Method to cache critical results

[Chan-Dong-Jun]

What's the problem this feature will solve?
There should be a method to cache the critical results.

Describe the solution you'd like
Introduce a method to cache the results at critical steps and write them to a dataframe before caching them.

[Chan-Dong-Jun]

Tutorial
https://github.com/Chan-Dong-Jun/mesa-forked/blob/3.0a1-cache-critical-results/mesa/Tutorial%20for%20Cacheable%20Model.ipynb
Source Code
https://github.com/Chan-Dong-Jun/mesa-forked/blob/3.0a1-cache-critical-results/mesa/cacheable_model.py

[Chan-Dong-Jun]

Main edit

 def save_critical_result(self, condition_function: Callable[[dict], bool]):
        model_vars = self.model.datacollector.model_vars
        self.cache_file_path.mkdir(parents=True, exist_ok=True)

        current_step = self.model._steps
        special_results_file = f"{self.cache_file_path}/special_results.parquet"
        if condition_function(model_vars):
            step_data = {key: [value[-1]] for key, value in model_vars.items()}
            step_data["Step"] = current_step
            special_results_df = pd.DataFrame(step_data)

            # Append the current step data to the Parquet file
            if os.path.exists(special_results_file):
                existing_data = pq.read_table(special_results_file).to_pandas()
                combined_data = pd.concat(
                    [existing_data, special_results_df], ignore_index=True
                )
                special_results_table = pa.Table.from_pandas(combined_data)
            else:
                special_results_table = pa.Table.from_pandas(special_results_df)

            pq.write_table(special_results_table, special_results_file)

condition_function (Callable[[dict], bool]): A function that takes a dictionary of model variables and returns a boolean indicating whether the current step's data should be saved.
save_critical_result takes in a condition function that caches model level results whenever the condition in the function is met.

[Chan-Dong-Jun]

Example