diffindiff 2.2.2__tar.gz → 2.2.4__tar.gz

{diffindiff-2.2.2 → diffindiff-2.2.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: diffindiff
-Version: 2.2.2
+Version: 2.2.4
 Summary: diffindiff: Python library for convenient Difference-in-Differences Analyses
 Author: Thomas Wieland
 Author-email: geowieland@googlemail.com
@@ -16,10 +16,11 @@ This Python library is designed for performing Difference-in-Differences (DiD) a
 Thomas Wieland [ORCID](https://orcid.org/0000-0001-5168-9846) [EMail](mailto:geowieland@googlemail.com) 
 
 
-## Updates v2.2.2
+## Updates v2.2.4
 - Bugfixes:
-  - Spillover treatment works now
-  - Verbose per default to False
+  - Spillover treatment really really works now (only relevant in rare cases)
+  - Fixed merging in diddata.DiffData.add_covariates() (only relevant in rare cases)
+  - Dropping missing values consequently (only relevant in rare cases)
 
 
 ## Features

{diffindiff-2.2.2 → diffindiff-2.2.4}/README.md

@@ -8,10 +8,11 @@ This Python library is designed for performing Difference-in-Differences (DiD) a
 Thomas Wieland [ORCID](https://orcid.org/0000-0001-5168-9846) [EMail](mailto:geowieland@googlemail.com) 
 
 
-## Updates v2.2.2
+## Updates v2.2.4
 - Bugfixes:
-  - Spillover treatment works now
-  - Verbose per default to False
+  - Spillover treatment really really works now (only relevant in rare cases)
+  - Fixed merging in diddata.DiffData.add_covariates() (only relevant in rare cases)
+  - Dropping missing values consequently (only relevant in rare cases)
 
 
 ## Features

{diffindiff-2.2.2 → diffindiff-2.2.4}/diffindiff/didanalysis.py

@@ -4,8 +4,8 @@
 # Author:      Thomas Wieland 
 #              ORCID: 0000-0001-5168-9846
 #              mail: geowieland@googlemail.com              
-# Version:     2.2.0
-# Last update: 2025-12-05 17:38
+# Version:     2.2.2
+# Last update: 2025-12-07 10:27
 # Copyright (c) 2025 Thomas Wieland
 #-----------------------------------------------------------------------
 
@@ -40,6 +40,9 @@ class DiffModel:
             timestamp
             ]    
 
+    def get_did_modeldata_df (self):
+        return pd.DataFrame(self.data[2])
+    
     def treatment_statistics(
         self,
         treatment: str = None,
@@ -890,7 +893,7 @@ class DiffModel:
         treatment_diagnostics = model_config["treatment_diagnostics"]
         no_treatments = model_config["no_treatments"]
         outcome_col = model_config["outcome_col"]
-        outcome_col_predicted = outcome_col+"_predicted"
+        outcome_col_predicted = f"{outcome_col}{config.PREDICTED_SUFFIX}"
 
         if TG_col is None and treatment is None:            
             if no_treatments == 1:
@@ -925,8 +928,7 @@ class DiffModel:
         if ("TG" in plot_intervals_groups and "CG" not in plot_intervals_groups) or ("CG" in plot_intervals_groups and "TG" not in plot_intervals_groups):
             lines_labels_required = lines_labels_required+1
         if "TG" in plot_intervals_groups and "CG" in plot_intervals_groups:
-            lines_labels_required = lines_labels_required+2
-            
+            lines_labels_required = lines_labels_required+2            
         assert len(lines_col) == lines_col_required, f"Parameter 'lines_col' must be a list with {lines_col_required} entries"
         assert len(lines_style) == lines_style_required, f"Parameter 'lines_style' must be a list with {lines_col_required} entries"
         assert len(lines_labels) == lines_labels_required, f"Parameter 'lines_labels' must be a list with {lines_labels_required} entries"
@@ -1389,6 +1391,13 @@ def did_analysis(
         *treatment_col        
         ]
     
