minimal lark parser

Author: wpbonelli

flopy4/mf6/__init__.py

@@ -1,3 +1,9 @@
+from flopy4.mf6 import (  # noqa: F401
+    gwf,
+    ims,
+    simulation,
+    tdis,
+)
 from flopy4.mf6.codec import dump
 from flopy4.mf6.component import Component
 from flopy4.mf6.converter import COMPONENT_CONVERTER

flopy4/mf6/codec/reader/__init__.py

@@ -1,12 +1,46 @@
 from os import PathLike
+from pathlib import Path
 from typing import Any
 
+from flopy4.mf6.codec.reader.parser import make_generic_parser
+from flopy4.mf6.codec.reader.transformer import GenericTransformer
+
 
 def load(path: str | PathLike) -> Any:
-    # TODO
-    pass
+    """
+    Load and parse an MF6 input file.
+
+    Parameters
+    ----------
+    path : str | PathLike
+        Path to the MF6 input file
+
+    Returns
+    -------
+    Any
+        Parsed MF6 input file structure
+    """
+    path = Path(path)
+    with open(path, "r") as f:
+        data = f.read()
+    return loads(data)
 
 
 def loads(data: str) -> Any:
-    # TODO
-    pass
+    """
+    Parse MF6 input file content from string.
+
+    Parameters
+    ----------
+    data : str
+        MF6 input file content as string
+
+    Returns
+    -------
+    Any
+        Parsed MF6 input file structure
+    """
+
+    parser = make_generic_parser()
+    transformer = GenericTransformer()
+    return transformer.transform(parser.parse(data))

flopy4/mf6/codec/reader/grammar/__init__.py

@@ -0,0 +1,42 @@
+from os import PathLike
+from pathlib import Path
+
+import jinja2
+from modflow_devtools.dfn import Dfn
+
+from .filters import get_block_variables, get_blocks, get_variables, is_recarray_block
+
+
+def _get_template_env():
+    loader = jinja2.FileSystemLoader(Path(__file__).parent / "templates")
+    env = jinja2.Environment(
+        loader=loader,
+        trim_blocks=True,
+        lstrip_blocks=True,
+        keep_trailing_newline=True,
+    )
+    env.filters["is_recarray_block"] = is_recarray_block
+    env.filters["get_block_variables"] = get_block_variables
+    return env
+
+
+_TEMPLATE_ENV = _get_template_env()
+
+
+def make_grammar(dfn: Dfn, outdir: PathLike):
+    """Generate a Lark grammar file for a single component."""
+    outdir = Path(outdir).expanduser().resolve().absolute()
+    template = _TEMPLATE_ENV.get_template("component.lark.j2")
+    blocks = get_blocks(dfn)
+    variables = get_variables(dfn)
+    target_path = outdir / f"{dfn['name'].replace('-', '')}.lark"
+    with open(target_path, "w") as f:
+        f.write(template.render(component=dfn["name"], blocks=blocks, variables=variables))
+
+
+def make_all_grammars(dfns: dict[str, Dfn], outdir: PathLike):
+    """Generate grammars for all components."""
+    outdir = Path(outdir).expanduser().resolve().absolute()
+    outdir.mkdir(parents=True, exist_ok=True)
+    for dfn in dfns.values():
+        make_grammar(dfn, outdir)

flopy4/mf6/codec/reader/grammar/mf6.lark

@@ -0,0 +1,19 @@
+start: [WS] [_NL*] (block [[WS] _NL*])+ [WS]
+block: "begin"i block_name _NL _content "end"i block_name _NL+
+block_name: CNAME [INT]
+_content: line*
+line: [WS] item+ _NL+
+item: word | NUMBER
+word: /[a-zA-Z0-9._'~,-\\(\\)]+/
+
+%import common.NEWLINE -> _NL
+%import common.WS
+%import common.WS_INLINE
+%import common.CNAME
+%import common.WORD
+%import common.NUMBER
+%import common.INT
+%import common.SH_COMMENT
+%import common._STRING_INNER
+%ignore WS_INLINE
+%ignore SH_COMMENT

flopy4/mf6/codec/reader/parser.py

@@ -0,0 +1,10 @@
+from pathlib import Path
+
+from lark import Lark
+
+
+def make_generic_parser() -> Lark:
+    grammar_path = Path(__file__).parent / "grammar" / "mf6.lark"
+    with open(grammar_path, "r") as f:
+        grammar = f.read()
+    return Lark(grammar, parser="lalr", debug=True)

flopy4/mf6/codec/reader/transformer.py

@@ -0,0 +1,52 @@
+from typing import Any
+
+from lark import Token, Transformer
+
+
+class GenericTransformer(Transformer):
+    """
+    Generic transformer for MF6 input files with the generic MF6 grammar.
+
+    This transformer works with the generic grammar to parse any MF6 input
+    file into structures of blocks each consisting only of lines of tokens.
+    """
+
+    def start(self, items: list[Any]) -> dict[str, Any]:
+        blocks = {}
+        for item in items:
+            if not isinstance(item, dict):
+                continue
+            block_name = next(iter(item.keys()))
+            blocks[block_name] = next(iter(item.values()))
+        return blocks
+
+    def block(self, items: list[Any]) -> dict[str, Any]:
+        return {items[0]: items[1 : (len(items) - 1)]}
+
+    def block_name(self, items: list[Any]) -> str:
+        return " ".join([str(item) for item in items if item is not None])
+
+    def line(self, items: list[Any]) -> list[Any]:
+        return items[1:]
+
+    def item(self, items: list[Any]) -> str | float | int:
+        return items[0]
+
+    def word(self, items: list[Token]) -> str:
+        return str(items[0])
+
+    def NUMBER(self, token: Token) -> int | float:
+        value = str(token)
+        try:
+            if "." in value or "e" in value.lower():
+                return float(value)
+            else:
+                return int(value)
+        except ValueError:
+            return float(value)
+
+    def CNAME(self, token: Token) -> str:
+        return str(token)
+
+    def INT(self, token: Token) -> int:
+        return int(token)

flopy4/mf6/gwf/dis.py

@@ -14,8 +14,8 @@
 @xattree
 class Dis(Package):
     length_units: str = field(
+        block="options",
         default=None,
-        metadata={"block": "options"},
     )
     nogrb: bool = field(block="options", default=False)
     xorigin: float = field(block="options", default=None)

pyproject.toml

@@ -38,6 +38,7 @@ dependencies = [
     "cattrs",
     "flopy",
     "Jinja2>=3.0",
+    "lark",
     "numpy>=1.20.3",
     "pandas>=2.0.0",
     "toml>=0.10",