Add Fortran namelist parser using the new Lark-based infrastructure.

Author: micaeljtoliveira

pyproject.toml

@@ -14,7 +14,6 @@ classifiers = [
     "Topic :: Utilities",
 ]
 dependencies = [
-    "f90nml",
     "lark",
     "ruamel.yaml",
 ]

src/access/parsers/fortran_nml.py

@@ -0,0 +1,70 @@
+# Copyright 2025 ACCESS-NRI and contributors. See the top-level COPYRIGHT file for details.
+# SPDX-License-Identifier: Apache-2.0
+
+"""Parser for Fortran namelists.
+
+Fortran namelists allow format-free I/O of variables by key-value assignements. Initially they were an extension to the
+languages, but became part of the standard in Fortran 90.
+"""
+
+from access.parsers.config import ConfigParser
+
+
+class FortranNMLParser(ConfigParser):
+    """Fortran Namelist parser.
+
+    Note: Currently array qualifiers, substrings and derived types are not implemented in the grammar.
+    """
+
+    @property
+    def case_sensitive_keys(self) -> bool:
+        return False
+
+    @property
+    def grammar(self) -> str:
+        return """
+?start: namelists
+
+?namelists: random_text? namelist (random_text? namelist)* random_text?
+
+namelist.2: nml_start key line_end? nml_lines nml_end -> key_block
+nml_start: ws* "&"
+nml_end:  ws* ("/"|/&end/i) line_end
+nml_lines: (nml_line | empty_line)* -> block
+
+?nml_line: assignment (ws* "," assignment)* (ws* separator)? line_end?
+
+?assignment: key_value | key_list | key_null
+
+key_value: ws* key ws* "=" ws* value
+key_list: ws* key ws* "=" ws* value ((line_break|ws* separator) ws* value)+
+key_null: ws* key ws* "=" ws*
+line_break: ws* separator line_end
+
+?value: logical
+      | integer
+      | float
+      | double
+      | complex
+      | double_complex
+      | string
+
+empty_line: line_end
+line_end: (fortran_comment|ws*) NEWLINE
+separator: ","
+
+random_text: (/.+/|NEWLINE)*
+ANYTHING: /.+/
+
+%import config.key
+%import config.logical
+%import config.integer
+%import config.float
+%import config.double
+%import config.complex
+%import config.double_complex
+%import config.string
+%import config.fortran_comment
+%import config.ws
+%import config.NEWLINE
+"""

