def get_scored_comparisons(self):
"""Use the EM algorithm to estimate model parameters and return match probabilities.
Note: Does not compute term frequency adjustments.
Returns:
DataFrame: A spark dataframe including a match probability column
"""
df_comparison = self._get_df_comparison()
df_gammas = add_gammas(df_comparison, self.settings, self.spark)
df_gammas = self.break_lineage_blocked_comparisons(
df_gammas, self.spark)
df_e = iterate(
df_gammas,
self.params,
self.settings,
self.spark,
compute_ll=False,
save_state_fn=self.save_state_fn,
)
# In case the user's break lineage function has persisted it
df_gammas.unpersist()
df_e = self.break_lineage_scored_comparisons(df_e, self.spark)
df_e_adj = self._make_term_frequency_adjustments(df_e)
df_e.unpersist()
return df_e_adj
def test_expectation(spark, sqlite_con_1, params_1, gamma_settings_1):
dfpd = pd.read_sql("select * from test1", sqlite_con_1)
df = spark.createDataFrame(dfpd)
gamma_settings_1["blocking_rules"] = [
"l.mob = r.mob",
"l.surname = r.surname",
]
df_comparison = block_using_rules(gamma_settings_1, df=df, spark=spark)
df_gammas = add_gammas(df_comparison, gamma_settings_1, spark)
# df_e = iterate(df_gammas, spark, params_1, num_iterations=1)
df_e = run_expectation_step(df_gammas, params_1, gamma_settings_1, spark)
df_e_pd = df_e.toPandas()
df_e_pd = df_e_pd.sort_values(["unique_id_l", "unique_id_r"])
correct_list = [
0.893617021,
0.705882353,
0.705882353,
0.189189189,
0.189189189,
0.893617021,
0.375,
0.375,
]
result_list = list(df_e_pd["match_probability"].astype(float))
for i in zip(result_list, correct_list):
assert i[0] == pytest.approx(i[1])
def test_tiny_numbers(spark, sqlite_con_1):
# Regression test, see https://github.com/moj-analytical-services/splink/issues/48
dfpd = pd.read_sql("select * from test1", sqlite_con_1)
df = spark.createDataFrame(dfpd)
settings = {
"link_type": "dedupe_only",
"proportion_of_matches": 0.4,
"comparison_columns": [
{
"col_name": "mob",
"num_levels": 2,
"m_probabilities": [5.9380419956766985e-25, 1 - 5.9380419956766985e-25],
"u_probabilities": [0.8, 0.2],
},
{"col_name": "surname", "num_levels": 2,},
],
"blocking_rules": ["l.mob = r.mob", "l.surname = r.surname",],
}
settings = complete_settings_dict(settings, spark=None)
df_comparison = block_using_rules(settings, df=df, spark=spark)
df_gammas = add_gammas(df_comparison, settings, spark)
params = Params(settings, spark="supress_warnings")
df_e = run_expectation_step(df_gammas, params, settings, spark)
def get_scored_comparisons(self):
"""Use the EM algorithm to estimate model parameters and return match probabilities.
Note: Does not compute term frequency adjustments.
Returns:
DataFrame: A spark dataframe including a match probability column
"""
df_comparison = self._get_df_comparison()
df_gammas = add_gammas(df_comparison, self.settings, self.spark)
df_gammas.persist()
df_e = iterate(
df_gammas,
self.params,
self.settings,
self.spark,
compute_ll=False,
save_state_fn=self.save_state_fn,
)
df_gammas.unpersist()
return df_e
def manually_apply_fellegi_sunter_weights(self):
"""Compute match probabilities from m and u probabilities specified in the splink settings object
Returns:
DataFrame: A spark dataframe including a match probability column
"""
df_comparison = self._get_df_comparison()
df_gammas = add_gammas(df_comparison, self.settings, self.spark)
return run_expectation_step(df_gammas, self.params, self.settings,
self.spark)
def manually_apply_fellegi_sunter_weights(self):
"""Compute match probabilities from m and u probabilities specified in the splink settings object
Returns:
DataFrame: A spark dataframe including a match probability column
"""
df_comparison = block_using_rules(self.settings_dict, self.df,
self.spark)
df_gammas = add_gammas(df_comparison, self.settings_dict, self.spark)
# see https://github.com/moj-analytical-services/splink/issues/187
df_gammas = self.break_lineage_blocked_comparisons(
df_gammas, self.spark)
