Merge pull request #1645 from GreenfishK/translate_algebra_tests

added three tests to cover changes made by the pull request #1361

rdflib/plugins/sparql/algebra.py

@@ -847,9 +847,15 @@ def find_nth(haystack, needle, n):
         with open("query.txt", "w") as file:
             file.write(filedata)
 
+    aggr_vars = collections.defaultdict(list)  # type: dict
+
     def convert_node_arg(node_arg):
         if isinstance(node_arg, Identifier):
-            return node_arg.n3()
+            if node_arg in aggr_vars.keys():
+                grp_var = aggr_vars[node_arg].pop(0).n3()
+                return grp_var
+            else:
+                return node_arg.n3()
         elif isinstance(node_arg, CompValue):
             return "{" + node_arg.name + "}"
         elif isinstance(node_arg, Expr):
@@ -959,7 +965,9 @@ def sparql_query_text(node):
                         raise ExpressionNotCoveredException(
                             "This expression might not be covered yet."
                         )
-                    agg_func_name = agg_func.name.split("_")[1]
+                    aggr_vars[agg_func.res].append(agg_func.vars)
+
+                    agg_func_name = agg_func.name.split('_')[1]
                     distinct = ""
                     if agg_func.distinct:
                         distinct = agg_func.distinct + " "
@@ -1022,7 +1030,7 @@ def sparql_query_text(node):
                     if isinstance(c.expr, Identifier):
                         var = c.expr.n3()
                         if c.order is not None:
-                            cond = var + "(" + c.order + ")"
+                            cond = c.order + "(" + var + ")"
                         else:
                             cond = var
                         order_conditions.append(cond)
@@ -1233,7 +1241,7 @@ def sparql_query_text(node):
                     "{Builtin_STRLEN}", "STRLEN(" + convert_node_arg(node.arg) + ")"
                 )
             elif node.name.endswith("Builtin_SUBSTR"):
-                args = [node.arg.n3(), node.start]
+                args = [convert_node_arg(node.arg), node.start]
                 if node.length:
                     args.append(node.length)
                 expr = "SUBSTR(" + ", ".join(args) + ")"
@@ -1294,7 +1302,7 @@ def sparql_query_text(node):
                 )
             elif node.name.endswith("Builtin_CONCAT"):
                 expr = "CONCAT({vars})".format(
-                    vars=", ".join(elem.n3() for elem in node.arg)
+                    vars=", ".join(convert_node_arg(elem) for elem in node.arg)
                 )
                 replace("{Builtin_CONCAT}", expr)
             elif node.name.endswith("Builtin_LANGMATCHES"):

test/test_translate_algebra.py

@@ -118,6 +118,16 @@ def _format_query(query: str) -> str:
         "test_graph_patterns__group",
         'Test if "group" gets properly translated into the query text.',
     ),
+    AlgebraTest(
+        "test_graph_patterns__group_and_substr",
+        'Test if a query with a variable that is used in the "GROUP BY" clause '
+        'and in the SUBSTR function gets properly translated into the query text.',
+    ),
+    AlgebraTest(
+        "test_graph_patterns__group_and_nested_concat",
+        'Test if a query with a nested concat expression in the select clause which '
+        'uses a group variable gets properly translated into the query text.',
+    ),
     AlgebraTest(
         "test_graph_patterns__having",
         'Test if "having" gets properly translated into the query text.',
@@ -266,26 +276,30 @@ def test_all_files_used(data_path: Path) -> None:
 @pytest.mark.parametrize("test_spec", [test.pytest_param() for test in algebra_tests])
 def test_roundtrip(test_spec: AlgebraTest, data_path: Path) -> None:
     """
-    Query remains the same over two successuive parse and translate cycles.
+    Query remains the same over two successive parse and translate cycles.
     """
     query_text = (data_path / test_spec.filename).read_text()
 
     logging.info("checking expectation: %s", test_spec.description)
     query_tree = parser.parseQuery(query_text)
     query_algebra = algebra.translateQuery(query_tree)
     query_from_algebra = translateAlgebra(query_algebra)
+    logging.debug(
+        "query_from_query_from_algebra = \n%s",
+        query_from_algebra,
+    )
 
     query_tree_2 = parser.parseQuery(query_from_algebra)
     query_algebra_2 = algebra.translateQuery(query_tree_2)
     query_from_query_from_algebra = translateAlgebra(query_algebra_2)
     logging.debug(
         "query_from_query_from_algebra = \n%s",
-        _format_query(query_from_query_from_algebra),
+        query_from_query_from_algebra,
     )
     assert (
         query_from_algebra == query_from_query_from_algebra
     ), f"failed expectation: {test_spec.description}"
 
-    # TODO: Execute the raw query (query_text) and the reconstitued query
+    # TODO: Execute the raw query (query_text) and the reconstituted query
     # (query_from_query_from_algebra) against a well defined graph and ensure
     # they yield the same result.

test/translate_algebra/test_functions__functions_on_strings.txt

@@ -12,6 +12,7 @@ select
 (concat("foo"@en, "bar"@en) as ?concat)
 (replace("abcd", "b", "Z") as ?replace)
 (regex(substr("foobar", 4), "bar", "bar") as ?regex)
+(regex(substr("foobar", 4, 1), "b", "b") as ?regex2)
 where {
 	?s ?p ?o .
     FILTER langMatches(lang(?o), "EN" )

test/translate_algebra/test_graph_patterns__group_and_nested_concat.txt

@@ -0,0 +1,7 @@
+PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
+
+SELECT
+(concat(str(?a), concat("foo"@en, "bar"@en)) as ?nested_concat)
+WHERE {
+  ?a rdf:value ?val .
+} GROUP BY ?a

test/translate_algebra/test_graph_patterns__group_and_substr.txt

@@ -0,0 +1,9 @@
+PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
+
+SELECT
+(SUM(?val) AS ?sum)
+(COUNT(?a) AS ?count)
+(SUBSTR(?a, 4, 1) as ?a_substr)
+WHERE {
+  ?a rdf:value ?val .
+} GROUP BY ?a

tox.ini

@@ -45,3 +45,9 @@ setenv =
     PYTHONHASHSEED = 0
 commands =
     sphinx-build -n -T -b html -d {envtmpdir}/doctrees docs docs/_build/html
+
+[pytest]
+# log_cli = true
+# log_cli_level = DEBUG
+log_cli_format = %(asctime)s [%(levelname)8s] %(message)s (%(filename)s:%(lineno)s)
+log_cli_date_format=%Y-%m-%d %H:%M:%S