Infer timeseries frequency after loading

tests/test_disaggregation.py

@@ -190,7 +190,6 @@ def test_chow_lin(self):
 
     def test_chow_lin_backcasting_error(self):
         # Test case for quarterly to monthly disaggregation backcasting error, issue #6
-
         expected = pd.read_csv("tests/data/R_Output_chow-lin_QtoM_2.csv")
 
         low_freq_data = pd.read_csv("tests/data/AL_Quarterly_Data_Modified.csv")
@@ -281,6 +280,29 @@ def test_chow_lin_two_indicator(self):
 
         self.assertEqual(expected, sales_q_chow_lin.to_frame())
 
+    def test_chow_lin_no_freq(self):
+        expected = pd.read_csv("tests/data/R_output_chow_lin_two_indicator.csv", index_col=0)
+        expected.index = self.exports_q.index
+        expected.columns = ["sales"]
+
+        df2 = self.exports_q.merge(self.imports_q, left_index=True, right_index=True)
+
+        sales_nofreq = self.sales_a.copy()
+        sales_nofreq.index.freq = None
+        assert sales_nofreq.index.freq is None
+
+        sales_q_chow_lin = disaggregate_series(
+            sales_nofreq,
+            df2.resample("QS-OCT").first().assign(constant=1),
+            method="chow-lin",
+            agg_func="sum",
+            optimizer_kwargs={"method": "l-bfgs-b"},
+            verbose=True,
+        )
+
+        self.assertEqual(expected, sales_q_chow_lin.to_frame())
+        assert sales_q_chow_lin.index.freq == "QS-OCT"
+
     def test_denton(self):
         expected = pd.read_csv("tests/data/R_output_denton.csv", index_col=0)
         expected.index = pd.date_range(start="1975-01-01", freq="QS-OCT", periods=144)

tsdisagg/ts_disagg.py

@@ -290,6 +290,9 @@ def prepare_input_dataframes(df1, df2, target_freq, method):
             "dataframe of high-frequency indicators must be provided."
         )
 
+    df1_out.index.freq = df1_out.index.inferred_freq
+    df2_out.index.freq = df2_out.index.inferred_freq
+
     df = pd.merge(df1_out, df2_out, left_index=True, right_index=True, how="outer")
     return df, df1_out, df2_out, time_conversion_factor