Rewrite all uses of compile(x) to x._compile()

This was done using:

    find . -name '*.py' | xargs sed -i -r -e 's/ compile\((\w+)\)/ \1._compile()/g'

And then unnecessary imports were removed using:

    just fix

ehrql/loaders.py

@@ -9,7 +9,7 @@
 import ehrql
 from ehrql.debugger import activate_debug_context
 from ehrql.measures import Measures
-from ehrql.query_language import Dataset, compile, modify_exception
+from ehrql.query_language import Dataset, modify_exception
 from ehrql.renderers import DISPLAY_RENDERERS
 from ehrql.serializer import deserialize
 from ehrql.utils.traceback_utils import get_trimmed_traceback
@@ -282,7 +282,7 @@ def get_variable_definitions_from_module(module):
         raise DefinitionError(
             "A population has not been defined; define one with define_population()"
         )
-    return compile(dataset)
+    return dataset._compile()
 
 
 def load_measure_definitions_unsafe(definition_file, user_args, **kwargs):

ehrql/query_engines/sandbox.py

@@ -2,14 +2,14 @@
 
 from ehrql.query_engines.in_memory_database import apply_function
 from ehrql.query_engines.local_file import LocalFileQueryEngine
-from ehrql.query_language import Dataset, DateDifference, compile
+from ehrql.query_language import Dataset, DateDifference
 from ehrql.query_model.introspection import get_table_nodes
 from ehrql.query_model.nodes import AggregateByPatient, Function
 
 
 class SandboxQueryEngine(LocalFileQueryEngine):
     def evaluate_dataset(self, dataset_definition):
-        variable_definitions = compile(dataset_definition)
+        variable_definitions = dataset_definition._compile()
         if not variable_definitions:
             return EmptyDataset()
         table_nodes = get_table_nodes(*variable_definitions.values())

tests/conftest.py

@@ -12,7 +12,6 @@
 from ehrql.query_engines.mssql import MSSQLQueryEngine
 from ehrql.query_engines.sqlite import SQLiteQueryEngine
 from ehrql.query_engines.trino import TrinoQueryEngine
-from ehrql.query_language import compile
 from tests.lib.orm_utils import make_orm_models
 
 from .lib.databases import (
@@ -196,7 +195,7 @@ def query_engine(self, dsn=False, **engine_kwargs):
         return self.query_engine_class(dsn, **engine_kwargs)
 
     def extract(self, dataset, **engine_kwargs):
-        variable_definitions = compile(dataset)
+        variable_definitions = dataset._compile()
         return self.extract_qm(variable_definitions, **engine_kwargs)
 
     def extract_qm(self, variable_definitions, **engine_kwargs):
@@ -207,7 +206,7 @@ def extract_qm(self, variable_definitions, **engine_kwargs):
         return [row._asdict() for row in sorted(results)]
 
     def dump_dataset_sql(self, dataset, **engine_kwargs):
-        variable_definitions = compile(dataset)
+        variable_definitions = dataset._compile()
         query_engine = self.query_engine(dsn=None, **engine_kwargs)
         return get_sql_strings(query_engine, variable_definitions)
 

tests/integration/backends/test_emis.py

@@ -7,7 +7,6 @@
 from ehrql import create_dataset
 from ehrql.backends.emis import EMISBackend
 from ehrql.query_engines.base_sql import get_setup_and_cleanup_queries
-from ehrql.query_language import compile
 from ehrql.tables import PatientFrame, Series, emis, table_from_rows
 from ehrql.tables.raw import emis as emis_raw
 from ehrql.utils.sqlalchemy_query_utils import CreateTableAs, GeneratedTable
@@ -606,7 +605,7 @@ class t(PatientFrame):
         backend=backend,
     )
 
-    variables = compile(dataset)
+    variables = dataset._compile()
 
     results_query = query_engine.get_query(variables)
     inline_tables = [
@@ -662,7 +661,7 @@ def test_temp_table_includes_organisation_hash(trino_database):
         backend=backend,
     )
 
-    variables = compile(dataset)
+    variables = dataset._compile()
     results_query = query_engine.get_query(variables)
     temp_tables = [
         ch

tests/integration/query_engines/test_local_file.py

@@ -2,7 +2,6 @@
 
 from ehrql import Dataset
 from ehrql.query_engines.local_file import LocalFileQueryEngine
-from ehrql.query_language import compile
 from ehrql.tables import EventFrame, PatientFrame, Series, table
 
