Add a test generator, copied from bash, and templetize [anagram]

bin/generate_tests

@@ -0,0 +1,212 @@
+#!/usr/bin/env python3
+
+"""Test generator v1."""
+
+import argparse
+import csv
+import datetime
+import io
+import json
+import os
+import pathlib
+import re
+import shlex
+import subprocess
+import sys
+import textwrap
+import tomllib
+
+import jinja2
+
+
+def problem_spec_dir() -> pathlib.Path:
+    """Detect and return the problem specs."""
+    cache_dir = os.getenv("XDG_CACHE_HOME", os.getenv("HOME") + "/.cache")
+    specs = pathlib.Path(cache_dir) / "exercism/configlet/problem-specifications"
+    if specs.exists():
+        return specs
+    cur = pathlib.Path(os.getcwd())
+    for i in cur.parents:
+        if i.name == "problem-specifications":
+            return i
+    raise LookupError("Could not find problem specs")
+
+
+def flatten_cases(cases: list[dict]) -> list[tuple[list[str], dict]]:
+    """Recursive flatten test cases, returning individual cases with parent descriptions."""
+    for case_or_group in cases:
+        if "cases" in case_or_group:
+            for groups, child_case in flatten_cases(case_or_group["cases"]):
+                yield ([case_or_group["description"]] + groups, child_case)
+        else:
+            yield ([], case_or_group)
+
+
+def get_cases(specs: pathlib.Path, exercise: pathlib.Path) -> list[dict]:
+    """Return flattened, filtered cases with additional metadata attached."""
+    canonical_path = specs / "exercises" / exercise.name / "canonical-data.json"
+    with open(canonical_path, "r", encoding="utf-8") as f:
+        canonical = json.load(f)
+    with open(exercise / ".meta" / "tests.toml", "rb") as f:
+        tests = tomllib.load(f)
+
+    reimplemented = {
+        test["reimplements"]
+        for test in tests.values()
+        if test.get("include", True) and "reimplements" in test
+    }
+    cases = []
+    for groups, case in flatten_cases(canonical["cases"]):
+        # Filter out test cases with include=false or not listed.
+        if case["uuid"] not in tests or case["uuid"] in reimplemented:
+            continue
+        if not tests[case["uuid"]].get("include", True):
+            continue
+        # Add metadata.
+        case["descriptions"] = groups + [case["description"]]
+        case["expect_error"] = isinstance(case["expected"], dict) and "error" in case["expected"]
+        if case["expect_error"]:
+            case["expect_error_msg"] = case["expected"]["error"]
+        cases.append(case)
+    return cases
+
+
+def filter_tojson(data, separators=(',', ':'), indent=None) -> str:
+    """Filter `tojson` that JSON encodes a string with flexible settings."""
+    return json.dumps(data, separators=separators, indent=indent)
+
+
+def filter_tocsv(data: list) -> str:
+    """Filter `tocsv` that CSV encodes an object."""
+    formatted_data = []
+    for datum in data:
+        if isinstance(datum, (int, float, str)):
+            formatted_data.append(datum)
+        elif isinstance(datum, list):
+            formatted_data.append("[" + filter_tocsv(datum) + "]")
+        else:
+            raise ValueError(f"Unsuported data type {type(datum)} in filter_tocsv()")
+    buff = io.StringIO()
+    writer = csv.writer(buff, dialect="unix")
+    writer.writerow(formatted_data)
+    buff.seek(0)
+    got = buff.read().removesuffix("\n")
+    return got
+
+
+def jinja_env(exercise: pathlib.Path) -> jinja2.Environment:
+    """Return a configured Jinja env with filters added."""
+    env = jinja2.Environment(loader=jinja2.FileSystemLoader(exercise / ".meta"))
+    # Shell quoting
+    env.filters["quote"] = shlex.quote
+    # JSON formatting, default to compact form (`jq -c`).
+    env.filters["tojson"] = filter_tojson
+    env.filters["tocsv"] = filter_tocsv
+    # String escaping, ANSI-C style.
+    env.filters["repr"] = repr
+    # Return a dict with only specified keys kepts.
+    env.filters["camel_to_snake"] = lambda x: re.sub(r"([a-z])([A-Z])", (lambda m: f"{m.group(1)}_{m.group(2).lower()}"), x)
+    env.filters["format_list"] = lambda x: shlex.quote(
+        "[" + ",".join(f'"{i}"' if isinstance(i, str) else str(i) for i in x) + "]"
+    )
+    return env
+
+
+def bool_to_str(obj):
+    """Convert boolean values to strings."""
+    if isinstance(obj, dict):
+        return {key: bool_to_str(val) for key, val in obj.items()}
+    if isinstance(obj, list):
+        return [bool_to_str(val) for val in obj]
+    if obj is True:
+        return "true"
+    if obj is False:
+        return "false"
+    return obj
+
+
+def generate(specs: pathlib.Path, exercise: pathlib.Path) -> None:
+    """Generate and write test file for a given spec and exercise."""
+    cases = get_cases(specs, exercise)
+    for case in cases:
+        case["expected"] = bool_to_str(case["expected"])
+
+    timestamp = datetime.datetime.now(tz=datetime.UTC).replace(microsecond=0).isoformat()
+    header = textwrap.dedent(f"""\
+        #!/usr/bin/env bats
+        load bats-extra
+
+        # generated on {timestamp}
+        """
+    ).strip()
+    data = {
+        "cases": cases,
+        "header": header,
+        "solution": json.loads((exercise / ".meta/config.json").read_text())["files"]["solution"][0]
+    }
+
+    # Render the template.
+    for out_file in ["create_test_table.sql", "data.csv"]:
+        base = out_file.split(".")[0]
+        if not (exercise / ".meta" / f"template.{base}.j2").exists():
+            continue
+        try:
+            template = jinja_env(exercise).get_template(f"template.{base}.j2")
+            out = template.render(data).strip("\n")
+        except jinja2.exceptions.TemplateAssertionError as e:
+            e.add_note(f"Error rendering template.{base}.j2 for {exercise.name}")
+            raise
+
+        # Write the test file.
+        (exercise / out_file).write_text(out + "\n")
+
+
+def argparser() -> argparse.ArgumentParser:
+    parser = argparse.ArgumentParser()
+    parser.add_argument(
+        "--no-pull",
+        action="store_false",
+        dest="pull",
+        help="Do not run `git pull` on the problem specs repo",
+    )
+    parser.add_argument(
+        "exercises",
+        nargs="*",
+        help="exercises to generate tests; if none supplied, generate all"
+    )
+    return parser
+
+
+def main():
+    """Main entrypoint."""
+    specs = problem_spec_dir()
+    args = argparser().parse_args()
+    if args.pull:
+        subprocess.check_call(["git", "pull"], cwd=specs)
+    exercises = args.exercises
+    # Generate all exercises with templates if none are specified as args.
+    if not exercises:
+        exercises = [
+            i.parent.parent
+            for i in pathlib.Path("exercises/practice").glob("*/.meta/template.j2")
+        ]
+    else:
+        # Turn strings to paths and make them relative to the practice exercises.
+        out = []
+        practice = pathlib.Path("exercises/practice")
+        for exercise in exercises:
+            path = pathlib.Path(exercise)
+            if not path.is_relative_to(practice):
+                path = practice / path
+            out.append(path)
+        exercises = out
+
+    for exercise in exercises:
+        exercise_path = pathlib.Path(exercise)
+        if not exercise_path.exists():
+            raise ValueError(f"Exercise {exercise_path} does not exist")
+        generate(specs, exercise_path)
+
+
+if __name__ == "__main__":
+    main()