+    data = tools.panel_index(
+        data = data,
+        unit_col = unit_col,
+        time_col = time_col,
+        verbose = verbose
+        )
+    
     treatment_diagnostics_results = helper.treatment_diagnostics(
         data = data,
         unit_col=unit_col,
@@ -1728,6 +1737,7 @@ def did_analysis(
         spillover = helper.create_spillover(
             data=data,
             unit_col=unit_col,
+            time_col=time_col,
             treatment_col=treatment_col,
             spillover_treatment=spillover_treatment,
             spillover_units=spillover_units
@@ -1925,7 +1935,14 @@ def ddd_analysis(
             )
           
         cols_relevant = cols_relevant + covariates
-   
+    
+    data = tools.panel_index(
+        data = data,
+        unit_col = unit_col,
+        time_col = time_col,
+        verbose = verbose
+        )
+    
     treatment_diagnostics_results = helper.treatment_diagnostics(
         data = data,
         unit_col=unit_col,

{diffindiff-2.2.2 → diffindiff-2.2.4}/diffindiff/didanalysis_helper.py

@@ -4,8 +4,8 @@
 # Author:      Thomas Wieland 
 #              ORCID: 0000-0001-5168-9846
 #              mail: geowieland@googlemail.com              
-# Version:     1.0.3
-# Last update: 2025-12-06 11:52
+# Version:     1.0.5
+# Last update: 2025-12-07 10:27
 # Copyright (c) 2025 Thomas Wieland
 #-----------------------------------------------------------------------
 
@@ -172,7 +172,9 @@ def create_specific_treatment_effects(
 def create_spillover(
     data: pd.DataFrame,
     unit_col: str,
+    time_col: str,
     treatment_col: list,
+    TT_col: str = None,
     spillover_treatment: list = [],
     spillover_units: list = [],
     verbose: bool = config.VERBOSE
@@ -189,16 +191,36 @@ def create_spillover(
     spillover_unit_vars = []
     spillover_treatment_vars = []
 
-    for i, treatment in enumerate(treatment_col):
+    for treatment in treatment_col:
+
+        if TT_col is None:
+            
+            TT_col = config.TT_COL
 
-        data[f"{config.SPILLOVER_UNIT_PREFIX}{config.DELIMITER}{treatment}"] = 0
-        data[f"{config.SPILLOVER_PREFIX}{config.DELIMITER}{treatment}"] = 0
+            data = tools.treatment_time_col(
+                data = data,
+                unit_col = unit_col,
+                time_col = time_col,
+                treatment_col = treatment,
+                create_TT_col = TT_col,
+                verbose = verbose
+                )[0]          
 
-        spillover_unit_vars.append(f"{config.SPILLOVER_UNIT_PREFIX}{config.DELIMITER}{treatment}")
-        spillover_treatment_vars.append(f"{config.SPILLOVER_PREFIX}{config.DELIMITER}{treatment}")
+        sp_unit_col = f"{config.SPILLOVER_UNIT_PREFIX}{config.DELIMITER}{treatment}"
+        sp_treatment_col = f"{config.SPILLOVER_PREFIX}{config.DELIMITER}{treatment}"
 
-        data.loc[data[unit_col].astype(str).isin(spillover_units), f"{config.SPILLOVER_UNIT_PREFIX}{config.DELIMITER}{treatment}"] = 1
-        data.loc[data[unit_col].astype(str).isin(spillover_units), f"{config.SPILLOVER_PREFIX}{config.DELIMITER}{treatment}"] = data.loc[data[unit_col].astype(str).isin(spillover_units), f"{config.SPILLOVER_UNIT_PREFIX}{config.DELIMITER}{treatment}"]*data.loc[data[unit_col].astype(str).isin(spillover_units), treatment]
+        data[sp_unit_col] = 0
+        data[sp_treatment_col] = 0
+
+        spillover_unit_vars.append(sp_unit_col)
+        spillover_treatment_vars.append(sp_treatment_col)
+
+        data.loc[
+            data[unit_col].astype(str).isin(spillover_units),
+            sp_unit_col
+            ] = 1
+        
+        data[sp_treatment_col] = data[sp_unit_col]*data[TT_col]
 
     spillover_treatment_vars_join = ' + '.join(spillover_treatment_vars)
 

{diffindiff-2.2.2 → diffindiff-2.2.4}/diffindiff/diddata.py

@@ -4,8 +4,8 @@
 # Author:      Thomas Wieland 
 #              ORCID: 0000-0001-5168-9846
 #              mail: geowieland@googlemail.com              
-# Version:     2.1.3
-# Last update: 2025-12-06 10:49
+# Version:     2.1.5
+# Last update: 2025-12-07 10:27
 # Copyright (c) 2025 Thomas Wieland
 #-----------------------------------------------------------------------
 