 
@@ -31,7 +30,7 @@ def test_local_file_query_engine():
     ).count_for_patient()
 
     dataset.define_population(patients.exists_for_patient())
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
 
     query_engine = LocalFileQueryEngine(FIXTURES)
     results = query_engine.get_results(variable_definitions)

tests/unit/dummy_data/test_generator.py

@@ -7,7 +7,7 @@
 from ehrql import Dataset
 from ehrql.dummy_data.generator import DummyDataGenerator, DummyPatientGenerator
 from ehrql.dummy_data.query_info import ColumnInfo, TableInfo
-from ehrql.query_language import compile, table_from_rows
+from ehrql.query_language import table_from_rows
 from ehrql.tables import Constraint, EventFrame, PatientFrame, Series, table
 
 
@@ -70,7 +70,7 @@ def test_dummy_data_generator():
     dataset.date = last_event.date
 
     # Generate some results
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
     generator = DummyDataGenerator(variable_definitions)
     generator.population_size = 7
     generator.batch_size = 4
@@ -97,7 +97,7 @@ def test_dummy_data_generator_timeout_with_some_results(patched_time):
     dataset = Dataset()
     dataset.define_population(patients.exists_for_patient())
 
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
     generator = DummyDataGenerator(variable_definitions)
     generator.population_size = 100
     generator.batch_size = 3
@@ -120,7 +120,7 @@ def test_dummy_data_generator_timeout_with_no_results(patched_time):
     dataset = Dataset()
     dataset.define_population(patients.sex != patients.sex)
 
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
     generator = DummyDataGenerator(variable_definitions)
     generator.timeout = 10
 
@@ -180,7 +180,7 @@ class inline_table(PatientFrame):
         )
 
     # Generate some results
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
     generator = DummyDataGenerator(variable_definitions)
     # We're asking for more results than we can possibly get (because there are only 6
     # patients in the inline table). We expect the attempt to timeout and just return 6
@@ -277,7 +277,7 @@ def test_get_random_int_with_range(dummy_patient_generator):
 def dummy_patient_generator():
     dataset = Dataset()
     dataset.define_population(patients.exists_for_patient())
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
     generator = DummyPatientGenerator(
         variable_definitions,
         random_seed="abc",

tests/unit/dummy_data/test_query_info.py

@@ -3,7 +3,6 @@
 from ehrql import Dataset, days
 from ehrql.codes import CTV3Code
 from ehrql.dummy_data.query_info import ColumnInfo, QueryInfo, TableInfo
-from ehrql.query_language import compile
 from ehrql.tables import (
     Constraint,
     EventFrame,
@@ -42,7 +41,7 @@ def test_query_info_from_variable_definitions():
         events.code == CTV3Code("abc00")
     ).exists_for_patient()
 
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
     query_info = QueryInfo.from_variable_definitions(variable_definitions)
 
     assert query_info == QueryInfo(
@@ -99,7 +98,7 @@ class test_table(PatientFrame):
         | test_table.value.contains("d")
     )
 
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
     query_info = QueryInfo.from_variable_definitions(variable_definitions)
     column_info = query_info.tables["test_table"].columns["value"]
 
@@ -127,7 +126,7 @@ class inline_table(PatientFrame):
         | inline_table.value.contains("d")
     )
 
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
     query_info = QueryInfo.from_variable_definitions(variable_definitions)
 
     assert query_info == QueryInfo(
@@ -150,7 +149,7 @@ def test_query_info_ignores_complex_comparisons():
     dataset.q1 = patients.date_of_birth.year.is_in([2000, 2010, 2020])
     dataset.q2 = patients.date_of_birth + days(100) == "2021-10-20"
     dataset.q3 = patients.date_of_birth == "2022-10-05"
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
 
     query_info = QueryInfo.from_variable_definitions(variable_definitions)
     column_info = query_info.tables["patients"].columns["date_of_birth"]

tests/unit/dummy_data_nextgen/test_generator.py

@@ -7,7 +7,7 @@
 from ehrql import Dataset
 from ehrql.dummy_data_nextgen.generator import DummyDataGenerator, DummyPatientGenerator
 from ehrql.dummy_data_nextgen.query_info import ColumnInfo, TableInfo
-from ehrql.query_language import compile, table_from_rows
+from ehrql.query_language import table_from_rows
 from ehrql.tables import Constraint, EventFrame, PatientFrame, Series, table
 
 
@@ -76,7 +76,7 @@ def test_dummy_data_generator():
     # Generate some results
     target_size = 7
 
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
     generator = DummyDataGenerator(variable_definitions, population_size=target_size)
     generator.batch_size = 4
     results = list(generator.get_results())
@@ -106,7 +106,7 @@ def test_dummy_data_generator_timeout_with_some_results(patched_time):
     dataset = Dataset()
     dataset.define_population(patients.exists_for_patient())
 
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
     generator = DummyDataGenerator(variable_definitions)
     generator.population_size = 100
     generator.batch_size = 3
@@ -132,7 +132,7 @@ def test_dummy_data_generator_timeout_with_no_results(patched_time):
         & (patients.date_of_death.day == 2)
     )
 
