diffindiff 2.2.5__tar.gz → 2.2.7__tar.gz

{diffindiff-2.2.5 → diffindiff-2.2.7}/PKG-INFO

@@ -1,12 +1,12 @@
 Metadata-Version: 2.1
 Name: diffindiff
-Version: 2.2.5
-Summary: diffindiff: Python library for convenient Difference-in-Differences Analyses
+Version: 2.2.7
+Summary: diffindiff: Python library for convenient Difference-in-Differences analyses
 Author: Thomas Wieland
 Author-email: geowieland@googlemail.com
 Description-Content-Type: text/markdown
 
-# diffindiff: A Python library for convenient difference-in-differences analyses
+# diffindiff: Python library for convenient Difference-in-Differences analyses
 
 This Python library is designed for performing Difference-in-Differences (DiD) analyses in a convenient way. It allows users to construct datasets, define treatment and control groups, and set treatment periods. DiD model analyses may be conducted with both datasets created by built-in functions and ready-to-use external datasets. Both simultaneous and staggered adoption are supported. The library allows for various extensions, such as two-way fixed effects models, group- or individual-specific effects, post-treatment periods, and triple-difference estimations. Additionally, it includes functions for visualizing results, such as plotting DiD coefficients with confidence intervals and illustrating the temporal evolution of staggered treatments. Furthermore, several functions for rigorous treatment setting and data diagnostics are incorporated.
 
@@ -20,14 +20,14 @@ Thomas Wieland [ORCID](https://orcid.org/0000-0001-5168-9846) [EMail](mailto:geo
 
 - 📦 PyPI: [diffindiff](https://pypi.org/project/diffindiff/)
 - 💻 GitHub Repository: [diffindiff_official](https://github.com/geowieland/diffindiff_official)
-- 📄 DOI (Zenodo): [10.5281/zenodo.18639559](https://doi.org/10.5281/zenodo.18639559)
+- 📄 DOI (Zenodo): [10.5281/zenodo.18656820](https://doi.org/10.5281/zenodo.18656820)
 
 
 ## Citation
 
 If you use this software, please cite:
 
-Wieland, T. (2026). diffindiff: A Python library for convenient difference-in-differences analyses (Version 2.2.4) [Computer software]. Zenodo. https://doi.org/10.5281/zenodo.18656820
+Wieland, T. (2026). diffindiff: A Python library for convenient difference-in-differences analyses (Version 2.2.7) [Computer software]. Zenodo. https://doi.org/10.5281/zenodo.18656820
 
 
 ## Installation
@@ -167,8 +167,9 @@ See the /tests directory for usage examples of most of the included functions.
   - Wooldridge JM (2012) *Introductory Econometrics. A Modern Approach*.
 
 
-## What's new (v2.2.5)
+## What's new (v2.2.7)
+- Functions
+  - diddata.DiffData.define_treatment() for constructing a new treatment from a column in the dataframe 
 - Bugfixes:
-  - Incorrect import
-- Other:
-  - Update README
+  - didtools.treatment_times() and didtools.is_multiple_treatment_period() now also identify continuous treatments correctly
+  - Fixed problematic type conversion in didtools.fit_metrics()

{diffindiff-2.2.5 → diffindiff-2.2.7}/README.md

@@ -1,4 +1,4 @@
-# diffindiff: A Python library for convenient difference-in-differences analyses
+# diffindiff: Python library for convenient Difference-in-Differences analyses
 
 This Python library is designed for performing Difference-in-Differences (DiD) analyses in a convenient way. It allows users to construct datasets, define treatment and control groups, and set treatment periods. DiD model analyses may be conducted with both datasets created by built-in functions and ready-to-use external datasets. Both simultaneous and staggered adoption are supported. The library allows for various extensions, such as two-way fixed effects models, group- or individual-specific effects, post-treatment periods, and triple-difference estimations. Additionally, it includes functions for visualizing results, such as plotting DiD coefficients with confidence intervals and illustrating the temporal evolution of staggered treatments. Furthermore, several functions for rigorous treatment setting and data diagnostics are incorporated.
 