return run_expectation_step(df_gammas, self.model, self.spark)
def test_iterate(spark, sqlite_con_1, params_1, gamma_settings_1):
original_params = copy.deepcopy(params_1.params)
dfpd = pd.read_sql("select * from test1", sqlite_con_1)
df = spark.createDataFrame(dfpd)
rules = [
"l.mob = r.mob",
"l.surname = r.surname",
]
gamma_settings_1["blocking_rules"] = rules
df_comparison = block_using_rules(gamma_settings_1, df=df, spark=spark)
df_gammas = add_gammas(df_comparison, gamma_settings_1, spark)
gamma_settings_1["max_iterations"] = 1
df_e = iterate(df_gammas, params_1, gamma_settings_1, spark)
assert params_1.params["λ"] == pytest.approx(0.540922141)
assert params_1.params["π"]["gamma_mob"]["prob_dist_match"]["level_0"][
"probability"] == pytest.approx(0.087438272, abs=0.0001)
assert params_1.params["π"]["gamma_surname"]["prob_dist_non_match"][
"level_1"]["probability"] == pytest.approx(0.160167628, abs=0.0001)
first_it_params = copy.deepcopy(params_1.params)
df_e_pd = df_e.toPandas()
df_e_pd = df_e_pd.sort_values(["unique_id_l", "unique_id_r"])
correct_list = [
0.658602114,
0.796821727,
0.796821727,
0.189486495,
0.189486495,
0.658602114,
0.495063367,
0.495063367,
]
result_list = list(df_e_pd["match_probability"].astype(float))
for i in zip(result_list, correct_list):
assert i[0] == pytest.approx(i[1], abs=0.0001)
# Does it still work with another iteration?
gamma_settings_1["max_iterations"] = 1
df_e = iterate(df_gammas, params_1, gamma_settings_1, spark)
assert params_1.params["λ"] == pytest.approx(0.534993426, abs=0.0001)
assert params_1.params["π"]["gamma_mob"]["prob_dist_match"]["level_0"][
"probability"] == pytest.approx(0.088546179, abs=0.0001)
assert params_1.params["π"]["gamma_surname"]["prob_dist_non_match"][
"level_1"]["probability"] == pytest.approx(0.109234086, abs=0.0001)
## Test whether the params object is correctly storing the iteration history
assert params_1.param_history[0] == original_params
assert params_1.param_history[1] == first_it_params
## Now test whether, when we
data = params_1._convert_params_dict_to_dataframe(original_params)
val1 = {
"gamma": "gamma_mob",
"match": 0,
"value_of_gamma": "level_0",
"probability": 0.8,
"value": 0,
"column": "mob",
}
val2 = {
"gamma": "gamma_surname",
"match": 1,
"value_of_gamma": "level_1",
"probability": 0.2,
"value": 1,
"column": "surname",
}
assert val1 in data
assert val2 in data
correct_list = [{
"iteration": 0,
"λ": 0.4
}, {
"iteration": 1,
"λ": 0.540922141
}]
result_list = params_1._iteration_history_df_lambdas()
for i in zip(result_list, correct_list):
assert i[0]["iteration"] == i[1]["iteration"]
assert i[0]["λ"] == pytest.approx(i[1]["λ"])
result_list = params_1._iteration_history_df_gammas()
val1 = {
"iteration": 0,
"gamma": "gamma_mob",
"match": 0,
"value_of_gamma": "level_0",
"probability": 0.8,
"value": 0,
"column": "mob",
}
assert val1 in result_list
val2 = {
"iteration": 1,
"gamma": "gamma_surname",
"match": 0,
"value_of_gamma": "level_1",
"probability": 0.160167628,
"value": 1,
"column": "surname",
}
for r in result_list:
if r["iteration"] == 1:
if r["gamma"] == "gamma_surname":
if r["match"] == 0:
if r["value"] == 1:
record = r
for k, v in record.items():
expected_value = val2[k]
if k == "probability":
assert v == pytest.approx(expected_value, abs=0.0001)
else:
assert v == expected_value
# Test whether saving and loading parameters works
import tempfile
dir = tempfile.TemporaryDirectory()
fname = os.path.join(dir.name, "params.json")
# print(params_1.params)
# import json
# print(json.dumps(params_1.to_dict(), indent=4))
params_1.save_params_to_json_file(fname)
from splink.params import load_params_from_json
p = load_params_from_json(fname)
assert p.params["λ"] == pytest.approx(params_1.params["λ"])
def test_add_gammas(spark):
rows = [
{
"unique_id_l": 1,
"unique_id_r": 2,
"fname_l": "robin",
"fname_r": "robin",
"sname_l": "linacre",
"sname_r": "linacre",
},
{
"unique_id_l": 3,
"unique_id_r": 4,
"fname_l": "robin",
