + ''
+      substituteInPlace tensorflow_datasets/core/utils/croissant_utils_test.py \
+        --replace-fail \
+          "references=mlc.Source(field='splits/name')," \
+          "references=mlc.Source(field='splits/name'), source=mlc.Source(field='splits/name')," \
+        --replace-fail \
+          "references=mlc.Source(field='labels/label')," \
+          "references=mlc.Source(field='labels/label'), source=mlc.Source(field='labels/label'),"
+    ''
     # TypeError: only 0-dimensional arrays can be converted to Python scalars
     + ''
       substituteInPlace tensorflow_datasets/datasets/smallnorb/smallnorb_dataset_builder.py \