Fix lambda inliner if parameter is not referenced (#1206)

This PR fixes a small bug in the lambda inliner such that parameters which are not referenced are correctly eliminated, i.e. now
```
(λ(x, y) → x)(x, y)
```
transforms to
```
x
```
Additionally tests for the inliner were introduced (didn't exist before).

src/gt4py/next/iterator/transforms/inline_lambdas.py

@@ -38,7 +38,7 @@ def inline_lambda(
         for i, param in enumerate(node.fun.params):
             # TODO(tehrengruber): allow inlining more complicated zero-op expressions like
             #  ignore_shift(...)(it_sym)  # noqa: E800
-            if ref_counts[param.id] != 1 and not isinstance(
+            if ref_counts[param.id] > 1 and not isinstance(
                 node.args[i], (ir.SymRef, ir.Literal, ir.OffsetLiteral)
             ):
                 eligible_params[i] = False

tests/next_tests/iterator_tests/test_inline_lambdas.py

@@ -0,0 +1,58 @@
+# GT4Py - GridTools Framework
+#
+# Copyright (c) 2014-2023, ETH Zurich
+# All rights reserved.
+#
+# This file is part of the GT4Py project and the GridTools framework.
+# GT4Py is free software: you can redistribute it and/or modify it under
+# the terms of the GNU General Public License as published by the
+# Free Software Foundation, either version 3 of the License, or any later
+# version. See the LICENSE.txt file at the top-level directory of this
+# distribution for a copy of the license or check <https://www.gnu.org/licenses/>.
+#
+# SPDX-License-Identifier: GPL-3.0-or-later
+
+import pytest
+
+from gt4py.next.ffront import itir_makers as im
+from gt4py.next.iterator.transforms.inline_lambdas import InlineLambdas
+
+
+test_data = [
+    ("identity1", im.call_(im.lambda__("x")("x"))("y"), im.ref("y")),
+    ("identity2", im.call_(im.lambda__("x")("x"))(im.plus_("y", "y")), im.plus_("y", "y")),
+    ("unused_param", im.call_(im.lambda__("x", "y")("x"))("x", "y"), im.ref("x")),
+    (
+        "composed_addition",
+        im.call_(im.lambda__("x")(im.plus_("x", "x")))(im.plus_("y", "y")),
+        {
+            True: im.call_(im.lambda__("x")(im.plus_("x", "x")))(im.plus_("y", "y")),  # stays as is
+            False: im.plus_(im.plus_("y", "y"), im.plus_("y", "y")),
+        },
+    ),
+    (
+        "name_collision",
+        im.call_(im.lambda__("x")(im.plus_("x", "x")))(im.plus_("x", "y")),
+        {
+            True: im.call_(im.lambda__("x")(im.plus_("x", "x")))(im.plus_("x", "y")),  # stays as is
+            False: im.plus_(im.plus_("x", "y"), im.plus_("x", "y")),
+        },
+    ),
+    (
+        "name_shadowing",
+        im.call_(
+            im.lambda__("x")(im.multiplies_(im.call_(im.lambda__("x")(im.plus_("x", 1)))(2), "x"))
+        )(im.plus_("x", "x")),
+        im.multiplies_(im.plus_(2, 1), im.plus_("x", "x")),
+    ),
+]
+
+
+@pytest.mark.parametrize("opcount_preserving", [True, False])
+@pytest.mark.parametrize("name,testee,expected", test_data)
+def test(name, opcount_preserving, testee, expected):
+    if isinstance(expected, dict):
+        expected = expected[opcount_preserving]
+
+    inlined = InlineLambdas.apply(testee, opcount_preserving=opcount_preserving)
+    assert inlined == expected