@@ -12,14 +12,14 @@ Thomas Wieland [ORCID](https://orcid.org/0000-0001-5168-9846) [EMail](mailto:geo
 
 - 📦 PyPI: [diffindiff](https://pypi.org/project/diffindiff/)
 - 💻 GitHub Repository: [diffindiff_official](https://github.com/geowieland/diffindiff_official)
-- 📄 DOI (Zenodo): [10.5281/zenodo.18639559](https://doi.org/10.5281/zenodo.18639559)
+- 📄 DOI (Zenodo): [10.5281/zenodo.18656820](https://doi.org/10.5281/zenodo.18656820)
 
 
 ## Citation
 
 If you use this software, please cite:
 
-Wieland, T. (2026). diffindiff: A Python library for convenient difference-in-differences analyses (Version 2.2.4) [Computer software]. Zenodo. https://doi.org/10.5281/zenodo.18656820
+Wieland, T. (2026). diffindiff: A Python library for convenient difference-in-differences analyses (Version 2.2.7) [Computer software]. Zenodo. https://doi.org/10.5281/zenodo.18656820
 
 
 ## Installation
@@ -159,8 +159,9 @@ See the /tests directory for usage examples of most of the included functions.
   - Wooldridge JM (2012) *Introductory Econometrics. A Modern Approach*.
 
 
-## What's new (v2.2.5)
+## What's new (v2.2.7)
+- Functions
+  - diddata.DiffData.define_treatment() for constructing a new treatment from a column in the dataframe 
 - Bugfixes:
-  - Incorrect import
-- Other:
-  - Update README
+  - didtools.treatment_times() and didtools.is_multiple_treatment_period() now also identify continuous treatments correctly
+  - Fixed problematic type conversion in didtools.fit_metrics()

{diffindiff-2.2.5 → diffindiff-2.2.7}/diffindiff/config.py

@@ -4,22 +4,25 @@
 # Author:      Thomas Wieland 
 #              ORCID: 0000-0001-5168-9846
 #              mail: geowieland@googlemail.com              
-# Version:     1.0.4
-# Last update: 2025-12-06 11:52
-# Copyright (c) 2025 Thomas Wieland
+# Version:     1.0.6
+# Last update: 2026-02-26 18:04
+# Copyright (c) 2025-2026 Thomas Wieland
 #-----------------------------------------------------------------------
 
 # Basic config:
 
-PACKAGE_VERSION = "2.2.2"
+PACKAGE_NAME = "diffindiff"
+PACKAGE_VERSION = "2.2.7"
 
-VERBOSE = False
+VERBOSE = True
 
 ROUND_STATISTIC = 3
 ROUND_PERCENT = 2
 
 AUTO_SWITCH_TO_PREPOST = True
 
+ACCEPT_CONTINUOUS_TREATMENTS = True
+
 # Description texts:
 
 DID_DESCRIPTION = "Difference-in-Differences Analysis"

{diffindiff-2.2.5 → diffindiff-2.2.7}/diffindiff/didanalysis.py

@@ -4,15 +4,14 @@
 # Author:      Thomas Wieland 
 #              ORCID: 0000-0001-5168-9846
 #              mail: geowieland@googlemail.com              
-# Version:     2.2.2
-# Last update: 2025-12-07 10:27
-# Copyright (c) 2025 Thomas Wieland
+# Version:     2.2.4
+# Last update: 2026-02-26 18:04
+# Copyright (c) 2024-2026 Thomas Wieland
 #-----------------------------------------------------------------------
 
 
 import pandas as pd
 import numpy as np
-from math import isnan
 import matplotlib.pyplot as plt
 from matplotlib.dates import DateFormatter
 import diffindiff.didtools as tools
@@ -930,7 +929,7 @@ class DiffModel:
         if "TG" in plot_intervals_groups and "CG" in plot_intervals_groups:
             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_style) == lines_style_required, f"Parameter 'lines_style' must be a list with {lines_style_required} entries"
         assert len(lines_labels) == lines_labels_required, f"Parameter 'lines_labels' must be a list with {lines_labels_required} entries"
             
         model_data = self.data[2]
@@ -1357,8 +1356,8 @@ def did_analysis(
     missing_replace_by_zero: bool = False,
     fit_by = "ols_fit",
     verbose: bool = config.VERBOSE
-    ):
-    
+    ):    
+   
     tools.check_columns(
         df = data,
         columns = [
@@ -1385,6 +1384,12 @@ def did_analysis(
         verbose = verbose
         )
     