tests/test_fortran_nml.py

@@ -0,0 +1,185 @@
+# Copyright 2025 ACCESS-NRI and contributors. See the top-level COPYRIGHT file for details.
+# SPDX-License-Identifier: Apache-2.0
+
+import pytest
+from lark.exceptions import UnexpectedCharacters, UnexpectedEOF
+
+from access.parsers.fortran_nml import FortranNMLParser
+
+
+@pytest.fixture(scope="module")
+def parser():
+    """Fixture instantiating the parser."""
+    return FortranNMLParser()
+
+
+@pytest.fixture()
+def fortran_nml():
+    """Fixture returning a dict holding the parsed content of a Fortran namelist file."""
+    return {
+        "LIST_A": {
+            "VAR": 4,
+            "LIST": [1, 2, 3],
+            "FLOAT": 1800.0,
+            "COMPLEX": 3.0 + 4.0j,
+            "NOVALUE": None,
+        },
+        "LIST_B": {
+            "BOOL": True,
+            "LIST": ["a", "b", "c"],
+            "DOUBLE": 1.0e-10,
+            "STRING": "a string",
+            "ANOTHER_LIST": [1, 2, 3, 4, 5, 6],
+        },
+        "LIST_C": {},
+    }
+
+
+@pytest.fixture()
+def fortran_nml_file():
+    """Fixture returning the content of a Fortran namelist file."""
+    return """
+&LIST_A ! This is a comment
+  Var = 4 , ! This is a comment after an assignment
+  LIST = 1, 2, 3
+
+! This is another comment
+
+  Float = 1800.0  ,
+  COMPLEX = (3.0, 4.0),
+
+  NOVALUE =
+/
+
+! Yet another comment
+ This is some random text 
+
+&LIST_B
+  Bool = .true.
+  LIST = "a", "b", "c", DOUBLE = 1d-10
+  STRING="a string"
+  ANOTHER_LIST = 1, 2,
+                 3,   ! Comment in line break
+                  4, 5, 6
+/
+
+&LIST_C
+/
+"""
+
+
+@pytest.fixture()
+def modified_fortran_nml_file():
+    """Fixture returning the content of the previous Fortran namelist file, but with some modifications."""
+    return """
+&LIST_A ! This is a comment
+  Var = 6 , ! This is a comment after an assignment
+  LIST = 1, 2, 3
+
+! This is another comment
+
+  Float = 900.0  ,
+  COMPLEX = (3.0, 4.0),
+
+  NOVALUE =
+/
+
+! Yet another comment
+ This is some random text 
+
+&LIST_B
+  Bool = .false.
+  LIST = "a", "b", "c", DOUBLE = 1d-10
+  STRING="another string"
+  ANOTHER_LIST = 1, 2,
+                 3,   ! Comment in line break
+                  4, 5, 6
+/
+
+&LIST_C
+/
+"""
+
+
+def test_valid_fortran_nml(parser):
+    """Test the basic grammar constructs"""
+    assert dict(parser.parse("&LIST TEST='a'/")) == {"LIST": {"TEST": "a"}}
+    assert dict(parser.parse("&LIST\nTEST='a'/")) == {"LIST": {"TEST": "a"}}
+    assert dict(parser.parse("&LIST\nTEST='a'\n/")) == {"LIST": {"TEST": "a"}}
+    assert dict(parser.parse("&LIST\nTEST='a'\n&end")) == {"LIST": {"TEST": "a"}}
+    assert dict(parser.parse("&LIST\nTEST='a'\n&End")) == {"LIST": {"TEST": "a"}}
+    assert dict(parser.parse(" &LIST\nTEST='a'/")) == {"LIST": {"TEST": "a"}}
+    assert dict(parser.parse("&LIST\nTEST = 'a' /")) == {"LIST": {"TEST": "a"}}
+    assert dict(parser.parse("&LIST\nTEST = 'a'/")) == {"LIST": {"TEST": "a"}}
+    assert dict(parser.parse("&LIST\nTEST= 'a'/")) == {"LIST": {"TEST": "a"}}
+    assert dict(parser.parse("&LIST\nTEST='a',\n/")) == {"LIST": {"TEST": "a"}}
+    assert dict(parser.parse("&LIST\nTEST='a' , \n/")) == {"LIST": {"TEST": "a"}}
+    assert dict(parser.parse("&LIST\nTEST = .true.\n/")) == {"LIST": {"TEST": True}}
+    assert dict(parser.parse("&LIST\nTEST = .false.\n/")) == {"LIST": {"TEST": False}}
+    assert dict(parser.parse("&LIST\nTEST='a', 'b'\n/")) == {"LIST": {"TEST": ["a", "b"]}}
+    assert dict(parser.parse("&LIST\nTEST='a','b'\n/")) == {"LIST": {"TEST": ["a", "b"]}}
+    assert dict(parser.parse("&LIST\nTEST='a','b',\n/")) == {"LIST": {"TEST": ["a", "b"]}}
+    assert dict(parser.parse("&LIST\nTEST='a', \n'b', \n'c'\n/")) == {"LIST": {"TEST": ["a", "b", "c"]}}
+    assert dict(parser.parse("&LIST\nVAR1=1, VAR2=2\n/")) == {"LIST": {"VAR1": 1, "VAR2": 2}}
+    assert dict(parser.parse("&LIST\nVAR1=1, VAR2=2,\n/")) == {"LIST": {"VAR1": 1, "VAR2": 2}}
+    assert dict(parser.parse("&LIST\nVAR1=1, 2, VAR2=3\n/")) == {"LIST": {"VAR1": [1, 2], "VAR2": 3}}
+    assert dict(parser.parse("&LIST\nVAR1=1, VAR2=2, 3\n/")) == {"LIST": {"VAR1": 1, "VAR2": [2, 3]}}
+    assert dict(parser.parse("&LIST\nVAR1=1, 2, VAR2=3, 4\n/")) == {"LIST": {"VAR1": [1, 2], "VAR2": [3, 4]}}
+    assert dict(parser.parse("&LIST\nVAR1=1, VAR2=2, VAR3=3\n/")) == {"LIST": {"VAR1": 1, "VAR2": 2, "VAR3": 3}}
+    assert dict(parser.parse("&LIST\nVAR1=1, VAR2=2, VAR3=3,\n/")) == {"LIST": {"VAR1": 1, "VAR2": 2, "VAR3": 3}}
+    assert dict(parser.parse(" &LIST\nTEST = \n /")) == {"LIST": {"TEST": None}}
+
+
+def test_invalid_fortran_nml(parser):
+    """Test checking that the parser catches malformed expressions"""
+    with pytest.raises(UnexpectedEOF):
+        parser.parse("&LIST\nTEST : 'a'\n/")
+
+    with pytest.raises(UnexpectedEOF):
+        parser.parse("&LIST\nTEST = true\n/")
+
+    with pytest.raises(UnexpectedEOF):
+        parser.parse("&LIST\nTEST = false\n/")
+
+    with pytest.raises(UnexpectedEOF):
+        parser.parse("%TEST\na=1\n%TEST")
+
+    with pytest.raises(UnexpectedEOF):
+        parser.parse("TEST%\na=1\nTEST%")
+
+    with pytest.raises(UnexpectedEOF):
+        parser.parse("&TEST\na=1 2\n/")
+
+    with pytest.raises(UnexpectedEOF):
+        parser.parse("&TEST\na=1 \n2\n/")
+
+    with pytest.raises(UnexpectedEOF):
+        parser.parse("BLOCK\n TEST ='a'/")
+
+    with pytest.raises(UnexpectedEOF):
+        parser.parse("&BLOCK\n VAR1=1\n&e")
+
+
+def test_fortran_nml_parse(parser, fortran_nml, fortran_nml_file):
+    """Test parsing a file."""
+    config = parser.parse(fortran_nml_file)
+    assert dict(config) == fortran_nml
+
+
+def test_fortran_nml_roundtrip(parser, fortran_nml_file):
+    """Test round-trip parsing."""
+    config = parser.parse(fortran_nml_file)
+
+    assert str(config) == fortran_nml_file
+
+
+def test_fortran_nml_roundtrip_with_mutation(parser, fortran_nml_file, modified_fortran_nml_file):
+    """Test round-trip parsing with mutation of the config."""
+    config = parser.parse(fortran_nml_file)
+
+    config["LIST_A"]["VAR"] = 6
+    config["LIST_A"]["FLOAT"] = 900.0
+    config["LIST_B"]["BOOL"] = False
+    config["LIST_B"]["STRING"] = "another string"
+
+    assert str(config) == modified_fortran_nml_file