Implement representer script

bin/run.py

@@ -0,0 +1,147 @@
+#!/usr/bin/env python3
+
+import argparse
+import json
+import pathlib
+import sys
+
+import gdtoolkit.parser
+from lark import Token, Tree
+
+def load_config(input_path: pathlib.Path) -> dict | None:
+    config_path = input_path / ".meta/config.json"
+    if not config_path.exists():
+        return None
+    return json.loads((input_path / ".meta/config.json").read_text())
+
+
+def solution_files(slug: str, input_path: pathlib.Path) -> list[pathlib.Path]:
+    config = load_config(input_path)
+    if config:
+        files = config.get("files", {}).get("solution", [])
+        if files:
+            return [input_path / file for file in files]
+    return [(input_path / slug.replace("-", "_")).with_suffix(".gd")]
+
+
+def load_ast(slug: str, input_path: pathlib.Path) -> list[Tree]:
+    return [
+        gdtoolkit.parser.parser.parse(file.read_text())
+        for file in solution_files(slug, input_path)
+    ]
+
+
+def drop_docstr(trees: list[Tree]) -> list[Tree]:
+    def skip(t):
+        if isinstance(t, Tree) and t.data == "docstr_stmt":
+            return True
+        if (
+            isinstance(t, Tree)
+            and t.data == "expr_stmt"
+            and len(t.children) == 1
+            and isinstance(t.children[0], Tree)
+            and t.children[0].data == "expr"
+            and len(t.children[0].children) == 1
+            and isinstance(t.children[0].children[0], Tree)
+            and t.children[0].children[0].data == "string"
+        ):
+            return True
+        return False
+
+    def translate(t):
+        if isinstance(t, Token):
+            return t
+        elif isinstance(t, Tree):
+            return Tree(t.data, [translate(i) for i in t.children if not skip(i)])
+        else:
+            raise ValueError(t)
+
+    return [translate(tree) for tree in trees]
+
+
+def replace_func_names(trees: list[Tree]) -> tuple[list[Tree], dict[str, str]]:
+    varmap = {}
+    for tree in trees:
+        func_names = [
+            next(i.find_token("NAME"))
+            for i in tree.find_data(Token("RULE", "func_header"))
+        ]
+        for var in func_names:
+            if var.value not in varmap:
+                varmap[var.value] = f"FUNC_{len(varmap) + 1:03d}"
+
+    def translate(t):
+        if isinstance(t, Token):
+            return Token(t.type, varmap.get(t.value, t.value))
+        elif isinstance(t, Tree):
+            return Tree(t.data, [translate(i) for i in t.children])
+        else:
+            raise ValueError(t)
+
+    revmap = {v: k for k, v in varmap.items()}
+
+    trees = [translate(tree) for tree in trees]
+    return trees, revmap
+
+
+def replace_var_names(trees: list[Tree]) -> tuple[list[Tree], dict[str, str]]:
+    varmap = {}
+    for tree in trees:
+        func_vars = [
+            next(i.find_token("NAME"))
+            for token_type in ["func_var_stmt", "func_arg_typed", "func_arg_regular"]
+            for i in tree.find_data(Token("RULE", token_type))
+        ]
+        func_vars.sort(key=lambda x: x.line)
+        for var in func_vars:
+            if var.value not in varmap:
+                varmap[var.value] = f"FUNCVAR_{len(varmap) + 1:03d}"
+
+    def translate(t):
+        if isinstance(t, Token):
+            return Token(t.type, varmap.get(t.value, t.value))
+        elif isinstance(t, Tree):
+            return Tree(t.data, [translate(i) for i in t.children])
+        else:
+            raise ValueError(t)
+
+    revmap = {v: k for k, v in varmap.items()}
+
+    trees = [translate(tree) for tree in trees]
+    return trees, revmap
+
+
+def write_output(output_path: pathlib.Path, trees: list[Tree], varmap: dict[str, str]) -> None:
+    if not output_path.exists():
+        output_path.mkdir(parents=Tree, exist_ok=True)
+
+    (output_path / "representation.json").write_text(json.dumps({"version": 1}, indent=2) + "\n")
+    (output_path / "mapping.json").write_text(json.dumps(varmap, indent=2, sort_keys=True) + "\n")
+    (output_path / "representation.txt").write_text("\n\n".join(tree.pretty() for tree in trees))
+
+
+
+def arg_parser() -> argparse.ArgumentParser:
+    parser = argparse.ArgumentParser()
+    parser.add_argument("slug")
+    parser.add_argument("input_dir")
+    parser.add_argument("output_dir")
+    return parser
+
+
+def main() -> None:
+    args = arg_parser().parse_args()
+    input_path = pathlib.Path(args.input_dir)
+    output_path = pathlib.Path(args.output_dir)
+
+    print(f"== Creating representation for {args.slug!r} ==")
+    trees = load_ast(args.slug, input_path)
+    trees = drop_docstr(trees)
+    trees, varmap = replace_var_names(trees)
+    trees, funcmap = replace_func_names(trees)
+    write_output(output_path, trees, varmap | funcmap)
+    print(f"== Done with {args.slug!r} ==")
+
+
+if __name__ == "__main__":
+    main()

bin/run.sh

@@ -1,4 +1,4 @@
-#!/usr/bin/env bash
+#!/usr/bin/env sh
 
 # Synopsis:
 # Run the representer on a solution.
@@ -15,30 +15,4 @@
 # Example:
 # ./bin/run.sh two-fer /absolute/path/to/two-fer/solution/folder/ /absolute/path/to/output/directory/
 