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
     generator = DummyDataGenerator(variable_definitions)
     generator.timeout = 10
 
@@ -192,7 +192,7 @@ class inline_table(PatientFrame):
         )
 
     # Generate some results
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
     generator = DummyDataGenerator(variable_definitions)
     # We're asking for more results than we can possibly get (because there are only 6
     # patients in the inline table). We expect the attempt to timeout and just return 6
@@ -336,7 +336,7 @@ def test_cannot_generate_data_outside_of_a_seed_block(dummy_patient_generator):
 def dummy_patient_generator():
     dataset = Dataset()
     dataset.define_population(patients.exists_for_patient())
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
     generator = DummyPatientGenerator(
         variable_definitions,
         random_seed="abc",

tests/unit/dummy_data_nextgen/test_query_info.py

@@ -3,7 +3,6 @@
 from ehrql import Dataset, days
 from ehrql.codes import CTV3Code
 from ehrql.dummy_data_nextgen.query_info import ColumnInfo, QueryInfo, TableInfo
-from ehrql.query_language import compile
 from ehrql.tables import (
     Constraint,
     EventFrame,
@@ -42,7 +41,7 @@ def test_query_info_from_variable_definitions():
         events.code == CTV3Code("abc00")
     ).exists_for_patient()
 
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
     query_info = QueryInfo.from_variable_definitions(variable_definitions)
 
     assert query_info == QueryInfo(
@@ -99,7 +98,7 @@ class test_table(PatientFrame):
         | test_table.value.contains("d")
     )
 
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
     query_info = QueryInfo.from_variable_definitions(variable_definitions)
     column_info = query_info.tables["test_table"].columns["value"]
 
@@ -123,7 +122,7 @@ class inline_table(PatientFrame):
         | inline_table.value.contains("d")
     )
 
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
     query_info = QueryInfo.from_variable_definitions(variable_definitions)
 
     assert query_info == QueryInfo(
@@ -146,7 +145,7 @@ def test_query_info_ignores_complex_comparisons():
     dataset.q1 = patients.date_of_birth.year.is_in([2000, 2010, 2020])
     dataset.q2 = patients.date_of_birth + days(100) == "2021-10-20"
     dataset.q3 = patients.date_of_birth == "2022-10-05"
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
 
     query_info = QueryInfo.from_variable_definitions(variable_definitions)
     column_info = query_info.tables["patients"].columns["date_of_birth"]

tests/unit/dummy_data_nextgen/test_specific_datasets.py

@@ -14,7 +14,6 @@
     EventFrame,
     PatientFrame,
     Series,
-    compile,
     table,
     table_from_rows,
 )
@@ -68,7 +67,7 @@ def test_queries_with_exact_one_shot_generation(patched_time, query):
 
     target_size = 1000
 
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
     generator = DummyDataGenerator(variable_definitions, population_size=target_size)
     generator.batch_size = target_size
     generator.timeout = 10
@@ -138,7 +137,7 @@ def test_combined_age_range_in_one_shot(patched_time, query, target_size):
 
     dataset.define_population(query)
 
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
     generator = DummyDataGenerator(variable_definitions, population_size=target_size)
     generator.batch_size = target_size
     generator.timeout = 10
@@ -221,7 +220,7 @@ def test_queries_not_yet_well_handled(patched_time, query):
     dataset.define_population(patients.exists_for_patient() & query)
 
     target_size = 1000
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
 
     generator = DummyDataGenerator(variable_definitions, population_size=target_size)
     generator.batch_size = target_size
@@ -251,7 +250,7 @@ def test_inline_table_query(patched_time):
     dataset.i = p.i
 
     target_size = 1000
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
 
     generator = DummyDataGenerator(variable_definitions, population_size=target_size)
     generator.batch_size = target_size
@@ -284,7 +283,7 @@ def test_will_raise_if_all_data_is_impossible(patched_time, query):
 
     dataset.define_population(query)
     target_size = 1000
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
 
     generator = DummyDataGenerator(variable_definitions, population_size=target_size)
     generator.timeout = 1
@@ -309,7 +308,7 @@ def test_generates_events_starting_from_birthdate():
 
     target_size = 1000
     dataset.configure_dummy_data(population_size=target_size)
-    variable_definitions = compile(dataset)
+    variable_definitions = dataset._compile()
 
     generator = DummyDataGenerator(variable_definitions, population_size=target_size)
 

tests/unit/test_query_language.py

@@ -102,7 +102,7 @@ def test_dataset():
     assert dataset.year_of_birth is year_of_birth
     assert dataset.dummy_data_config.population_size == 123
 
-    assert compile(dataset) == {
+    assert dataset._compile() == {
         "year_of_birth": Function.YearFromDate(
             source=SelectColumn(
                 name="date_of_birth",