Causal Estimation

To learn the causal relationship between the random variables in the causal graph, we use the dowhy to identify the causal effect, then use the EconML to estimate the causal effect. You can find the code for above pipeline in estimate.py. You can directly execute the script to estimate the causal effect.

Usage:

python estimate.py [-h] [--causal_graph CAUSAL_GRAPH]
                   [--collected_data COLLECTED_DATA]
                   [--workflow_data WORKFLOW_DATA]
                   [--output_dir OUTPUT_DIR]

Here is the arguments table for the script:

Argument	Description	Special Remark
causal_graph	Path to causal graph file (should be txt file).	`str`, causal graph file generated by `run_blip.py`.
collected_data	Path to collected data file (should be csv file).	`str`, the `combined.csv` generated by `collect_data.py`.
workflow_data	Path to Chaos Mesh workflow data file (should be csv file).	`str`, the Chaos Mesh workflow data generated by `parse_chaos.py`.
output_dir	Path to output directory.	`str`, It should have two subfolder, `deepiv/` and `dml/` to store different kind of models respectively.