Added from_path to InitialEstimates

[cbl]

This pull request adds a from_path method to InitialEstimates to reduce complexity of loading the required files and doing necessary checks.

Before

data_df = pd.read_csv("hodgkins_disease.csv")
surv_0_df = pd.read_csv("initial_estimates/surv_0.csv")
surv_1_df = pd.read_csv("initial_estimates/surv_1.csv")
cens_surv_0_df = pd.read_csv("initial_estimates/cens_surv_0.csv")
cens_surv_1_df = pd.read_csv("initial_estimates/cens_surv_1.csv")
haz1_0_df = pd.read_csv("initial_estimates/haz1_0.csv")
haz1_1_df = pd.read_csv("initial_estimates/haz1_1.csv")
haz2_0_df = pd.read_csv("initial_estimates/haz2_0.csv")
haz2_1_df = pd.read_csv("initial_estimates/haz2_1.csv")
prop_0_df = pd.read_csv("initial_estimates/prop_0.csv", header=None)
prop_1_df = pd.read_csv("initial_estimates/prop_1.csv", header=None)

# List of dataframes to check
dataframes = [surv_0_df, surv_1_df, cens_surv_0_df, cens_surv_1_df, haz1_0_df, haz1_1_df, haz2_0_df, haz2_1_df]

# Check if all dataframes have the same columns
all_col_equal = all(df.index.equals(dataframes[0].index) for df in dataframes)

dataframes += [data_df, prop_0_df, prop_1_df]
# Check if all dataframes have the same index
all_indices_equal = all(df.index.equals(dataframes[0].index) for df in dataframes)


print("All dataframes have the same index and columns:", all_indices_equal and all_col_equal)

initial_estimates = {
    0: InitialEstimates(times = surv_0_df.columns.astype(float),
                        g_star_obs= 1 - data_df["chemo"].values,
                        propensity_scores=prop_0_df.values.squeeze(),
                        hazards=np.stack([haz1_0_df.values, haz2_0_df.values], axis=-1),
                        event_free_survival_function=surv_0_df.values,
                        censoring_survival_function=cens_surv_0_df.values),
    1: InitialEstimates(times = surv_1_df.columns.astype(float),
                        g_star_obs= data_df["chemo"].values,
                        propensity_scores=prop_1_df.values.squeeze(),
                        hazards=np.stack([haz1_1_df.values, haz2_1_df.values], axis=-1),
                        event_free_survival_function=surv_1_df.values,
                        censoring_survival_function=cens_surv_1_df.values),
}

After

data_df = pd.read_csv("hodgkins_disease.csv")
initial_estimates = InitialEstimates.from_path(data_df, 'initial_estimates')