Add support for TIGER years 2000 and 2009-2023 (incl.)

Small refactor in us_census module, to remove awkward concept of "CensusYear".
NAU-CCL · Jul 8, 2024 · 77cc8a1 · 77cc8a1
1 parent 9a2608a
commit 77cc8a1
Show file tree

Hide file tree

Showing 4 changed files with 132 additions and 95 deletions.
diff --git a/doc/devlog/2024-05-03.ipynb b/doc/devlog/2024-05-03.ipynb
@@ -8,11 +8,13 @@
     "\n",
     "_Author: Tyler Coles_\n",
     "\n",
-    "Testing our us_census functions for loading canonical sets of IDs for Census granularities from state to block group for the years 2000, 2010, and 2020.\n",
+    "Testing our us_census functions for loading canonical sets of IDs for Census granularities from state to block group for all supported TIGER years (2000 and 2009-2023).\n",
     "\n",
     "Since this is our source of truth for these delineations, we want to make sure we're getting complete data. One thing we can test is that at each level of granularity (above block group) each node should contain at least one child node. That is every state should contain a county, every county a tract, and every tract a block group. Otherwise we know something is missing.\n",
     "\n",
-    "(This may seem like a trivial test, but in fact it discovered that my original assumptions about how TIGER provides the data were invalid and has already saved us from bugs!)"
+    "(This may seem like a trivial test, but in fact it discovered that my original assumptions about how TIGER provides the data were invalid and has already saved us from bugs!)\n",
+    "\n",
+    "WARNING: this will take a very long time if you don't have the TIGER files cached."
    ]
   },
   {
@@ -22,13 +24,14 @@
    "outputs": [],
    "source": [
     "import epymorph.geography.us_census as c\n",
+    "import epymorph.geography.us_tiger as t\n",
     "\n",
     "\n",
     "class Fail(Exception):\n",
     "    pass\n",
     "\n",
     "\n",
-    "def test_year(year: c.CensusYear) -> None:\n",
+    "def test_year(year: int) -> None:\n",
     "    # 1. test that we have 52 states\n",
     "    states = c.get_us_states(year).geoid\n",
     "\n",
@@ -80,46 +83,28 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "Census year 2020 passed!\n"
-     ]
-    }
-   ],
-   "source": [
-    "test_year(2020)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 3,
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Census year 2010 passed!\n"
-     ]
-    }
-   ],
-   "source": [
-    "test_year(2010)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 4,
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Census year 2000 passed!\n"
+      "Census year 2000 passed!\n",
+      "Census year 2009 passed!\n",
+      "Census year 2010 passed!\n",
+      "Census year 2011 passed!\n",
+      "Census year 2012 passed!\n",
+      "Census year 2013 passed!\n",
+      "Census year 2014 passed!\n",
+      "Census year 2015 passed!\n",
+      "Census year 2016 passed!\n",
+      "Census year 2017 passed!\n",
+      "Census year 2018 passed!\n",
+      "Census year 2019 passed!\n",
+      "Census year 2020 passed!\n",
+      "Census year 2021 passed!\n",
+      "Census year 2022 passed!\n",
+      "Census year 2023 passed!\n"
      ]
     }
    ],
    "source": [
-    "test_year(2000)"
+    "for year in t.TIGER_YEARS:\n",
+    "    test_year(year)"
    ]
   }
  ],

diff --git a/epymorph/geo/adrio/census/adrio_census.py b/epymorph/geo/adrio/census/adrio_census.py
@@ -12,7 +12,8 @@
 
 from epymorph.data_shape import Shapes
 from epymorph.data_type import CentroidDType
-from epymorph.error import DataResourceException, GeoValidationException
+from epymorph.error import (DataResourceException, GeographyError,
+                            GeoValidationException)
 from epymorph.geo.adrio.adrio import ADRIO, ADRIOMaker
 from epymorph.geo.spec import TimePeriod, Year
 from epymorph.geography.us_census import (BLOCK_GROUP, COUNTY, STATE, TRACT,
@@ -22,7 +23,7 @@
                                           StateScopeAll, TractScope)
 from epymorph.geography.us_tiger import (get_block_groups_geo,
                                          get_counties_geo, get_states_geo,
-                                         get_tracts_geo)
+                                         get_tracts_geo, is_tiger_year)
 from epymorph.simulation import AttributeDef, geo_attrib
 
 
@@ -197,18 +198,22 @@ def fetch_acs5(self, variables: list[str], scope: CensusScope, year: int) -> Dat
     def fetch_sf(self, scope: CensusScope) -> GeoDataFrame:
         """Utility function to fetch shape files from Census for specified regions."""
         # call appropriate pygris function based on granularity and sort result
+        scope_year = scope.year
+        if not is_tiger_year(scope_year):
+            raise GeographyError(f"Unsupported year: {scope_year}")
+
         match scope:
             case StateScopeAll() | StateScope():
-                df = get_states_geo(year=scope.year)
+                df = get_states_geo(year=scope_year)
 
             case CountyScope():
-                df = get_counties_geo(year=scope.year)
+                df = get_counties_geo(year=scope_year)
 
             case TractScope():
-                df = get_tracts_geo(year=scope.year)
+                df = get_tracts_geo(year=scope_year)
 
             case BlockGroupScope():
-                df = get_block_groups_geo(year=scope.year)
+                df = get_block_groups_geo(year=scope_year)
 
             case _:
                 raise DataResourceException("Unsupported query.")