[datasets] Tighten the requirements of benchmark name regex.

Do not permit that the benchmark name can be missing for a benchmark
URI to be considered well formed, as we no longer support dataset-only
URIs.

Issue facebookresearch#45.

compiler_gym/datasets/benchmark.py

@@ -22,21 +22,19 @@
 # Regular expression that matches the full two-part URI prefix of a dataset:
 #     {{protocol}}://{{dataset}}
 #
-# A trailing slash is permitted.
+# An optional trailing slash is permitted.
 #
-# Example matches: "benchmark://foo-v0", "benchmark://foo-v0/".
+# Example matches: "benchmark://foo-v0", "generator://bar-v0/".
 DATASET_NAME_RE = re.compile(
     r"(?P<dataset>(?P<dataset_protocol>[a-zA-z0-9-_]+)://(?P<dataset_name>[a-zA-z0-9-_]+-v(?P<dataset_version>[0-9]+)))/?"
 )
 
 # Regular expression that matches the full three-part format of a benchmark URI:
 #     {{protocol}}://{{dataset}}/{{id}}
 #
-# The {{id}} is optional.
-#
-# Example matches: "benchmark://foo-v0/" or "benchmark://foo-v0/program".
+# Example matches: "benchmark://foo-v0/foo" or "generator://bar-v1/foo/bar.txt".
 BENCHMARK_URI_RE = re.compile(
-    r"(?P<dataset>(?P<dataset_protocol>[a-zA-z0-9-_]+)://(?P<dataset_name>[a-zA-z0-9-_]+-v(?P<dataset_version>[0-9]+)))(/(?P<benchmark_name>[^\s]*))?$"
+    r"(?P<dataset>(?P<dataset_protocol>[a-zA-z0-9-_]+)://(?P<dataset_name>[a-zA-z0-9-_]+-v(?P<dataset_version>[0-9]+)))/(?P<benchmark_name>[^\s]+)$"
 )
 
 

tests/datasets/benchmark_test.py

@@ -28,54 +28,58 @@ def _rgx_match(regex, groupname, string) -> str:
     return match.group(groupname)
 
 
-@pytest.mark.parametrize("regex", (DATASET_NAME_RE, BENCHMARK_URI_RE))
-def test_benchmark_uri_protocol(regex):
-    assert not regex.match("B?://cbench-v1/")  # Invalid characters
-    assert not regex.match("cbench-v1/")  # Missing protocol
-
+def test_benchmark_uri_protocol():
     assert (
-        _rgx_match(regex, "dataset_protocol", "benchmark://cbench-v1/") == "benchmark"
+        _rgx_match(DATASET_NAME_RE, "dataset_protocol", "benchmark://cbench-v1/")
+        == "benchmark"
     )
     assert (
-        _rgx_match(regex, "dataset_protocol", "Generator13://gen-v11/") == "Generator13"
+        _rgx_match(DATASET_NAME_RE, "dataset_protocol", "Generator13://gen-v11/")
+        == "Generator13"
     )
 
 
-def test_benchmark_uri_dataset():
-    assert not BENCHMARK_URI_RE.match("benchmark://cBench?v0/")  # Invalid character
-    assert not BENCHMARK_URI_RE.match("benchmark://cBench/")  # Missing version suffix
+def test_invalid_benchmark_uris():
+    # Invalid protocol
+    assert not DATASET_NAME_RE.match("B?://cbench-v1/")  # Invalid characters
+    assert not DATASET_NAME_RE.match("cbench-v1/")  # Missing protocol
+
+    # Invalid dataset name
+    assert not BENCHMARK_URI_RE.match("benchmark://cbench?v0/foo")  # Invalid character
+    assert not BENCHMARK_URI_RE.match(
+        "benchmark://cbench/foo"
+    )  # Missing version suffix
+    assert not BENCHMARK_URI_RE.match("benchmark://cbench-v0")  # Missing benchmark ID
+    assert not BENCHMARK_URI_RE.match("benchmark://cbench-v0/")  # Missing benchmark ID
+
+    # Invalid benchmark ID
+    assert not BENCHMARK_URI_RE.match("benchmark://cbench-v1/ whitespace")  # Whitespace
+    assert not BENCHMARK_URI_RE.match("benchmark://cbench-v1/\t")  # Whitespace
+
 
+def test_benchmark_uri_dataset():
     assert (
-        _rgx_match(BENCHMARK_URI_RE, "dataset_name", "benchmark://cbench-v1/")
+        _rgx_match(BENCHMARK_URI_RE, "dataset_name", "benchmark://cbench-v1/foo")
         == "cbench-v1"
     )
     assert (
-        _rgx_match(BENCHMARK_URI_RE, "dataset_name", "Generator13://gen-v11/")
+        _rgx_match(BENCHMARK_URI_RE, "dataset_name", "Generator13://gen-v11/foo")
         == "gen-v11"
     )
 
 
 def test_benchmark_dataset_name():
     assert (
-        _rgx_match(BENCHMARK_URI_RE, "dataset", "benchmark://cbench-v1/")
+        _rgx_match(BENCHMARK_URI_RE, "dataset", "benchmark://cbench-v1/foo")
         == "benchmark://cbench-v1"
     )
     assert (
-        _rgx_match(BENCHMARK_URI_RE, "dataset", "Generator13://gen-v11/")
+        _rgx_match(BENCHMARK_URI_RE, "dataset", "Generator13://gen-v11/foo")
         == "Generator13://gen-v11"
     )
 
 
 def test_benchmark_uri_id():
-    assert not BENCHMARK_URI_RE.match("benchmark://cbench-v1/ whitespace")  # Whitespace
-    assert not BENCHMARK_URI_RE.match("benchmark://cbench-v1/\t")  # Whitespace
-
-    assert (
-        _rgx_match(BENCHMARK_URI_RE, "benchmark_name", "benchmark://cbench-v1") is None
-    )
-    assert (
-        _rgx_match(BENCHMARK_URI_RE, "benchmark_name", "benchmark://cbench-v1/") == ""
-    )
     assert (
         _rgx_match(BENCHMARK_URI_RE, "benchmark_name", "benchmark://cbench-v1/foo")
         == "foo"
@@ -84,6 +88,12 @@ def test_benchmark_uri_id():
         _rgx_match(BENCHMARK_URI_RE, "benchmark_name", "benchmark://cbench-v1/foo/123")
         == "foo/123"
     )
+    assert (
+        _rgx_match(
+            BENCHMARK_URI_RE, "benchmark_name", "benchmark://cbench-v1/foo/123.txt"
+        )
+        == "foo/123.txt"
+    )
     assert (
         _rgx_match(
             BENCHMARK_URI_RE,