def test_regenerate_blk_blk_xwalk(self):
known_ids = numpy.array([
"G10000100401001000",
"G10000100401001001",
"G10000100401001002",
"G10000100401001003",
"G10000100401001003",
])
xwalk_name = base_xwalk_name_fmat % (blk, _90, blk, _10, gj)
path_in = data_dir + xwalk_name + ".%s" % ZIP
path_out = data_dir
dtype = nhgisxwalk.str_types(nhgisxwalk.ID_COLS)
nhgisxwalk.regenerate_blk_blk_xwalk(path_in,
path_out,
"GJOIN2010",
dtype,
remove_unpacked=True)
# read in the crosswalk
gjoin = "GJOIN%s"
gj_src, gj_trg = gjoin % _90, gjoin % _10
data_types = nhgisxwalk.str_types([gj_src, gj_trg])
from_csv_kws = {
"path": path_out,
"archived": True,
"remove_unpacked": True,
}
read_csv_kws = {"dtype": data_types}
read_xwalk = nhgisxwalk.xwalk_df_from_csv(xwalk_name, **from_csv_kws,
**read_csv_kws)
observed_ids = read_xwalk["GJOIN2010"].head().values
numpy.testing.assert_array_equal(known_ids, observed_ids)
def fetch_base_xwalk(sg, tg, sy, ty):
base_xwalk_name = base_xwalk_name_fmat % (sg, sy, tg, ty, gj)
data_types = nhgisxwalk.str_types(["GJOIN%s" % sy, "GJOIN%s" % ty])
from_csv_kws = {
"path": data_dir,
"archived": True,
"remove_unpacked": True
}
read_csv_kws = {"dtype": data_types}
base_xwalk = nhgisxwalk.xwalk_df_from_csv(base_xwalk_name, **from_csv_kws,
**read_csv_kws)
return base_xwalk
def test_xwalk_write_read_csv_from_class(self):
write_xwalk = nhgisxwalk.GeoCrossWalk(
base_xwalk_blk2000_blk2010,
source_year=_00,
target_year=_10,
source_geo=bgp,
target_geo=tr,
base_source_table=tab_data_path_2000,
input_var=input_vars_2000_SF1b,
weight_var=input_var_tags,
keep_base=False,
stfips=stfips,
)
nhgisxwalk.xwalk_df_to_csv(cls=write_xwalk)
read_xwalk = nhgisxwalk.xwalk_df_from_csv(write_xwalk.xwalk_name)
known_values = write_xwalk.xwalk["wt_pop"].values
observed_values = read_xwalk["wt_pop"].values
numpy.testing.assert_allclose(known_values, observed_values)
def test_prepare_data_product(self):
xwalk_name = prod_xwalk_name_fmat % (bgp, _90, co, _10)
path_out = data_dir + xwalk_name
nhgisxwalk.prepare_data_product(self.example_df,
xwalk_name,
path_out,
remove=True)
# read in the crosswalk
from_csv_kws = {
"path": data_dir,
"archived": True,
"remove_unpacked": True,
}
read_xwalk = nhgisxwalk.xwalk_df_from_csv(xwalk_name, **from_csv_kws)
known = numpy.array([1.0, 0.3, 0.7, 1.0, 1.0])
observed = read_xwalk["wt"].values
numpy.testing.assert_array_equal(known, observed)
def test_split_xwalk(self):
known_ids = numpy.array([
"G10000100401001000",
"G10000100401001001",
"G10000100401001002",
"G10000100401001003",
"G10000100401001003",
])
xwalk_name = base_xwalk_name_fmat % (blk, _90, blk, _10, gj)
xwalk_path = data_dir + xwalk_name + "_state"
sorter = nhgisxwalk.SORT_BYS[xwalk_name]
nhgisxwalk.split_xwalk(
base_xwalk_blk1990_blk2010,
"GJOIN2010",
xwalk_name,
gj,
fpath=xwalk_path,
sort_by=sorter,
)
# read in the crosswalk
gjoin = "GJOIN%s"
gj_src, gj_trg = gjoin % _90, gjoin % _10
data_types = nhgisxwalk.str_types([gj_src, gj_trg])
from_csv_kws = {
"path": xwalk_path + "/",
"archived": True,
"remove_unpacked": True,
}
read_csv_kws = {"dtype": data_types}
read_xwalk = nhgisxwalk.xwalk_df_from_csv(xwalk_name + "_%s" % stfips,
**from_csv_kws,
**read_csv_kws)
observed_ids = read_xwalk["GJOIN2010"].head().values
numpy.testing.assert_array_equal(known_ids, observed_ids)
def test_generate_data_product(self):
# records known data values
knw_str_vals = numpy.array([
[
"G100001090444999990421009999999219012",
"G1000010042100",
"10001042100",
],
[
"G100001090444999990421009999999999921",
"G1000010042100",
"10001042100",
],
[
"G100001090444999990421009999999999921",
"G1000010042201",
"10001042201",
],
[
"G100001090444999990421009999999999922",
"G1000010042100",
"10001042100",
],
])
knw_num_vals = numpy.array([
[1.0, 1.0, 1.0, 1.0],
[0.99766436, 0.99716625, 0.99714829, 0.99727768],
[0.00233564, 0.00283375, 0.00285171, 0.00272232],
[1.0, 1.0, 1.0, 1.0],
])
# generate the product
xwalk_args = {
"source_year": _90,
"target_year": _10,
"source_geo": bgp,
"target_geo": tr,
"base_source_table": tab_data_path_1990,
"supp_source_table": supplement_data_path_90,
"input_var": input_vars_1990,
"weight_var": input_var_tags,
"keep_base": False,
"add_geoid": True,
}
nhgisxwalk.generate_data_product(base_xwalk_blk1990_blk2010,
xwalk_args, data_dir)
# record observed
obs_xwalk = nhgisxwalk.GeoCrossWalk(
base_xwalk_blk1990_blk2010,
source_year=_90,
target_year=_10,
source_geo=bgp,
target_geo=tr,
base_source_table=tab_data_path_1990,
supp_source_table=supplement_data_path_90,
input_var=input_vars_1990,
weight_var=input_var_tags,
)
# test
id_cols = ["bgp1990gj", "tr2010gj", "tr2010ge"]
data_types = nhgisxwalk.str_types(id_cols)
from_csv_kws = {
"path": data_dir,
"archived": True,
"remove_unpacked": True,
}
read_csv_kws = {"dtype": data_types}
read_xwalk = nhgisxwalk.xwalk_df_from_csv(obs_xwalk.xwalk_name,
**from_csv_kws,
**read_csv_kws)
ix1, ix2 = 13, 17
obs_str_vals = obs_xwalk.xwalk[id_cols][ix1:ix2].values
wgt_cols = ["wt_pop", "wt_fam", "wt_hh", "wt_hu"]
obs_num_vals = obs_xwalk.xwalk[wgt_cols][ix1:ix2].values
numpy.testing.assert_equal(knw_str_vals, obs_str_vals)
numpy.testing.assert_allclose(knw_num_vals, obs_num_vals, atol=6)