Merge pull request #517 from PSLmodels/cleanup-area-deadwood

Cleanup area deadwood

Author: donboyd5

tmd/areas/AREA_WEIGHTING_GUIDE.md

@@ -624,7 +624,6 @@ tmd/areas/
 ├── developer_tools.py          # Difficulty / dual / relaxation cascade
 ├── quality_report.py           # Multi- and single-area quality reports
 ├── sweep_params.py             # Solver parameter grid search
-├── make_all.py                 # End-to-end driver used in CI
 ├── targets/                    # Per-area target CSVs (output)
 │   ├── states/
 │   ├── cds_118/

tmd/areas/README.md

@@ -184,7 +184,6 @@ tmd/areas/
 ├── batch_weights.py                parallel-runner support
 ├── solver_overrides.py             read per-area override YAML
 ├── sweep_params.py                 parameter grid-search utility
-├── make_all.py                     end-to-end driver used in CI
 ├── prepare/                        shares + targets building blocks
 │   ├── data/                       SOI inputs and CD crosswalks
 │   ├── recipes/                    target specs

tmd/areas/batch_weights.py

@@ -310,6 +310,25 @@ def _list_target_areas(target_dir=None):
     return areas
 
 
+def _time_of_newest_other_dependency():
+    """
+    Return mtime of the newest non-target dependency that should
+    invalidate cached area-weight files.
+    """
+    from tmd.areas import AREAS_FOLDER
+    from tmd.storage import STORAGE_FOLDER
+    from tmd.imputation_assumptions import POPULATION_FILE
+
+    deps = [
+        AREAS_FOLDER / "create_area_weights.py",
+        STORAGE_FOLDER / "output" / "tmd.csv.gz",
+        STORAGE_FOLDER / "output" / "tmd_weights.csv.gz",
+        STORAGE_FOLDER / "output" / "tmd_growfactors.csv",
+        STORAGE_FOLDER / "input" / POPULATION_FILE,
+    ]
+    return max(dpath.stat().st_mtime for dpath in deps)
+
+
 def _filter_areas(areas, area_filter):
     """Filter areas by type: 'states', 'cds', or 'all'."""
     if area_filter == "all":
@@ -366,9 +385,7 @@ def run_batch(
 
     # Filter to out-of-date areas unless force=True
     if not force:
-        from tmd.areas.make_all import time_of_newest_other_dependency
-
-        newest_dep = time_of_newest_other_dependency()
+        newest_dep = _time_of_newest_other_dependency()
         todo = []
         for area in areas:
             wpath = weight_dir / f"{area}_tmd_weights.csv.gz"

tmd/areas/create_area_weights.py

@@ -26,8 +26,8 @@
 
 This module is a library, not a CLI.  It is called by
 ``tmd.areas.solve_weights`` (parallel batch solver, the production
-entry point), ``tmd.areas.developer_tools`` (relaxation cascade),
-and ``tmd.areas.make_all`` (CI driver).  Key public entry points:
+entry point) and ``tmd.areas.developer_tools`` (relaxation cascade).
+Key public entry points:
 
     create_area_weights_file(area, ...)   — solve one area, write
                                              ``<area>_tmd_weights.csv.gz``

tmd/areas/make_all.py

@@ -106,21 +106,13 @@ def build_agi_labels(area_type: AreaType) -> pd.DataFrame:
 
 # --- SOI file naming patterns ---
 
-# State SOI CSV files by year (2-digit year substituted into pattern)
+# State SOI CSV files by year
 SOI_STATE_CSV_PATTERNS: Dict[int, str] = {
-    2015: "15in54cmcsv.csv",
-    2016: "16in54cmcsv.csv",
-    2017: "17in54cmcsv.csv",
-    2018: "18in55cmagi.csv",
-    2019: "19in55cmcsv.csv",
-    2020: "20in55cmcsv.csv",
-    2021: "21in55cmcsv.csv",
     2022: "22in55cmcsv.csv",
 }
 
 # CD SOI CSV files by year
 SOI_CD_CSV_PATTERNS: Dict[int, str] = {
-    2021: "21incd.csv",
     2022: "22incd.csv",
 }