"fname_r": "robin",
"sname_l": "linacrr",
"sname_r": "linacre",
},
{
"unique_id_l": 5,
"unique_id_r": 6,
"fname_l": None,
"fname_r": None,
"sname_l": None,
"sname_r": "linacre",
},
{
"unique_id_l": 7,
"unique_id_r": 8,
"fname_l": "robin",
"fname_r": "julian",
"sname_l": "linacre",
"sname_r": "smith",
},
]
df = spark.createDataFrame(Row(**x) for x in rows)
gamma_settings = {
"link_type":
"dedupe_only",
"proportion_of_matches":
0.5,
"comparison_columns": [
{
"col_name": "fname",
"num_levels": 2
},
{
"col_name":
"sname",
"num_levels":
3,
"case_expression":
"""
case
when sname_l is null or sname_r is null then -1
when sname_l = sname_r then 2
when substr(sname_l,1, 3) = substr(sname_r, 1, 3) then 1
else 0
end
as gamma_sname
""",
},
],
"blocking_rules": [],
"retain_matching_columns":
False,
}
df_gammas = add_gammas(df, gamma_settings, spark)
correct_answer = [
{
"unique_id_l": 1,
"unique_id_r": 2,
"gamma_fname": 1,
"gamma_sname": 2
},
{
"unique_id_l": 3,
"unique_id_r": 4,
"gamma_fname": 1,
"gamma_sname": 1
},
{
"unique_id_l": 5,
"unique_id_r": 6,
"gamma_fname": -1,
"gamma_sname": -1
},
{
"unique_id_l": 7,
"unique_id_r": 8,
"gamma_fname": 0,
"gamma_sname": 0
},
]
pd_correct = pd.DataFrame(correct_answer)
pd_correct = pd_correct.sort_values(["unique_id_l", "unique_id_r"])
pd_correct = pd_correct.astype(int)
pd_result = df_gammas.toPandas()
pd_result = pd_result.sort_values(["unique_id_l", "unique_id_r"])
pd_result = pd_result.astype(int)
assert_frame_equal(pd_correct, pd_result)
gamma_settings["retain_matching_columns"] = True
df_gammas = add_gammas(df, gamma_settings, spark)
result = df_gammas.toPandas()
col_names = list(result.columns)
correct_col_names = [
"unique_id_l",
"unique_id_r",
"fname_l",
"fname_r",
"gamma_fname",
"sname_l",
"sname_r",
"gamma_sname",
]
assert col_names == correct_col_names
# With source datset
gamma_settings["source_dataset_column_name"] = "source_ds"
df = df.withColumn("source_ds_l", lit("ds"))
df = df.withColumn("source_ds_r", lit("ds"))
df_gammas = add_gammas(df, gamma_settings, spark)
result = df_gammas.toPandas()
col_names = list(result.columns)
correct_col_names = [
"source_ds_l",
"unique_id_l",
"source_ds_r",
"unique_id_r",
"fname_l",
"fname_r",
"gamma_fname",
"sname_l",
"sname_r",
"gamma_sname",
]
assert col_names == correct_col_names
def test_expectation_and_maximisation(spark):
settings = {
"link_type":
"dedupe_only",
"proportion_of_matches":
0.4,
"comparison_columns": [
{
"col_name": "mob",
"num_levels": 2,
"m_probabilities": [0.1, 0.9],
"u_probabilities": [0.8, 0.2],
},
{
"custom_name": "surname",
"custom_columns_used": ["surname"],
"num_levels": 3,
"case_expression": """
case
when surname_l is null or surname_r is null then -1
when surname_l = surname_r then 2
when substr(surname_l,1, 3) = substr(surname_r, 1, 3) then 1
else 0
end
as gamma_surname
""",
"m_probabilities": [0.1, 0.2, 0.7],
"u_probabilities": [0.5, 0.25, 0.25],
},
],
"blocking_rules": [
"l.mob = r.mob",
"l.surname = r.surname",
],
"retain_intermediate_calculation_columns":
True,
}
rows = [
{
"unique_id": 1,
"mob": 10,
"surname": "Linacre"
},
{
"unique_id": 2,
"mob": 10,
"surname": "Linacre"
},
{
"unique_id": 3,
"mob": 10,
"surname": "Linacer"
},
{
"unique_id": 4,
"mob": 7,
"surname": "Smith"
},
{
"unique_id": 5,
"mob": 8,
"surname": "Smith"
},
{
"unique_id": 6,
"mob": 8,
"surname": "Smith"
},
{
"unique_id": 7,
"mob": 8,
"surname": "Jones"
},
]
df_input = spark.createDataFrame(Row(**x) for x in rows)
df_input.persist()
params = Model(settings, spark)
df_comparison = block_using_rules(
params.current_settings_obj.settings_dict, df_input, spark)
df_gammas = add_gammas(df_comparison,
params.current_settings_obj.settings_dict, spark)
df_gammas.persist()
df_e = run_expectation_step(df_gammas, params, spark)
df_e = df_e.sort("unique_id_l", "unique_id_r")
df_e.persist()
################################################
# Test probabilities correctly assigned
################################################