@@ -478,7 +478,7 @@ class DiffData:
         ):
         
         if unit_col is None and time_col is None:
-            raise ValueError("unit_col and/or time_col must be stated")
+            raise ValueError("Parameter 'unit_col' and/or 'time_col' must be stated")
         
         if verbose:
             if len(variables) > 0:
@@ -488,16 +488,16 @@ class DiffData:
 
         did_modeldata = self.get_did_modeldata_df()
         
+        additional_df = tools.panel_index(
+            data=additional_df,
+            unit_col=unit_col,
+            time_col=time_col,
+            verbose=verbose
+            )
+        
         existing_variables = []
         
-        if unit_col is not None and time_col is not None:
-
-            additional_df = tools.panel_index(
-                data=additional_df,
-                unit_col=unit_col,
-                time_col=time_col,
-                verbose=verbose
-                )
+        if unit_col is not None and time_col is not None:            
             
             if variables is None:
 
@@ -659,15 +659,15 @@ class DiffData:
         
         new_merge = tools.panel_index(
             data=new_merge,
-            unit_col=unit_id_col,
-            time_col=time_col,
+            unit_col=config.UNIT_COL,
+            time_col=config.TIME_COL,
             verbose=verbose
             )
 
         did_modeldata_old = tools.panel_index(
             data=did_modeldata_old,
-            unit_col=unit_id_col,
-            time_col=time_col,
+            unit_col=config.UNIT_COL,
+            time_col=config.TIME_COL,
             verbose=verbose
             )
 
@@ -1055,24 +1055,12 @@ def merge_data(
         treatment_data_df, 
         how = "cross"
         )
-    
-    if drop_missing or missing_replace_by_zero:
-        modeldata_ismissing = tools.is_missing(
-            data = did_modeldata, 
-            drop_missing = drop_missing,
-            missing_replace_by_zero = missing_replace_by_zero,
-            verbose = False
-            )
-        did_modeldata = modeldata_ismissing[2]
 
     did_modeldata[treatment_name] = did_modeldata[TG_col] * did_modeldata[TT_col]
 
     if treatment_config["after_treatment_period"]:
         did_modeldata[after_treatment_name] = did_modeldata[TG_col] * did_modeldata[ATT_col]
 
-    if np.dtype(did_modeldata[config.TIME_COL]) != np.dtype(outcome_data[time_col]):
-        print(f"WARNING: Time columns of treatment data and outcome data differ: {str(np.dtype(did_modeldata[config.TIME_COL]))}, {str(np.dtype(outcome_data[time_col]))}. This might induce an error while building the model dataset.")
-                
     did_modeldata = tools.panel_index(
         data=did_modeldata,
         unit_col=config.UNIT_COL,
@@ -1086,7 +1074,7 @@ def merge_data(
         time_col=time_col,
         verbose=verbose
         )
-      
+
     if keep_columns:
         outcome_data_short = outcome_data
     else:
@@ -1097,6 +1085,15 @@ def merge_data(
         on=config.UNIT_TIME_COL, 
         how="left"
         )
+    
+    if drop_missing or missing_replace_by_zero:
+        modeldata_ismissing = tools.is_missing(
+            data = did_modeldata, 
+            drop_missing = drop_missing,
+            missing_replace_by_zero = missing_replace_by_zero,
+            verbose = False
+            )
+        did_modeldata = modeldata_ismissing[2]
  
     outcome_col_original = outcome_col
     unit_time_col_original = unit_id_col, time_col
@@ -1230,7 +1227,7 @@ def create_counterfactual(
         unit_col = unit_col,
         time_col = time_col,
         treatment_col = treatment_col
-        )
+        )[0]
     units = tools.unique(units_tt[unit_col])  
     
     if not isnotreatment[0]:

{diffindiff-2.2.2 → diffindiff-2.2.4}/diffindiff/didtools.py

@@ -4,8 +4,8 @@
 # Author:      Thomas Wieland 
 #              ORCID: 0000-0001-5168-9846
 #              mail: geowieland@googlemail.com              
-# Version:     2.1.1
-# Last update: 2025-12-06 10:48
+# Version:     2.1.4
+# Last update: 2025-12-07 10:27
 # Copyright (c) 2025 Thomas Wieland
 #-----------------------------------------------------------------------
 