+    tools.is_numeric(
+        df = data,
+        columns = treatment_col,
+        verbose = verbose
+        )
+    
     cols_relevant = [
         unit_col,
         time_col,
@@ -1808,7 +1813,7 @@ def did_analysis(
         }
     
     if bonferroni:
-        confint_alpha = confint_alpha/no_treatments   
+        confint_alpha = confint_alpha/no_treatments
 
     if fit_by == "ml":
         fit_result = helper.ml_fit(
@@ -1825,7 +1830,7 @@ def did_analysis(
             cluster_SE_by = cluster_SE_by,
             verbose = verbose
         )
-    
+           
     model_results = helper.extract_model_results(
         fit_result = fit_result,
         TG_col = TG_col,

{diffindiff-2.2.5 → diffindiff-2.2.7}/diffindiff/didanalysis_helper.py

@@ -4,9 +4,9 @@
 # Author:      Thomas Wieland 
 #              ORCID: 0000-0001-5168-9846
 #              mail: geowieland@googlemail.com              
-# Version:     1.0.5
-# Last update: 2025-12-07 10:27
-# Copyright (c) 2025 Thomas Wieland
+# Version:     1.0.7
+# Last update: 2025-02-26 18:02
+# Copyright (c) 2025-2026 Thomas Wieland
 #-----------------------------------------------------------------------
 
 import pandas as pd
@@ -203,7 +203,7 @@ def create_spillover(
                 time_col = time_col,
                 treatment_col = treatment,
                 create_TT_col = TT_col,
-                verbose = verbose
+                verbose = False
                 )[0]          
 
         sp_unit_col = f"{config.SPILLOVER_UNIT_PREFIX}{config.DELIMITER}{treatment}"
@@ -396,7 +396,11 @@ def treatment_diagnostics(
         )
     
     if verbose:
-        print(f"There are {no_treatments} treatments (simultaneous: {no_treatments-staggered_count}, staggered: {staggered_count}) with {untreated[0]} treated and {untreated[1]} untreated units.")
+        
+        if no_treatments > 1:
+            print(f"There are {no_treatments} treatments (simultaneous: {no_treatments-staggered_count}, staggered: {staggered_count}) with {untreated[0]} treated and {untreated[1]} untreated units.")
+        else:
+            print(f"There is {no_treatments} treatment (staggered: {staggered_count}) with {untreated[0]} treated and {untreated[1]} untreated units.")
 
     return [
         treatment_diagnostics_results,
@@ -918,10 +922,9 @@ def create_timestamp(function):
     now = datetime.now()
 
     timestamp_dict = {
-        "package_version": f"diffindiff {config.PACKAGE_VERSION}",
+        "package_version": f"{config.PACKAGE_NAME} {config.PACKAGE_VERSION}",
         "function": function,
         "datetime": now.strftime("%Y-%m-%d %H-%M-%S")
     }
 
-    return timestamp_dict
-
+    return timestamp_dict

{diffindiff-2.2.5 → diffindiff-2.2.7}/diffindiff/diddata.py

@@ -4,9 +4,9 @@
 # Author:      Thomas Wieland 
 #              ORCID: 0000-0001-5168-9846
 #              mail: geowieland@googlemail.com              
-# Version:     2.1.5
-# Last update: 2025-12-07 10:27
-# Copyright (c) 2025 Thomas Wieland
+# Version:     2.1.8
+# Last update: 2026-02-26 18:30
+# Copyright (c) 2024-2026 Thomas Wieland
 #-----------------------------------------------------------------------
 
 
@@ -76,29 +76,34 @@ class DiffGroups:
         verbose: bool = config.VERBOSE       
         ):
 
-        groups_config = self.data[1]
+        groups_config = self.data[1]        
         
         if groups_config["DDD"]:
-            raise ValueError("DiffGroups object already includes a benefit group")
-
-        if verbose:
-            print(f"Adding benefit group with {len(group_benefit)} units to groups data", end = " ... ")
+            
+            print("DiffGroups object already includes a benefit group. No segmentation added.")
+            
+            groups = self
+            
+        else:
 