df = df_e.toPandas()
cols_to_keep = [
"prob_gamma_mob_match",
"prob_gamma_mob_non_match",
"prob_gamma_surname_match",
"prob_gamma_surname_non_match",
]
pd_df_result = df[cols_to_keep][:4]
df_correct = [
{
"prob_gamma_mob_match": 0.9,
"prob_gamma_mob_non_match": 0.2,
"prob_gamma_surname_match": 0.7,
"prob_gamma_surname_non_match": 0.25,
},
{
"prob_gamma_mob_match": 0.9,
"prob_gamma_mob_non_match": 0.2,
"prob_gamma_surname_match": 0.2,
"prob_gamma_surname_non_match": 0.25,
},
{
"prob_gamma_mob_match": 0.9,
"prob_gamma_mob_non_match": 0.2,
"prob_gamma_surname_match": 0.2,
"prob_gamma_surname_non_match": 0.25,
},
{
"prob_gamma_mob_match": 0.1,
"prob_gamma_mob_non_match": 0.8,
"prob_gamma_surname_match": 0.7,
"prob_gamma_surname_non_match": 0.25,
},
]
pd_df_correct = pd.DataFrame(df_correct)
assert_frame_equal(pd_df_correct, pd_df_result)
################################################
# Test match probabilities correctly calculated
################################################
result_list = list(df["match_probability"])
# See https://github.com/moj-analytical-services/splink/blob/master/tests/expectation_maximisation_test_answers.xlsx
# for derivation of these numbers
correct_list = [
0.893617021,
0.705882353,
0.705882353,
0.189189189,
0.189189189,
0.893617021,
0.375,
0.375,
]
assert result_list == pytest.approx(correct_list)
################################################
# Test new probabilities correctly calculated
################################################
run_maximisation_step(df_e, params, spark)
new_lambda = params.current_settings_obj["proportion_of_matches"]
# See https://github.com/moj-analytical-services/splink/blob/master/tests/expectation_maximisation_test_answers.xlsx
# for derivation of these numbers
assert new_lambda == pytest.approx(0.540922141)
rows = [
["mob", 0, 0.087438272, 0.441543191],
["mob", 1, 0.912561728, 0.558456809],
["surname", 0, 0.173315146, 0.340356209],
["surname", 1, 0.326240275, 0.160167628],
["surname", 2, 0.500444578, 0.499476163],
]
settings_obj = params.current_settings_obj
for r in rows:
cc = settings_obj.get_comparison_column(r[0])
level_dict = cc.level_as_dict(r[1])
assert level_dict["m_probability"] == pytest.approx(r[2])
assert level_dict["u_probability"] == pytest.approx(r[3])
################################################
# Test revised probabilities correctly used
################################################
df_e = run_expectation_step(df_gammas, params, spark)
df_e = df_e.sort("unique_id_l", "unique_id_r")
result_list = list(df_e.toPandas()["match_probability"])
correct_list = [
0.658602114,
0.796821727,
0.796821727,
0.189486495,
0.189486495,
0.658602114,
0.495063367,
0.495063367,
]
assert result_list == pytest.approx(correct_list)
run_maximisation_step(df_e, params, spark)
new_lambda = params.current_settings_obj["proportion_of_matches"]
assert new_lambda == pytest.approx(0.534993426)
rows = [
["mob", 0, 0.088546179, 0.435753788],
["mob", 1, 0.911453821, 0.564246212],
["surname", 0, 0.231340865, 0.27146747],
["surname", 1, 0.372351177, 0.109234086],
["surname", 2, 0.396307958, 0.619298443],
]
settings_obj = params.current_settings_obj
for r in rows:
cc = settings_obj.get_comparison_column(r[0])
level_dict = cc.level_as_dict(r[1])
assert level_dict["m_probability"] == pytest.approx(r[2])
assert level_dict["u_probability"] == pytest.approx(r[3])
################################################
# Test whether saving and loading params works
# (If we load params, does the expectation step yield same answer)
################################################
import tempfile
dir = tempfile.TemporaryDirectory()
fname = os.path.join(dir.name, "params.json")
df_e = run_expectation_step(df_gammas, params, spark)
params.save_model_to_json_file(fname)
from splink.model import load_model_from_json
p = load_model_from_json(fname)
df_e_2 = run_expectation_step(df_gammas, p, spark)
assert_frame_equal(df_e.toPandas(), df_e_2.toPandas())