Parse periods in Converter

src/obr_fetcher/converter.py

+import re
 from collections.abc import Iterator
 from typing import TYPE_CHECKING
 
@@ -8,7 +9,9 @@ from dbnomics_data_model.model import (
     DatasetMetadata,
     DatasetReference,
     Observation,
+    Period,
     ProviderMetadata,
+    QuarterPeriod,
     Series,
 )
 
@@ -81,9 +84,23 @@ class Converter:
         df = self._source_data_loader.load_march24_dataset()
 
         for series_name, column in df.items():
-            observations = [Observation.create(period=period, value=value) for period, value in column.items()]
+            observations = [
+                Observation.create(period=self._parse_period(source_period), value=value)
+                for source_period, value in column.items()
+            ]
             yield Series.create(
                 code="E",
                 name=series_name,
                 observations=observations,
             )
+
+    def _parse_period(self, source_period: str) -> Period:
+        quarter_period_re = re.compile(r"(\d{4})Q(\d)")
+        match = quarter_period_re.fullmatch(source_period)
+        if match is None:
+            msg = f"Not a period: {source_period!r}"
+            raise ValueError(msg)
+
+        year = int(match.group(1))
+        quarter = int(match.group(2))
+        return QuarterPeriod(year, quarter)

src/obr_fetcher/source_data_loader.py

@@ -77,26 +77,9 @@ class SourceDataLoader:
     def _load_march24_from_xlsx(self, xlsx_file: Path) -> DataFrame:
         df = pd.read_excel(xlsx_file, header=None, sheet_name="1.6")
 
-        employment_df = df.iloc[2:88, 1:3]
+        employment_df = df.iloc[2:88, 1:3]  # keep "quarterly" section only
         employment_df.rename(columns=employment_df.iloc[0], inplace=True)
         employment_df = employment_df[1:]
-        employment_df.rename(columns={employment_df.columns[0]: "Period"}, inplace=True)
-
-        formatted_periods: list[str] = []
-        for period in employment_df["Period"]:
-            res = ""
-
-            for char in period:
-                if char == "Q":
-                    res += "-"
-
-                res += char
-
-            formatted_periods.append(res)
-
-        employment_df["Period"] = formatted_periods
-
-        employment_df.replace(float("nan"), "NA", inplace=True)
         employment_df.set_index(employment_df.columns[0], inplace=True)
 
         return employment_df