-        groups_data = self.data[0]        
+            if verbose:
+                print(f"Adding benefit group with {len(group_benefit)} units to groups data", end = " ... ")                
+            
+            groups_data = self.data[0]
+            
+            groups_data[config.BG_COL] = 0
+            groups_data.loc[groups_data[config.UNIT_COL].astype(str).isin(group_benefit), config.BG_COL] = 1
+            
+            groups_config["DDD"] = True
 
-        groups_data[config.BG_COL] = 0
-        groups_data.loc[groups_data[config.UNIT_COL].astype(str).isin(group_benefit), config.BG_COL] = 1
-        
-        groups_config["DDD"] = True
+            groups = DiffGroups(
+                groups_data, 
+                groups_config,
+                timestamp = helper.create_timestamp(function="add_segmentation")
+                )
 
-        groups = DiffGroups(
-            groups_data, 
-            groups_config,
-            timestamp = helper.create_timestamp(function="add_segmentation")
-            )
-
-        if verbose:
-            print("OK")
+            if verbose:
+                print("OK")
 
         return groups
 
@@ -255,9 +260,16 @@ def create_treatment(
     after_treatment_period: bool = False,
     verbose = config.VERBOSE
     ): 
+       
+    check_dates = tools.check_date_format(
+        dates = study_period+treatment_period,
+        date_format = date_format
+    )
+    if check_dates[0]:
+        raise ValueError(f"Study and/or treatment period include invalid dates: {', '.join(check_dates[1])}.")
 
     TT_col = config.TT_COL
-
+    
     if treatment_name is not None:
         
         if not isinstance(treatment_name, str):
@@ -474,7 +486,7 @@ class DiffData:
         variables: list = None,
         unit_col: str = None,
         time_col: str = None,
-        verbose: bool = config.VERBOSE
+        verbose: bool = False
         ):
         
         if unit_col is None and time_col is None:
@@ -567,6 +579,7 @@ class DiffData:
         
         self.data[0] = did_modeldata
         self.data[5] = variables
+        self.data[7][len(self.data[7])] = helper.create_timestamp(function="add_covariates")
 
         if verbose:
             print("OK")
@@ -610,7 +623,6 @@ class DiffData:
         groups_data_old = did_groups_old.get_data()
 
         did_modeldata_old = self.get_did_modeldata_df()
-        unit_id_col, time_col = self.get_unit_time_cols()
         outcome_col_original = self.data[3]
         unit_time_col_original = self.get_unit_time_cols()
         covariates = self.get_covariates()
@@ -716,21 +728,157 @@ class DiffData:
             timestamp = helper.create_timestamp(function="add_treatment")
             )
 
