add support for '0=contains()' in ADQL

Author: kimakan

src/queryparser/adql/ADQLParser.g4

@@ -41,6 +41,7 @@ comp_op:                        EQ | NOT_EQ | LTH | GTH | GRET | LEET ;
 comparison_predicate:           value_expression comp_op value_expression ;
 concatenation_operator:         CONCAT ;
 contains:                       CONTAINS LPAREN geometry_value_expression COMMA geometry_value_expression RPAREN ;
+contains_predicate:             INT EQ contains;
 coord_sys:                      string_value_expression ;
 coord_value:                    point_value | column_reference ;
 coord1:                         COORD1 LPAREN coord_value RPAREN ;
@@ -136,7 +137,8 @@ polygon:                        POLYGON LPAREN
                                 coordinates COMMA
                                 coordinates ( COMMA coordinates )+ RPAREN ;
 predicate:
-          comparison_predicate
+          contains_predicate
+        | comparison_predicate
         | between_predicate
         | in_predicate
         | like_predicate

src/queryparser/adql/adqltranslator.py

@@ -299,29 +299,6 @@ def visitPolygon(self, ctx):
         self.contexts[ctx] = ctx_text
 
 
-class ADQLContainsVisitor(ADQLParserVisitor):
-    def __init__(self):
-        self.contains = []
-
-    def visitContains(self, ctx):
-        if ctx.getText().lower()[:8] == 'contains':
-            self.contains.append(ctx)
-
-
-class ADQLComparisonPredicateVisitor(ADQLParserVisitor):
-    # Get rid of "1=" in 1=CONTAINS() statement
-    def visitComparison_predicate(self, ctx):
-        contains_visitor = ADQLContainsVisitor()
-        contains_visitor.visit(ctx)
-        if len(contains_visitor.contains):
-            if ctx.children[0].getText().lower()[:8] == 'contains':
-                ctx.children[1].removeLastChild()
-                ctx.children[2].removeLastChild()
-            elif ctx.children[2].getText().lower()[:8] == 'contains':
-                ctx.children[0].removeLastChild()
-                ctx.children[1].removeLastChild()
-
-
 class ADQLFunctionsTranslationVisitor(ADQLParserVisitor):
     """
     Run this visitor after the geometry has already been processed.
@@ -371,6 +348,23 @@ def visitCentroid(self, ctx):
         _remove_children(ctx)
         self.contexts[ctx] = ctx_text
 
+
+    def visitContains_predicate(self, ctx):
+        comp_value = ctx.children[0].getText()
+        if comp_value == '1' or comp_value == '0':
+            self.visitContains(ctx.children[2])
+            ctx_text = self.contexts[ctx.children[2]]
+            if self.output_sql == 'mysql':
+                ctx_text = f"{comp_value} = {ctx_text}"
+            elif self.output_sql == 'postgresql' and comp_value == '0':
+                ctx_text = ctx_text.replace('@', '!@')
+        else:
+            raise QueryError('The function CONTAINS allows comparison to 1 or 0 only.')
+
+        _remove_children(ctx)
+        self.contexts[ctx] = ctx_text
+
+
     def visitContains(self, ctx):
         arg = (self.contexts[ctx.children[2].children[0]],
                self.contexts[ctx.children[4].children[0]])
@@ -646,9 +640,6 @@ def to_postgresql(self):
 
         self.parse()
 
-        comp_visitor = ADQLComparisonPredicateVisitor()
-        comp_visitor.visit(self.tree)
-
         translator_visitor = ADQLGeometryTranslationVisitor(
                 output_sql='postgresql')
         translator_visitor.visit(self.tree)

src/queryparser/testing/test_postgresql.py

@@ -1,6 +1,6 @@
 # -*- coding: utf-8 -*-
 
-from .utils import _test_parsing, _test_indexed_adql_translation
+from .utils import _test_parsing
 
 from queryparser.postgresql import PostgreSQLQueryProcessor
 from queryparser.exceptions import QueryError, QuerySyntaxError
@@ -35,7 +35,3 @@ def test_postrgresql_query(t):
     with pytest.raises(QueryError):
         PostgreSQLQueryProcessor(t)
 
-
-@pytest.mark.parametrize("t", tests['indexed_adql_postgres_tests'])
-def test_indexed_adql_postrgresql_query(t):
-    _test_indexed_adql_translation(t + ['postgresql'])

src/queryparser/testing/tests.yaml

@@ -773,24 +773,6 @@ postgresql_tests:
         - ['gdr2.gaia_source']
 
