From a49d45d93929e26c4a02e758e2c86ddac523d731 Mon Sep 17 00:00:00 2001
From: Andrea Soria <andrea@huggingface.co>
Date: Thu, 11 Jul 2024 11:56:18 -0400
Subject: [PATCH] Skip smart update when language is updated

---
 libs/libcommon/src/libcommon/constants.py    |  2 ++
 libs/libcommon/src/libcommon/orchestrator.py | 12 ++++++------
 2 files changed, 8 insertions(+), 6 deletions(-)

diff --git a/libs/libcommon/src/libcommon/constants.py b/libs/libcommon/src/libcommon/constants.py
index a38e61e504..6ebfd4f781 100644
--- a/libs/libcommon/src/libcommon/constants.py
+++ b/libs/libcommon/src/libcommon/constants.py
@@ -103,3 +103,5 @@
     5000.0,
     float("inf"),
 )
+
+YAML_FIELDS_TO_CHECK = ["dataset_info", "configs", "viewer", "language"]
diff --git a/libs/libcommon/src/libcommon/orchestrator.py b/libs/libcommon/src/libcommon/orchestrator.py
index 53fe0858c9..b5fef81d87 100644
--- a/libs/libcommon/src/libcommon/orchestrator.py
+++ b/libs/libcommon/src/libcommon/orchestrator.py
@@ -19,6 +19,7 @@
     DATASET_CONFIG_NAMES_KIND,
     DEFAULT_DIFFICULTY_MAX,
     DIFFICULTY_BONUS_BY_FAILED_RUNS,
+    YAML_FIELDS_TO_CHECK,
 )
 from libcommon.dtos import JobInfo, JobResult, Priority
 from libcommon.processing_graph import ProcessingGraph, ProcessingStep, ProcessingStepDoesNotExist, processing_graph
@@ -882,12 +883,11 @@ def __post_init__(self) -> None:
         }:  # TODO: maybe support .huggingface.yaml later
             raise SmartUpdateImpossibleBecauseOfUpdatedFiles(", ".join(self.files_impacted_by_commit)[:1000])
         self.updated_yaml_fields_in_dataset_card = self.get_updated_yaml_fields_in_dataset_card()
-        if "dataset_info" in self.updated_yaml_fields_in_dataset_card:
-            raise SmartUpdateImpossibleBecauseOfUpdatedYAMLField("dataset_info")
-        if "configs" in self.updated_yaml_fields_in_dataset_card:
-            raise SmartUpdateImpossibleBecauseOfUpdatedYAMLField("configs")
-        if "viewer" in self.updated_yaml_fields_in_dataset_card:
-            raise SmartUpdateImpossibleBecauseOfUpdatedYAMLField("viewer")
+
+        for yaml_field in YAML_FIELDS_TO_CHECK:
+            if yaml_field in self.updated_yaml_fields_in_dataset_card:
+                raise SmartUpdateImpossibleBecauseOfUpdatedYAMLField(yaml_field)
+
         # We update the cache entries and the storage (assets + cached assets)
         # We don't update the jobs because they might be creating artifacts that won't be updated by this code,
         # so we let them finish and restart later.