exercises/practice/anagram/.meta/template.create_test_table.j2

@@ -0,0 +1,34 @@
+DROP TABLE IF EXISTS tests;
+
+CREATE TABLE IF NOT EXISTS tests (
+  -- uuid and description are taken from the test.toml file
+  uuid TEXT PRIMARY KEY,
+  description TEXT NOT NULL,
+  -- The following section is needed by the online test-runner
+  status TEXT DEFAULT 'fail',
+  message TEXT,
+  output TEXT,
+  test_code TEXT,
+  task_id INTEGER DEFAULT NULL,
+  -- Here are columns for the actual tests
+  subject TEXT NOT NULL,
+  candidates TEXT NOT NULL, -- json array of strings
+  expected TEXT NOT NULL
+);
+
+INSERT INTO
+  tests (uuid, description, subject, candidates, expected)
+VALUES
+{%- for case in cases %}
+  (
+    '{{ case["uuid"] }}',
+    '{{ case["description"] }}',
+    '{{ case["input"]["subject"] }}',
+    '[{{ case["input"]["candidates"] | tocsv }}]',
+    '[{{ case["expected"] | tocsv }}]'
+{%- if loop.last %}
+  );
+{%- else %}
+  ),
+{%- endif %}
+{%- endfor %}

exercises/practice/anagram/.meta/template.data.j2

@@ -0,0 +1,3 @@
+{%- for case in cases -%}
+{{ [case["input"]["subject"], case["input"]["candidates"]] | tocsv }},""
+{% endfor %}