Upload datasets/task_metadata.json with huggingface_hub

datasets/task_metadata.json

@@ -53,14 +53,14 @@
   "expert": {
     "task_id": "expert_sales_pipeline",
     "difficulty": "expert",
-    "description": "Clean a sales dataset by running a complete expert pipeline: 1) Remove duplicate transactions, 2) Standardize region values to Title Case, 3) Fill missing values correctly, 4) Fix all data types, 5) Remove outliers in sale_amount, 6) Rename all columns to snake_case, 7) Validate final schema. This task requires correct ORDER of operations. Call finish when done.",
+    "description": "Clean a sales dataset by running expert pipeline in ORDER: 1) Remove duplicate transactions, 2) Rename columns to snake_case, 3) Fill missing values using mode strategy, 4) Fix all data types, 5) Remove outliers using IQR method, 6) Validate final schema. Call finish when done.",
     "max_steps": 25,
     "operations_allowed": [
       "remove_duplicates",
+      "rename_columns",
       "fill_missing",
       "fix_dtype",
       "remove_outliers",
-      "rename_columns",
       "validate_schema",
       "finish"
     ],