-        did_data_new = DiffData(
-            did_modeldata = did_modeldata_new, 
-            diff_groups = groups_new, 
-            diff_treatment = treatment_new, 
-            outcome_col_original = outcome_col_original,
-            unit_time_col_original = unit_time_col_original,
-            covariates = covariates,
-            treatment_cols = treatment_cols_new,
-            timestamp = helper.create_timestamp(function="add_segmentation")
+        if verbose:
+            print("OK")
+
+        self.data[0] = did_modeldata_new
+        self.data[1] = groups_new
+        self.data[2] = treatment_new
+        self.data[3] = outcome_col_original
+        self.data[4] = unit_time_col_original
+        self.data[5] = covariates
+        self.data[6] = treatment_cols_new
+        self.data[7][len(self.data[7])] = helper.create_timestamp(function="add_treatment")
+        
+        return self
+
+    def define_treatment(
+        self,
+        treatment_name,
+        after_treatment_period: bool = False,
+        after_treatment_name = None,
+        verbose: bool = config.VERBOSE
+        ):
+
+        if not treatment_name:            
+            raise ValueError("When adding a treatment from the data, you need to specify a treatment column with parameter treament_name = [your_treatment].")
+        
+        if treatment_name not in self.get_did_modeldata_df().columns:
+            raise KeyError(f"Column '{treatment_name}' not in data frame")
+
+        did_treatment_old = self.get_did_treatment()
+        treatment_config_old = did_treatment_old.get_config()
+        treatment_meta_old = did_treatment_old.get_metadata()        
+        no_treatments_old = treatment_meta_old["no_treatments"]
+
+        did_groups_old = self.get_did_groups()
+        groups_config_old = did_groups_old.get_config()
+        groups_data_old = did_groups_old.get_data()
+
+        did_modeldata_old = self.get_did_modeldata_df()
+        outcome_col_original = self.data[3]
+        unit_time_col_original = self.get_unit_time_cols()
+        covariates = self.get_covariates()
+
+        treatment_cols = self.get_treatment_cols()
+        treatment_cols_new = treatment_cols
+
+        no_treatments = no_treatments_old+1
+        key_counter = no_treatments-1        
+        
+        tt = tools.treatment_times(
+            data = did_modeldata_old,
+            unit_col=config.UNIT_COL,
+            time_col=config.TIME_COL,
+            treatment_col=treatment_name,
+            verbose=verbose
+        )
+        
+        tt_date = [datetime.strptime(t, treatment_meta_old["date_format"]) for t in tt[1]]
+        treatment_period_start = min(tt_date)
+        treatment_period_end = max(tt_date)
+        treatment_period_start = treatment_period_start.strftime("%Y-%m-%d")
+        treatment_period_end = treatment_period_end.strftime("%Y-%m-%d")        
+        
+        is_notreatment_result = tools.is_notreatment(
+            data = did_modeldata_old,
+            unit_col=config.UNIT_COL,
+            treatment_col=treatment_name,
+            verbose = verbose 
             )
+        
+        treatment_group = is_notreatment_result[1]
+        control_group = is_notreatment_result[2]
+        
+        if verbose:
+            print(f"Constructing treatment from column '{treatment_name}'", end = " ... ")
+
+        new_groups = create_groups(
+            treatment_group = treatment_group, 
+            control_group = control_group,
+            treatment_name = treatment_name,
+            verbose=False
+            )
+        new_groups_data_df = new_groups.get_data()[0]        
+        new_groups_config = new_groups.get_config()
+        TG_col = new_groups_config[0]["TG_col"]
+
+        new_treatment = create_treatment(
+            study_period = [treatment_meta_old["study_period_start"], treatment_meta_old["study_period_end"]],
+            treatment_period = [treatment_period_start, treatment_period_end],
+            freq = treatment_meta_old["frequency"],
+            date_format = treatment_meta_old["date_format"],
+            treatment_name = treatment_name,
+            pre_post = treatment_meta_old["pre_post"],
+            after_treatment_period = after_treatment_period,
+            verbose=False
+            )
+        
+        new_treatment_data_df = new_treatment.get_data()
+        
+        new_treatment_config = new_treatment.get_config()        
+        TT_col = new_treatment_config[0]["TT_col"]
+        ATT_col = new_treatment_config[0]["ATT_col"]
+       
+        treatment_cols_new[key_counter] = {
+            "TT_col": TT_col, 
+            "ATT_col": ATT_col, 
+            "treatment_name": treatment_name, 
+            "after_treatment_name": after_treatment_name                
+            }
+        
+        groups_config_new = groups_config_old
+        groups_config_new[key_counter] = new_groups_config[0] 
+        groups_data_new = groups_data_old
+        groups_data_old.append(new_groups_data_df) 
+        groups_new = DiffGroups(
+            groups_data_new, 
+            groups_config_new,
+            timestamp = helper.create_timestamp(function="define_treatment")
+            )
+
+        treatment_meta_new = treatment_meta_old
+        treatment_meta_new["no_treatments"] = no_treatments
+        treatment_config_new = treatment_config_old
+        treatment_config_new[key_counter] = new_treatment_config[0]
+        
+        treatment_new = DiffTreatment(
+            new_treatment_data_df, 
+            treatment_config_new,
+            treatment_meta_new,
+            timestamp = helper.create_timestamp(function="define_treatment")
+            )       
 
