Merge pull request #278 from PSLmodels/simplify-output

Simplify output in order to reduce potential confusion about tmd data files

tmd/create_taxcalc_cached_files.py

@@ -40,9 +40,11 @@ def create_cached_files():
         varray = calc.array(vname)
         fpath = STORAGE_FOLDER / "output" / f"cached_{vname}.npy"
         np.save(fpath, varray, allow_pickle=False)
+
+    # provide timestamp for Makefile
     fpath = STORAGE_FOLDER / "output" / "cached_files"
     with open(fpath, "w", encoding="utf-8") as cfiles:
-        cfiles.write("  ")  # provides timestamp for Makefile
+        cfiles.write("  ")
 
     return 0
 

tmd/create_taxcalc_input_variables.py

@@ -16,6 +16,7 @@
 
 
 TAXYEAR = 2021
+DUMP_ALL_UNROUNDED_VARIABLES = False
 
 
 def create_variable_file(write_file=True):
@@ -33,10 +34,9 @@ def create_variable_file(write_file=True):
     vdf = create_tmd_2021()
     vdf.FLPDYR = TAXYEAR
     vdf.agi_bin = 0
-    weights = vdf.s006.copy()
-    if write_file:
-        # save a copy containing both input and output variables
-        fname = STORAGE_FOLDER / "output" / "tmd_2021.csv"
+    # optionally dump all input and output variables unrounded
+    if write_file and DUMP_ALL_UNROUNDED_VARIABLES:
+        fname = STORAGE_FOLDER / "allvars_unrounded_2021.csv"
         print(f"Writing PUF+CPS file... [{fname}]")
         vdf.to_csv(fname, index=False)
     # streamline dataframe so that it includes only input variables
@@ -52,6 +52,7 @@ def create_variable_file(write_file=True):
     )
     vdf.drop(columns=rec.IGNORED_VARS, inplace=True)
     # round all float variables to nearest integer except for weights
+    weights = vdf.s006.copy()
     vdf = vdf.astype(int)
     vdf.s006 = weights
     for var in ["e00200", "e00900", "e02100"]:

tmd/datasets/tmd.py

@@ -58,4 +58,3 @@ def create_tmd_2021():
 
 if __name__ == "__main__":
     tmd = create_tmd_2021()
-    tmd.to_csv(STORAGE_FOLDER / "output" / "tmd_2021.csv", index=False)

tmd/storage/output/README.md

@@ -4,15 +4,3 @@ Three national files suitable for input to Tax-Calculator:
 - tmd.csv.gz
 - tmd_weights.csv.gz
 - tmd_growfactors.csv
-
-## Warning about `tmd_2021.csv` file
-
-There is a special-purpose `tmd_2021.csv` file that includes 2021
-Tax-Calculator output variables and the pre-optimization weight,
-`s006_original`.  The weights and input variables in this file are not
-rounded for Tax-Calculator input (as they are in the `tmd.csv` file),
-and therefore, there has always been minor differences between the
-content of `tmd_2021.csv` and `tmd.csv` files.  As a result, using the
-`tmd_2021.csv` file is not recommended.  There are plans to remove the
-`tmd_2021.csv` file in the future.
-