chore.

examples/mi/machine_learning.py

@@ -2,29 +2,30 @@ import fire
 
 from metagpt.roles.mi.interpreter import Interpreter
 
-DATA_DIR = "examples/mi/data"
 WINE_REQ = "Run data analysis on sklearn Wine recognition dataset, include a plot, and train a model to predict wine class (20% as validation), and show validation accuracy."
 
-# sales_forecast data from https://www.kaggle.com/datasets/aslanahmedov/walmart-sales-forecast/data,
-# new_train, new_test from train.csv.
+# DATA_DIR = "your/path/to/data"
+DATA_DIR = "examples/mi/data/WalmartSalesForecast2"
+# sales_forecast data from https://www.kaggle.com/datasets/aslanahmedov/walmart-sales-forecast/data
 SALES_FORECAST_REQ = f"""
 # Goal
-Use time series regression machine learning to make predictions for Dept sales of the stores as accurate as possible.
+Train a model to predict sales for each department in every store (split the last 40 weeks records as validation dataset,
+the others is train dataset), include plot sales trends, holiday effects, distribution of sales across stores/departments,
+using box on the train dataset, print metric and plot scatter plots of groud truth and predictions on validation data.
+save predictions on test data.
 
 # Datasets Available
-- train_data: {DATA_DIR}/WalmartSalesForecast/new_train.csv
-- test_data: {DATA_DIR}/WalmartSalesForecast/new_test.csv
-- additional data: {DATA_DIR}/WalmartSalesForecast/features.csv; To merge on train, test data.
-- stores data: {DATA_DIR}/WalmartSalesForecast/stores.csv; To merge on train, test data.
+- train_data: {DATA_DIR}/train.csv
+- test_data: {DATA_DIR}/test.csv, no label data.
+- additional data: {DATA_DIR}/features.csv
+- stores data: {DATA_DIR}/stores.csv
 
 # Metric
 The metric of the competition is weighted mean absolute error (WMAE) for test data.
 
 # Notice
 - *print* key variables to get more information for next task step.
-- Perform data analysis by plotting sales trends, holiday effects, distribution of sales across stores/departments using box/violin on the train data.
-- Make sure the DataFrame.dtypes must be int, float or bool, and drop date column.
-- Plot scatter plots of  groud truth and predictions on test data.
+- Only When you fit the model, make the DataFrame.dtypes to be int, float or bool, and drop date column.
 """
 
 requirements = {"wine": WINE_REQ, "sales_forecast": SALES_FORECAST_REQ}
@@ -32,7 +33,12 @@ requirements = {"wine": WINE_REQ, "sales_forecast": SALES_FORECAST_REQ}
 
 async def main(auto_run: bool = True, use_case: str = "wine"):
     mi = Interpreter(auto_run=auto_run)
-    await mi.run(requirements[use_case])
+    if use_case == "wine":
+        requirement = requirements[use_case]
+    else:
+        assert DATA_DIR != "your/path/to/data", f"Please set DATA_DIR for the use_case: {use_case}!"
+        requirement = requirements[use_case]
+    await mi.run(requirement)
 
 
 if __name__ == "__main__":