         if verbose:
             print("OK")
-
-        return did_data_new
+            
+        if treatment_name in covariates:
+            
+            if verbose:
+                print(f"NOTE: Column '{treatment_name}' was defined as covariate before and is now removed from covariates list.")
+            
+            covariates.remove(treatment_name)
+
+        self.data[0] = did_modeldata_old
+        self.data[1] = groups_new
+        self.data[2] = treatment_new
+        self.data[3] = outcome_col_original
+        self.data[4] = unit_time_col_original
+        self.data[5] = covariates
+        self.data[6] = treatment_cols_new
+        self.data[7][len(self.data[7])] = helper.create_timestamp(function="define_treatment")        
+        
+        return self
 
     def add_segmentation(
         self,
@@ -967,8 +1115,8 @@ class DiffData:
                 if value["after_treatment_name"] is not None:
                     after_treatment_col[key] = value["after_treatment_name"]
                 if value["ATT_col"] is not None:
-                    ATT_col[key] = value["ATT_col"]
-                    
+                    ATT_col[key] = value["ATT_col"]            
+            
             did_results = didanalysis.did_analysis(
                 data = did_modeldata,
                 TG_col = TG_col,
@@ -1016,9 +1164,15 @@ def merge_data(
     keep_columns: bool = False,
     verbose: bool = config.VERBOSE
     ):
-
-    if verbose:
-        print("Merging groups and treatment data", end = " ... ")
+    
+    tools.check_columns(
+        df = outcome_data,
+        columns = [
+            unit_id_col, 
+            time_col, 
+            outcome_col
+            ]
+        )
 
     groups_data_df = diff_groups.get_data()
     groups_data_df = groups_data_df[0]
@@ -1075,6 +1229,9 @@ def merge_data(
         verbose=verbose
         )
 
+    if verbose:
+        print("Merging groups and treatment data", end = " ... ")
+
     if keep_columns:
         outcome_data_short = outcome_data
     else:
@@ -1108,7 +1265,8 @@ def merge_data(
             }
         }
 
-    timestamp = helper.create_timestamp(function="merge_data")
+    timestamp = {}
+    timestamp[0] = helper.create_timestamp(function="merge_data")
 
     did_data_all = DiffData(
         did_modeldata, 
@@ -1175,8 +1333,6 @@ def create_data(
         verbose = verbose
         )
 
-    did_data_all.timestamp = helper.create_timestamp(function="create_data")
-
     return did_data_all
 
 def create_counterfactual(

{diffindiff-2.2.5 → diffindiff-2.2.7}/diffindiff/didtools.py

@@ -4,15 +4,16 @@
 # Author:      Thomas Wieland 
 #              ORCID: 0000-0001-5168-9846
 #              mail: geowieland@googlemail.com              
-# Version:     2.1.4
-# Last update: 2025-12-07 10:27
-# Copyright (c) 2025 Thomas Wieland
+# Version:     2.1.6
+# Last update: 2026-02-26 18:33
+# Copyright (c) 2025-2026 Thomas Wieland
 #-----------------------------------------------------------------------
 
 
 import pandas as pd
 import numpy as np
 import re
+from datetime import datetime
 from collections.abc import Iterable
 from statsmodels.formula.api import ols
 from sklearn.ensemble import BaggingRegressor, RandomForestRegressor, GradientBoostingRegressor
@@ -23,7 +24,6 @@ from xgboost import XGBRegressor
 from lightgbm import LGBMRegressor
 from sklearn.linear_model import LinearRegression
 from sklearn.model_selection import train_test_split
-from huff.goodness_of_fit import modelfit, modelfit_cat, modelfit_plot
 import diffindiff.config as config
 
 
@@ -46,6 +46,30 @@ def check_columns(
         if missing_columns:
             raise KeyError(f"Data do not contain column(s): {', '.join(missing_columns)}")
 