@@ -54,25 +54,39 @@ def panel_index(
     ):
 
     to_str = []
-    unit_x_time = True
 
-    if data[unit_col].dtype != 'object':
-        data[unit_col] = data[unit_col].astype(str)
-        to_str.append(unit_col)
+    if unit_col is not None:
+        if data[unit_col].dtype != 'object':
+            data[unit_col] = data[unit_col].astype(str)
+            to_str.append(unit_col)
+    else:
+        if verbose:
+            print("NOTE: No unit column was stated")
+
+    if time_col is not None:
+        if data[time_col].dtype != 'object':
+            data[time_col] = data[time_col].astype(str)
+            to_str.append(time_col)
+    else:
+        if verbose:
+            print("NOTE: No time column was stated")
 
-    if data[time_col].dtype != 'object':
-        data[time_col] = data[time_col].astype(str)
-        to_str.append(time_col)
+    if verbose and len(to_str) > 0:
+        print(f"NOTE: The following columns were converted to str: {', '.join(to_str)}.")
 
     if config.UNIT_TIME_COL not in data.columns:
-        unit_x_time = False
-        data[config.UNIT_TIME_COL] = data[unit_col]+config.DELIMITER+data[time_col]
+        
+        if unit_col is not None and time_col is not None:
 
-    if verbose:
-        if len(to_str) > 0:
-            print(f"NOTE: The following columns were converted to str: {', '.join(to_str)}.")
-        if not unit_x_time:
-            print(f"NOTE: The following unit-time-index column was included: {config.UNIT_TIME_COL}.")
+            data[config.UNIT_TIME_COL] = data[unit_col]+config.DELIMITER+data[time_col]
+            
+            if verbose:
+                print(f"NOTE: The following unit-time-index column was created: {config.UNIT_TIME_COL}.")
+
+        else:
+
+            if verbose:
+                print("No unit-time-index column was created.")
 
     return data
 
@@ -170,8 +184,8 @@ def is_binary(
     if verbose:
         print("OK")
 
-    if not binary:
-        print(f"NOTE: treatment column '{treatment_col}' is not binary. Likely treatment format is: {treatment_format}.")
+        if not binary:
+            print(f"NOTE: treatment column '{treatment_col}' is not binary. Likely treatment format is: {treatment_format}.")
     
     return [
         binary,
@@ -268,8 +282,8 @@ def is_simultaneous(
     if verbose:
         print("OK")
 
-    if not simultaneous and data_isnotreatment[0]:
-        print(f"NOTE: treatment '{treatment_col}' is not simultaneous.")
+        if not simultaneous and data_isnotreatment[0]:
+            print(f"NOTE: treatment '{treatment_col}' is not simultaneous.")
 
     if simultaneous and not data_isnotreatment[0]:
         print(f"WARNING: treatment '{treatment_col}' is simultaneous and does not include a {config.NO_TREATMENT_CG_DESCRIPTION}")
@@ -303,8 +317,8 @@ def is_notreatment(
     if verbose:
         print("OK")
 
-    if not no_treatment:
-        print(f"NOTE: treatment '{treatment_col}' does not include a {config.NO_TREATMENT_CG_DESCRIPTION}.")
+        if not no_treatment:
+            print(f"NOTE: treatment '{treatment_col}' does not include a {config.NO_TREATMENT_CG_DESCRIPTION}.")
 
     return [
         no_treatment, 
@@ -342,8 +356,8 @@ def treatment_group_col(
     if verbose:
         print("OK")
 
-    if create_TG_col_exists:
-        print(f"NOTE: Column {create_TG_col} already exists. Saved treatment group in column {config.TG_COL}{config.DELIMITER}{treatment_col}.")
+        if create_TG_col_exists:
+            print(f"NOTE: Column {create_TG_col} already exists. Saved treatment group in column {config.TG_COL}{config.DELIMITER}{treatment_col}.")
 
     return [
         data, 
@@ -351,6 +365,32 @@ def treatment_group_col(
         create_TG_col
         ]
 
+def treatment_time_col(
+    data: pd.DataFrame,
+    unit_col: str,
+    time_col: str,
+    treatment_col: str,
+    create_TT_col: str = "TT",
+    verbose: bool = config.VERBOSE
+    ):
+
+    tt = treatment_times(
+        data = data,
+        unit_col = unit_col,
+        time_col = time_col,
+        treatment_col = treatment_col,
+        verbose = verbose
+        )[1]
+    
+    data[create_TT_col] = 0
+    data.loc[data[time_col].isin(tt), create_TT_col] = 1
+
+    return [
+        data,
+        tt,
+        create_TT_col
+    ]          
+
 def untreated_units(
     data: pd.DataFrame,
     unit_col: str,
@@ -460,10 +500,11 @@ def is_prepost(
     if verbose:
         print("OK")
     
