Merge pull request #40 from arnoox/diff-script

feat: add transpiler output baseline and diff tool scripts

Author: arnoox

.gitignore

@@ -29,6 +29,9 @@ docs/.ub_cache/
 *.profraw
 *.profdata
 
+# Transpiler output baselines (created by scripts/transpile_diff.sh)
+.herkos-baselines/
+
 # Temporary files
 *.tmp
 *.bak

scripts/transpile_diff.py

@@ -0,0 +1,323 @@
+#!/usr/bin/env python3
+"""Baseline and diff tool for herkos transpiler output.
+
+Creates named snapshots of the .rs files generated by herkos-tests and
+lets you diff any two baselines—or a baseline against the current build.
+"""
+
+import argparse
+import os
+import shutil
+import subprocess
+import sys
+import tempfile
+from datetime import datetime, timezone
+from pathlib import Path
+
+
+# ---------------------------------------------------------------------------
+# Paths
+# ---------------------------------------------------------------------------
+
+SCRIPT_DIR = Path(__file__).resolve().parent
+REPO_ROOT = SCRIPT_DIR.parent
+BASELINES_DIR = REPO_ROOT / ".herkos-baselines"
+
+
+def check_repo_root() -> None:
+    if not (REPO_ROOT / "Cargo.toml").exists():
+        sys.exit("ERROR: Cargo.toml not found. Please run this script from the repo root.")
+
+
+# ---------------------------------------------------------------------------
+# OUT_DIR discovery
+# ---------------------------------------------------------------------------
+
+def find_out_dir(profile: str) -> Path | None:
+    """Find the herkos-tests OUT_DIR that actually contains .rs files."""
+    build_dir = REPO_ROOT / "target" / profile / "build"
+    if not build_dir.is_dir():
+        return None
+    for candidate in build_dir.glob("herkos-tests-*/out"):
+        if candidate.is_dir() and any(candidate.glob("*.rs")):
+            return candidate
+    return None
+
+
+# ---------------------------------------------------------------------------
+# Git helpers
+# ---------------------------------------------------------------------------
+
+def git(*args: str) -> str:
+    result = subprocess.run(
+        ["git", "-C", str(REPO_ROOT), *args],
+        capture_output=True, text=True, check=True,
+    )
+    return result.stdout.strip()
+
+
+# ---------------------------------------------------------------------------
+# Snapshot
+# ---------------------------------------------------------------------------
+
+def cmd_snapshot(name: str | None, profile: str) -> None:
+    out_dir = find_out_dir(profile)
+    if out_dir is None:
+        sys.exit(
+            f"ERROR: No populated OUT_DIR found in target/{profile}/build/herkos-tests-*/out/\n"
+            "       Run 'cargo build -p herkos-tests' first."
+        )
+
+    if name is None:
+        branch = git("rev-parse", "--abbrev-ref", "HEAD").replace("/", "_")
+        commit = git("rev-parse", "--short", "HEAD")
+        ts = datetime.now(timezone.utc).strftime("%Y%m%dT%H%M%SZ")
+        name = f"{branch}__{commit}__{ts}"
+
+    dest = BASELINES_DIR / name / "files"
+    if (BASELINES_DIR / name).exists():
+        sys.exit(
+            f"ERROR: Baseline '{name}' already exists. Delete it first:\n"
+            f"       scripts/transpile_diff.py delete {name}"
+        )
+
+    dest.mkdir(parents=True)
+    _copy_rs_files(out_dir, dest)
+
+    rs_count = sum(1 for _ in dest.rglob("*.rs"))
+    branch = git("rev-parse", "--abbrev-ref", "HEAD")
+    commit_sha = git("rev-parse", "HEAD")
+    commit_short = git("rev-parse", "--short", "HEAD")
+
+    meta = BASELINES_DIR / name / "meta.env"
+    meta.write_text(
+        f"TIMESTAMP={datetime.now(timezone.utc).strftime('%Y-%m-%dT%H:%M:%SZ')}\n"
+        f"BRANCH={branch}\n"
+        f"COMMIT_SHA={commit_sha}\n"
+        f"COMMIT_SHORT={commit_short}\n"
+        f"HERKOS_OPTIMIZE={os.environ.get('HERKOS_OPTIMIZE', '0')}\n"
+        f"PROFILE={profile}\n"