+def is_numeric(
+    df: pd.DataFrame, 
+    columns: list,
+    verbose: bool = config.VERBOSE
+    ):
+    
+    if len(columns) > 0:
+        
+        if verbose:
+            print(f"Checking if column(s) {', '.join(columns)} are numeric", end=" ... ")
+        
+        non_numeric_columns = []
+        
+        for col in columns:
+        
+            if not pd.api.types.is_numeric_dtype(df[col]):
+                non_numeric_columns.append(col)
+        
+        if verbose:
+            print("OK")
+        
+        if non_numeric_columns:
+            raise KeyError(f"Data contain non-numeric column(s): {', '.join(non_numeric_columns)}")
+
 def panel_index(
     data: pd.DataFrame,
     unit_col: str,
@@ -527,8 +551,11 @@ def is_multiple_treatment_period(
         unit_treatment = data_sub[treatment_col]
 
         groups = (unit_treatment != unit_treatment.shift()).cumsum()
-
-        periods_count = (unit_treatment == 1).groupby(groups).any().sum()
+        
+        if config.ACCEPT_CONTINUOUS_TREATMENTS:
+            periods_count = (unit_treatment > 0).groupby(groups).any().sum()
+        else:
+            periods_count = (unit_treatment == 1).groupby(groups).any().sum()
 
         unit_treatment_periods[unit] = int(periods_count)
 
@@ -636,25 +663,31 @@ def treatment_times(
         verbose=verbose
         )
     
-    is_multiple_treatment_period(
+    is_multiple_treatment_period_result = 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]))
-
+    if config.ACCEPT_CONTINUOUS_TREATMENTS:
+        tt = list(unique(data.loc[data[treatment_col] > 0, time_col]))
+    else:
+        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"])
     
     for unit in units:
         
-        data_unit_tt = data[(data[unit_col] == unit) & (data[treatment_col] == 1)]
+        if config.ACCEPT_CONTINUOUS_TREATMENTS:
+            data_unit_tt = data[(data[unit_col] == unit) & (data[treatment_col] > 0)]
+        else:
+            data_unit_tt = data[(data[unit_col] == unit) & (data[treatment_col] == 1)]
         
         if data_unit_tt.empty:
             continue
@@ -678,7 +711,7 @@ def treatment_times(
         
     if verbose:
         print("OK")
-        
+            
     return [
         units_tt,
         tt
@@ -796,9 +829,9 @@ def fit_metrics(
 
     assert observed_no == expected_no, "Error while calculating fit metrics: Observed and expected differ in length"
     
-    if not pd.api.types.is_numeric_dtype(observed):
+    if not pd.api.types.is_numeric_dtype(observed) or not np.issubdtype(observed.dtype, np.number):
         raise ValueError("Error while calculating fit metrics: Observed column is not numeric")
-    if not pd.api.types.is_numeric_dtype(expected):
+    if not pd.api.types.is_numeric_dtype(expected) or not np.issubdtype(expected.dtype, np.number):
         raise ValueError("Error while calculating fit metrics: Expected column is not numeric")
     
     if outcome_col is not None:
@@ -810,8 +843,8 @@ def fit_metrics(
     
     if remove_nan:
         
-        observed = observed.reset_index(drop=True)
-        expected = expected.reset_index(drop=True)
+        observed = np.array(observed)
+        expected = np.array(expected)
 
         obs_exp = pd.DataFrame(
             {
@@ -947,4 +980,30 @@ def bool_to_YN(val):
     if isinstance(val, bool):
         return "YES" if val else "NO"    
     else:        
-        return val
+        return val
+    
+def check_date_format(
+    dates: list = None, 
+    date_format: str = "%Y-%m-%d"
+    ):
+    
+    if dates is None:
+        dates = []
+    
+    invalid_dates_included = False
+    invalid_dates = []
+    
+    for date in dates:
+        try:
+            datetime.strptime(date, date_format)
+        except (ValueError, TypeError):
+            invalid_dates.append(date)
+    
+    if len(invalid_dates) > 0:
+        invalid_dates_included = True
+        invalid_dates = [str(d) for d in invalid_dates]
+    
+    return [
+        invalid_dates_included,
+        invalid_dates
+    ]

{diffindiff-2.2.5 → diffindiff-2.2.7}/diffindiff/tests/tests_diffindiff.py

@@ -4,9 +4,9 @@
 # Author:      Thomas Wieland 
 #              ORCID: 0000-0001-5168-9846
 #              mail: geowieland@googlemail.com              
-# Version:     2.0.10
-# Last update: 2025-12-05 17:23
-# Copyright (c) 2025 Thomas Wieland
+# Version:     2.0.11
+# Last update: 2026-02-20 17:44
+# Copyright (c) 2025-2026 Thomas Wieland
 #-----------------------------------------------------------------------
 