 
-# Each test below consists of:
-#
-# - ADQL query string
-# - translated query string with replaced indexed object
-# - dictionary of indexed objects to be replaced and their specifications
-
-indexed_adql_postgres_tests:
-    -
-        - SELECT ra FROM gdr2.gaia_source AS gaia
-          WHERE 1=CONTAINS(POINT('ICRS', gaia.ra, gaia.dec), 
-                           CIRCLE('ICRS', 245.8962, -26.5222, 0.5));
-        - SELECT ra FROM gdr2.gaia_source AS gaia WHERE gaia.pos @ scircle(spoint(RADIANS(245.8962), RADIANS(-26.5222)), RADIANS(0.5));
-        - spoint:
-          -
-            - !!python/tuple ['gdr2', 'gaia_source', 'ra']
-            - !!python/tuple ['gdr2', 'gaia_source', 'dec']
-            - 'pos'
-
 
 # Each test below consists of:
 #
@@ -814,6 +796,10 @@ adql_mysql_tests:
         - SELECT TOP 10 ra, dec FROM db.tab WHERE 1=CONTAINS(POINT('ICRS', ra, dec), BOX('ICRS', -3.0, 5.0, 4.0, 10.0));
         - SELECT `ra`, `dec` FROM `db`.`tab` WHERE 1 = srcontainsl(spoint(RADIANS(`ra`), RADIANS(`dec`)), sbox(spoint(RADIANS(-5.000000000000),RADIANS(0.000000000000)),spoint(RADIANS(-1.000000000000),RADIANS(10.000000000000)))) LIMIT 10;
 
+    -
+        - SELECT TOP 10 ra, dec FROM db.tab WHERE 0=CONTAINS(POINT('ICRS', ra, dec), BOX('ICRS', -3.0, 5.0, 4.0, 10.0));
+        - SELECT `ra`, `dec` FROM `db`.`tab` WHERE 0 = srcontainsl(spoint(RADIANS(`ra`), RADIANS(`dec`)), sbox(spoint(RADIANS(-5.000000000000),RADIANS(0.000000000000)),spoint(RADIANS(-1.000000000000),RADIANS(10.000000000000)))) LIMIT 10;
+
     -
         - SELECT TOP 10 DISTANCE(POINT('ICRS', ra, dec), POINT('ICRS', 13.66, -58.3)) FROM db.tab;
         - SELECT DEGREES(sdist(spoint(RADIANS(`ra`), RADIANS(`dec`)), spoint(RADIANS(13.66), RADIANS(-58.3)))) FROM `db`.`tab` LIMIT 10;
@@ -874,6 +860,14 @@ adql_postgresql_tests:
         - SELECT TOP 10 ra, dec FROM db.tab WHERE 1=CONTAINS(POINT(ra, dec), BOX(-3.0, 5.0, 4.0, 10.0));
         - SELECT ra, dec FROM db.tab WHERE spoint(RADIANS(ra), RADIANS(dec)) @ sbox(spoint(RADIANS(-5.000000000000),RADIANS(0.000000000000)),spoint(RADIANS(-1.000000000000),RADIANS(10.000000000000))) LIMIT 10;
 
+    -
+        - SELECT TOP 10 ra, dec FROM db.tab WHERE 1=CONTAINS(POINT(ra, dec), CIRCLE(POINT(-3.0, 4.0), 10.0));
+        - SELECT ra, dec FROM db.tab WHERE spoint(RADIANS(ra), RADIANS(dec)) @ scircle(spoint(RADIANS(-3.0), RADIANS(4.0)), RADIANS(10.0)) LIMIT 10;
+
+    -
+        - SELECT TOP 10 ra, dec FROM db.tab WHERE 0=CONTAINS(POINT(ra, dec), CIRCLE(POINT(-3.0, 4.0), 10.0));
+        - SELECT ra, dec FROM db.tab WHERE spoint(RADIANS(ra), RADIANS(dec)) !@ scircle(spoint(RADIANS(-3.0), RADIANS(4.0)), RADIANS(10.0)) LIMIT 10;
+
     -
         - SELECT TOP 10 LOG10(ra), LOG(dec) FROM db.tab WHERE 1=CONTAINS(POINT('ICRS', ra, dec), POLYGON('ICRS', 10.0, -10.5, 20.0, 20.5, 30.0, 30.5));
         - SELECT LOG(ra), LN(dec) FROM db.tab WHERE spoint(RADIANS(ra), RADIANS(dec)) @ spoly('{(10.0d,-10.5d),(20.0d,20.5d),(30.0d,30.5d)}') LIMIT 10;

src/queryparser/testing/utils.py

@@ -51,16 +51,3 @@ def _test_parsing(query_processor, test, translate=False):
     if tables is not None:
         assert set(tables) == set(qp_tables)
 
-
-
-def _test_indexed_adql_translation(test):
-    query, translated_query, iob, output = test
-    adt = ADQLQueryTranslator(query)
-
-    if translated_query is not None:
-        if output == 'postgresql':
-            tq = adt.to_postgresql()
-            qp = PostgreSQLQueryProcessor()
-            qp.set_query(tq)
-            qp.process_query(indexed_objects=iob)
-            assert translated_query.strip() == qp.query