-# If any required arguments is missing, print the usage and exit
-if [ -z "$1" ] || [ -z "$2" ] || [ -z "$3" ]; then
-    echo "usage: ./bin/run.sh exercise-slug /absolute/path/to/two-fer/solution/folder/ /absolute/path/to/output/directory/"
-    exit 1
-fi
-
-slug="$1"
-input_dir="${2%/}"
-output_dir="${3%/}"
-representation_txt_file="${output_dir}/representation.txt"
-representation_json_file="${output_dir}/representation.json"
-mapping_file="${output_dir}/mapping.json"
-
-# Create the output directory if it doesn't exist
-mkdir -p "${output_dir}"
-
-echo "${slug}: creating representation..."
-
-# Create the representation for the solution
-# TODO: replace the below commands with your own commands 
-# to create the representation.txt, representation.json and
-# mapping.json files
-echo '' > "${representation_txt_file}"
-echo '{"version": 1}' > "${representation_json_file}"
-echo '{}' > "${mapping_file}"
-
-echo "${slug}: done"
+python3 "${0%.*}.py" "$@"

tests/multiple-solution-files/.meta/config.json

@@ -0,0 +1,13 @@
+{
+  "authors": [
+    "person"
+  ],
+  "files": {
+    "solution": [
+      "one.gd",
+      "two.gd"
+    ],
+    "test": [],
+    "example": []
+  }
+}

tests/multiple-solution-files/expected_mapping.json

@@ -1 +1,4 @@
-{}
+{
+  "FUNC_001": "foo",
+  "FUNC_002": "bar"
+}

tests/multiple-solution-files/expected_representation.json

@@ -1 +1,3 @@
-{"version": 1}
+{
+  "version": 1
+}

tests/multiple-solution-files/expected_representation.txt

@@ -1 +1,17 @@
+start
+  func_def
+    func_header
+      FUNC_001
+      func_args
+    return_stmt
+      expr	1
 
+
+start
+  func_def
+    func_header
+      FUNC_002
+      func_args
+    return_stmt
+      expr
+        string	"Hello"

tests/multiple-solution-files/one.gd

@@ -0,0 +1,2 @@
+func foo():
+	return 1

tests/multiple-solution-files/two.gd

@@ -0,0 +1,2 @@
+func bar():
+	return "Hello"

tests/normalize-whitespace/expected_mapping.json

@@ -1 +1,5 @@
-{}
+{
+  "FUNCVAR_001": "a",
+  "FUNCVAR_002": "b",
+  "FUNC_001": "foo"
+}

tests/normalize-whitespace/expected_representation.json

@@ -1 +1,3 @@
-{"version": 1}
+{
+  "version": 1
+}

tests/normalize-whitespace/expected_representation.txt

@@ -1 +1,21 @@
-
+start
+  func_def
+    func_header
+      FUNC_001
+      func_args
+    func_var_stmt
+      func_var_assigned
+        FUNCVAR_001
+        expr
+          arith_expr
+            1
+            +
+            2
+    func_var_stmt
+      func_var_assigned
+        FUNCVAR_002
+        expr
+          arith_expr
+            3
+            +
+            4

tests/normalize-whitespace/normalize_whitespace.gd

@@ -0,0 +1,5 @@
+func foo():
+	var a = 1 + 2
+
+
+	var b    =    3   + 4

tests/remove-comments/expected_mapping.json

@@ -1 +1,5 @@
-{}
+{
+  "FUNCVAR_001": "a",
+  "FUNCVAR_002": "b",
+  "FUNC_001": "foo"
+}

tests/remove-comments/expected_representation.json

@@ -1 +1,3 @@
-{"version": 1}
+{
+  "version": 1
+}

tests/remove-comments/expected_representation.txt

@@ -1 +1,20 @@
-
+start
+  func_def
+    func_header
+      FUNC_001
+      func_args
+    func_var_stmt
+      func_var_assigned
+        FUNCVAR_001
+        expr
+          arith_expr
+            1
+            +
+            2
+    func_var_stmt
+      func_var_assigned
+        FUNCVAR_002
+        expr
+          string	"Hello World"
+    return_stmt
+      expr	FUNCVAR_001

tests/remove-comments/remove_comments.gd

@@ -0,0 +1,10 @@
+"""Module docstring."""
+# Module comment
+
+func foo():
+	"""Function docstring."""
+	# Function comment
+	var a = 1 + 2
+	"Some other string"
+	var b = "Hello World"
+	return a

tests/rename-identifiers/expected_mapping.json

@@ -0,0 +1,7 @@
+{
+  "FUNCVAR_001": "name",
+  "FUNCVAR_002": "suffix",
+  "FUNCVAR_003": "greeting",
+  "FUNC_001": "greet",
+  "FUNC_002": "greet_bob"
+}

tests/rename-identifiers/expected_representation.json

@@ -0,0 +1,3 @@
+{
+  "version": 1
+}

tests/rename-identifiers/expected_representation.txt

@@ -0,0 +1,34 @@
+start
+  func_def
+    func_header
+      FUNC_001
+      func_args
+        func_arg_typed
+          FUNCVAR_001
+          str
+        func_arg_regular	FUNCVAR_002
+    func_var_stmt
+      func_var_assigned
+        FUNCVAR_003
+        expr
+          string	"Hello "
+    expr_stmt
+      expr
+        standalone_call
+          print
+          arith_expr
+            FUNCVAR_003
+            +
+            FUNCVAR_001
+            +
+            FUNCVAR_002
+  func_def
+    func_header
+      FUNC_002
+      func_args
+    expr_stmt
+      expr
+        standalone_call
+          FUNC_001
+          string	"Bob"
+          string	"!"

tests/rename-identifiers/rename_identifiers.gd

@@ -0,0 +1,6 @@
+func greet(name: str, suffix):
+	var greeting = "Hello "
+	print(greeting + name + suffix)
+
+func greet_bob():
+	greet("Bob", "!")