fasttrack: fasttrack/latest/infant_mortality_vaccination_shattock.csv

etl/steps/data/grapher/fasttrack/latest/infant_mortality_vaccination_shattock.py

@@ -1,3 +1,5 @@
+import pandas as pd
+
 from etl.helpers import PathFinder, create_dataset, get_metadata_path
 from etl.snapshot import Snapshot
 
@@ -11,12 +13,28 @@ def run(dest_dir: str) -> None:
     # load data
     tb = snap.read_csv()
 
+    # add dimensions with dim_ prefix
+    dims = [c for c in tb.columns if c.startswith("dim_")]
+    dims_without_prefix = [c[4:] for c in dims]
+
+    if dims:
+        tb = tb.rename(columns={d: dw for d, dw in zip(dims, dims_without_prefix)})
+
+    if uses_dates(tb["year"]):
+        tb = tb.rename(columns={"year": "date"}).format(["country", "date"] + dims_without_prefix)
+    else:
+        tb = tb.format(["country", "year"] + dims_without_prefix)
+
     # add table, update metadata from *.meta.yml and save
-    ds = create_dataset(dest_dir, tables=[tb.set_index(["country", "year"])], default_metadata=snap.metadata)
+    ds = create_dataset(dest_dir, tables=[tb], default_metadata=snap.metadata)
 
     # override metadata if necessary
     meta_path = get_metadata_path(dest_dir).with_suffix(".override.yml")
     if meta_path.exists():
         ds.update_metadata(meta_path)
 
     ds.save()
+
+
+def uses_dates(s: pd.Series) -> bool:
+    return pd.to_datetime(s, errors="coerce", format="%Y-%m-%d").notnull().all()

snapshots/fasttrack/latest/infant_mortality_vaccination_shattock.csv.dvc

@@ -1,14 +1,14 @@
 meta:
   origin:
     producer: |-
-      Shattock et al. (2024). Contribution of vaccination to improved child survival: modelling 50 years of the Expanded Programme on Immunization.
+      Shattock et al. (2024). Contribution of vaccination to improved survival and health: modelling 50 years of the Expanded Programme on Immunization.
     title: infant_mortality_vaccination_shattock
     citation_full: |-
-      Shattock et al. (2024). Contribution of vaccination to improved child survival: modelling 50 years of the Expanded Programme on Immunization. The Lancet.
+      Shattock et al. (2024). Contribution of vaccination to improved survival and health: modelling 50 years of the Expanded Programme on Immunization. The Lancet.
     url_main: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(24)00850-X/fulltext
     url_download: |-
       https://docs.google.com/spreadsheets/d/e/2PACX-1vSLcIzJwPFHDM0x41c0X0h4_YNiw43W3YtVQfpoNtVKOjCZuOyyFcd0JM1O-RjCW_DASDOJkWRIfntY/pub?output=csv
-    date_accessed: '2024-05-03'
+    date_accessed: '2024-11-28'
   name: Infant mortality with and without vaccinations (Shattock et al, 2024)
   description: Estimates of the infant mortality rates with and without vital vaccinations.
   license: {}