 

{diffindiff-2.2.5 → diffindiff-2.2.7}/diffindiff.egg-info/PKG-INFO

@@ -1,12 +1,12 @@
 Metadata-Version: 2.1
 Name: diffindiff
-Version: 2.2.5
-Summary: diffindiff: Python library for convenient Difference-in-Differences Analyses
+Version: 2.2.7
+Summary: diffindiff: Python library for convenient Difference-in-Differences analyses
 Author: Thomas Wieland
 Author-email: geowieland@googlemail.com
 Description-Content-Type: text/markdown
 
-# diffindiff: A Python library for convenient difference-in-differences analyses
+# diffindiff: Python library for convenient Difference-in-Differences analyses
 
 This Python library is designed for performing Difference-in-Differences (DiD) analyses in a convenient way. It allows users to construct datasets, define treatment and control groups, and set treatment periods. DiD model analyses may be conducted with both datasets created by built-in functions and ready-to-use external datasets. Both simultaneous and staggered adoption are supported. The library allows for various extensions, such as two-way fixed effects models, group- or individual-specific effects, post-treatment periods, and triple-difference estimations. Additionally, it includes functions for visualizing results, such as plotting DiD coefficients with confidence intervals and illustrating the temporal evolution of staggered treatments. Furthermore, several functions for rigorous treatment setting and data diagnostics are incorporated.
 
@@ -20,14 +20,14 @@ Thomas Wieland [ORCID](https://orcid.org/0000-0001-5168-9846) [EMail](mailto:geo
 
 - 📦 PyPI: [diffindiff](https://pypi.org/project/diffindiff/)
 - 💻 GitHub Repository: [diffindiff_official](https://github.com/geowieland/diffindiff_official)
-- 📄 DOI (Zenodo): [10.5281/zenodo.18639559](https://doi.org/10.5281/zenodo.18639559)
+- 📄 DOI (Zenodo): [10.5281/zenodo.18656820](https://doi.org/10.5281/zenodo.18656820)
 
 
 ## Citation
 
 If you use this software, please cite:
 
-Wieland, T. (2026). diffindiff: A Python library for convenient difference-in-differences analyses (Version 2.2.4) [Computer software]. Zenodo. https://doi.org/10.5281/zenodo.18656820
+Wieland, T. (2026). diffindiff: A Python library for convenient difference-in-differences analyses (Version 2.2.7) [Computer software]. Zenodo. https://doi.org/10.5281/zenodo.18656820
 
 
 ## Installation
@@ -167,8 +167,9 @@ See the /tests directory for usage examples of most of the included functions.
   - Wooldridge JM (2012) *Introductory Econometrics. A Modern Approach*.
 
 
-## What's new (v2.2.5)
+## What's new (v2.2.7)
+- Functions
+  - diddata.DiffData.define_treatment() for constructing a new treatment from a column in the dataframe 
 - Bugfixes:
-  - Incorrect import
-- Other:
-  - Update README
+  - didtools.treatment_times() and didtools.is_multiple_treatment_period() now also identify continuous treatments correctly
+  - Fixed problematic type conversion in didtools.fit_metrics()

{diffindiff-2.2.5 → diffindiff-2.2.7}/diffindiff.egg-info/requires.txt

@@ -9,4 +9,3 @@ xgboost
 lightgbm
 patsy
 openpyxl
-huff>=1.6.6

{diffindiff-2.2.5 → diffindiff-2.2.7}/setup.py

@@ -7,8 +7,8 @@ def read_README():
     
 setup(
     name='diffindiff',
-    version='2.2.5',
-    description='diffindiff: Python library for convenient Difference-in-Differences Analyses',
+    version='2.2.7',
+    description='diffindiff: Python library for convenient Difference-in-Differences analyses',
     packages=find_packages(include=["diffindiff", "diffindiff.tests"]),
     include_package_data=True,
     long_description=read_README(),
@@ -30,8 +30,7 @@ setup(
         'xgboost',
         'lightgbm',
         'patsy',
-        'openpyxl',
-        'huff>=1.6.6'
+        'openpyxl'
     ],
     test_suite='tests',
 )