+        f"OUT_DIR={out_dir}\n"
+        f"RS_FILE_COUNT={rs_count}\n"
+    )
+
+    print(f"Snapshot '{name}' saved ({rs_count} .rs files).")
+    print(f"  Branch:  {branch}")
+    print(f"  Commit:  {commit_short}")
+    print(f"  OUT_DIR: {out_dir}")
+
+
+def _copy_rs_files(src: Path, dest: Path) -> None:
+    """Copy all .rs files (except *_src.rs) from src to dest, preserving structure."""
+    for rs_file in src.rglob("*.rs"):
+        if rs_file.name.endswith("_src.rs"):
+            continue
+        rel = rs_file.relative_to(src)
+        target = dest / rel
+        target.parent.mkdir(parents=True, exist_ok=True)
+        shutil.copy2(rs_file, target)
+
+
+# ---------------------------------------------------------------------------
+# List
+# ---------------------------------------------------------------------------
+
+def cmd_list() -> None:
+    if not BASELINES_DIR.is_dir() or not any(BASELINES_DIR.iterdir()):
+        print(f"No baselines found in {BASELINES_DIR}")
+        return
+
+    header = f"{'NAME':<42}  {'TIMESTAMP':<22}  {'BRANCH':<20}  {'COMMIT':<10}  FILES"
+    print(header)
+    print("─" * 100)
+
+    for meta_path in sorted(BASELINES_DIR.glob("*/meta.env")):
+        name = meta_path.parent.name
+        meta = _read_meta(meta_path)
+        print(
+            f"{name:<42}  "
+            f"{meta.get('TIMESTAMP', '?'):<22}  "
+            f"{meta.get('BRANCH', '?'):<20}  "
+            f"{meta.get('COMMIT_SHORT', '?'):<10}  "
+            f"{meta.get('RS_FILE_COUNT', '?')}"
+        )
+
+
+def _read_meta(path: Path) -> dict[str, str]:
+    meta: dict[str, str] = {}
+    for line in path.read_text().splitlines():
+        if "=" in line:
+            k, _, v = line.partition("=")
+            meta[k.strip()] = v.strip()
+    return meta
+
+
+# ---------------------------------------------------------------------------
+# Compare
+# ---------------------------------------------------------------------------
+
+def cmd_compare(a: str, b: str, profile: str) -> None:
+    dir_a = _resolve_baseline(a)
+    tmp_dir: str | None = None
+
+    if b == "latest":
+        out_dir = find_out_dir(profile)
+        if out_dir is None:
+            sys.exit(
+                f"ERROR: 'latest' requested but no populated OUT_DIR found in target/{profile}/.\n"
+                "       Run 'cargo build -p herkos-tests' or specify a baseline name."
+            )
+        tmp_dir = tempfile.mkdtemp()
+        _copy_rs_files(out_dir, Path(tmp_dir))
+        dir_b = Path(tmp_dir)
+        label_b = f"latest ({profile} build)"
+    else:
+        dir_b = _resolve_baseline(b)
+        label_b = b
+
+    try:
+        _run_compare(a, dir_a, label_b, dir_b)
+    finally:
+        if tmp_dir:
+            shutil.rmtree(tmp_dir, ignore_errors=True)
+
+
+def _resolve_baseline(name: str) -> Path:
+    path = BASELINES_DIR / name / "files"
+    if not path.is_dir():
+        sys.exit(
+            f"ERROR: Baseline '{name}' not found in {BASELINES_DIR}/\n"
+            "       Run 'scripts/transpile_diff.py list' to see available baselines."
+        )
+    return path
+
+
+def _run_compare(label_a: str, dir_a: Path, label_b: str, dir_b: Path) -> None:
+    print("=== herkos transpiler diff ===")
+    print(f"  A: {label_a}")
+    print(f"  B: {label_b}")
+    print()
+
+    files_a = {p.relative_to(dir_a) for p in dir_a.rglob("*.rs")}
+    files_b = {p.relative_to(dir_b) for p in dir_b.rglob("*.rs")}
+
+    added   = sorted(files_b - files_a)
+    removed = sorted(files_a - files_b)
+    changed = sorted(
+        f for f in files_a & files_b
+        if (dir_a / f).read_bytes() != (dir_b / f).read_bytes()
+    )
+
+    print(f"Summary:  Changed: {len(changed)}  |  Added: {len(added)}  |  Removed: {len(removed)}")
+    print()
+
+    if changed:
+        print("Changed files:")
+        for f in changed:
+            print(f"  ~ {f}")
+        print()
+    if added:
+        print("Added files (in B, not in A):")
+        for f in added:
+            print(f"  + {f}")
+        print()
+    if removed:
+        print("Removed files (in A, not in B):")
+        for f in removed:
+            print(f"  - {f}")
+        print()
+
+    if not (changed or added or removed):
+        print("No differences found.")
+        return
+
+    print("Full diff:")
+    print("─" * 60)
+
+    diff_proc = subprocess.Popen(
+        ["git", "-C", str(REPO_ROOT), "diff", "--no-index", "--color=always",
+         str(dir_a), str(dir_b)],
+        stdout=subprocess.PIPE,
+    )
+
+    if sys.stdout.isatty():
+        pager = subprocess.Popen(["less", "-R"], stdin=diff_proc.stdout)
+        diff_proc.stdout.close()  # type: ignore[union-attr]
+        pager.wait()
+    else:
+        shutil.copyfileobj(diff_proc.stdout, sys.stdout.buffer)  # type: ignore[arg-type]
+
+    diff_proc.wait()  # exit code 1 is normal when differences exist
+
+
+# ---------------------------------------------------------------------------
+# Delete
+# ---------------------------------------------------------------------------
+
+def cmd_delete(name: str) -> None:
+    path = BASELINES_DIR / name
+    if not path.is_dir():
+        sys.exit(f"ERROR: Baseline '{name}' not found in {BASELINES_DIR}/")
+    shutil.rmtree(path)
+    print(f"Deleted baseline '{name}'.")
+
+
+# ---------------------------------------------------------------------------
+# CLI
+# ---------------------------------------------------------------------------
+
+def main() -> None:
+    check_repo_root()
+
+    parser = argparse.ArgumentParser(
+        prog="scripts/transpile_diff.py",
+        description="Baseline and diff tool for herkos transpiler output.",
+        formatter_class=argparse.RawDescriptionHelpFormatter,
+        epilog="""
+examples:
+  scripts/transpile_diff.py snapshot before-optimizer
+  scripts/transpile_diff.py snapshot
+  scripts/transpile_diff.py list
+  scripts/transpile_diff.py compare before-optimizer
+  scripts/transpile_diff.py compare before-optimizer after-optimizer
+  scripts/transpile_diff.py --release compare main__abc
+  scripts/transpile_diff.py delete before-optimizer
+""",
+    )
+    parser.add_argument(
+        "--release", action="store_true",
+        help="Use target/release/ instead of target/debug/ when discovering OUT_DIR.",
+    )
+
+    sub = parser.add_subparsers(dest="subcommand", metavar="subcommand")
+    sub.required = True
+
+    p_snap = sub.add_parser("snapshot", help="Save current transpiler output as a named baseline.")
+    p_snap.add_argument("name", nargs="?", default=None,
+                        help="Baseline name (auto-generated if omitted).")
+
+    sub.add_parser("list", help="Show all saved baselines with metadata.")
+
+    p_cmp = sub.add_parser("compare", help="Diff baseline A vs B (default B: latest build).")
+    p_cmp.add_argument("a", metavar="A", help="First baseline name.")
+    p_cmp.add_argument("b", metavar="B", nargs="?", default="latest",
+                       help="Second baseline name, or 'latest' (default).")
+
+    p_del = sub.add_parser("delete", help="Remove a baseline.")
+    p_del.add_argument("name", help="Baseline name to delete.")
+
+    args = parser.parse_args()
+    profile = "release" if args.release else "debug"
+
+    if args.subcommand == "snapshot":
+        cmd_snapshot(args.name, profile)
+    elif args.subcommand == "list":
+        cmd_list()
+    elif args.subcommand == "compare":
+        cmd_compare(args.a, args.b, profile)
+    elif args.subcommand == "delete":
+        cmd_delete(args.name)
+
+
+if __name__ == "__main__":
+    main()

scripts/transpile_diff.sh

@@ -0,0 +1,2 @@
+#!/usr/bin/env bash
+exec python3 "$(dirname "$0")/transpile_diff.py" "$@"