-    if prepost:
-        print("NOTE: Panel data is pre-post.")
-    else:
-        print("NOTE: Panel data is multi-period panel data.")
+    if verbose:
+        if prepost:
+            print("NOTE: Panel data is pre-post.")
+        else:
+            print("NOTE: Panel data is multi-period panel data.")
             
     return prepost
 
@@ -502,8 +543,8 @@ def is_multiple_treatment_period(
     if verbose:
         print("OK")
     
-    if units_multiple > 0:
-        print(f"NOTE: There are {units_multiple} observational units with multiple treatment periods with respect to treatment '{treatment_col}'.")
+        if units_multiple > 0:
+            print(f"NOTE: There are {units_multiple} observational units with multiple treatment periods with respect to treatment '{treatment_col}'.")
 
     return [
         multiple_treatment_period,
@@ -591,12 +632,22 @@ def treatment_times(
             unit_col, 
             time_col, 
             treatment_col
-            ]
+            ],
+        verbose=verbose
         )
     
+    is_multiple_treatment_period(
+        data = data,
+        unit_col = unit_col,
+        treatment_col = treatment_col,
+        verbose = verbose
+        )[0]
+    
     if verbose:
         print(f"Identifying treatment times for treatment '{treatment_col}'", end = " ... ")
     
+    tt = list(unique(data.loc[data[treatment_col] == 1, time_col]))
+
     units = unique(data[unit_col])
     
     units_tt = pd.DataFrame(columns = [unit_col, "treatment_min", "treatment_max"])
@@ -628,7 +679,10 @@ def treatment_times(
     if verbose:
         print("OK")
         
-    return units_tt
+    return [
+        units_tt,
+        tt
+    ]
 
 def model_wrapper(
     y,
@@ -833,8 +887,6 @@ def fit_metrics(
         RSQ_ADJ = (1-(1-RSQ)*((observations-1)/(observations-indep_vars_no-1)))
         
     else:
-        
-        print("NOTE: As no number of independent vars was stated, no Adj. R-Squared is calculated.")
             
         RSQ_ADJ = np.nan        
 
@@ -854,8 +906,13 @@ def fit_metrics(
     if verbose:
         print("OK")
     
-    if len(obs_exp_clean) < len(observed) or len(obs_exp_clean) < len(expected):
-        print("NOTE: Vectors 'observed' and/or 'expected' contain NaNs which were dropped.")
+    if verbose:
+        
+        if RSQ_ADJ == np.nan:
+            print("NOTE: As no number of independent vars was stated, no Adj. R-Squared is calculated.")
+
+        if len(obs_exp_clean) < len(observed) or len(obs_exp_clean) < len(expected):
+            print("NOTE: Vectors 'observed' and/or 'expected' contain NaNs which were dropped.")
             
     modelfit_results = [
         model_residuals,

{diffindiff-2.2.2 → diffindiff-2.2.4}/diffindiff.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: diffindiff
-Version: 2.2.2
+Version: 2.2.4
 Summary: diffindiff: Python library for convenient Difference-in-Differences Analyses
 Author: Thomas Wieland
 Author-email: geowieland@googlemail.com
@@ -16,10 +16,11 @@ This Python library is designed for performing Difference-in-Differences (DiD) a
 Thomas Wieland [ORCID](https://orcid.org/0000-0001-5168-9846) [EMail](mailto:geowieland@googlemail.com) 
 
 
-## Updates v2.2.2
+## Updates v2.2.4
 - Bugfixes:
-  - Spillover treatment works now
-  - Verbose per default to False
+  - Spillover treatment really really works now (only relevant in rare cases)
+  - Fixed merging in diddata.DiffData.add_covariates() (only relevant in rare cases)
+  - Dropping missing values consequently (only relevant in rare cases)
 
 
 ## Features

{diffindiff-2.2.2 → diffindiff-2.2.4}/setup.py

@@ -8,7 +8,7 @@ def read_README():
     
 setup(
     name='diffindiff',
-    version='2.2.2',
+    version='2.2.4',
     description='diffindiff: Python library for convenient Difference-in-Differences Analyses',
     packages=find_packages(include=["diffindiff", "diffindiff.tests"]),
     include_package_data=True,