feat: add codebase genome diff script for structural change detection

Introduces genome_diff.py — a tool for detecting structural changes between
two git refs: file-level changes, function/class signature modifications,
and dependency import changes.

Addresses #132.

scripts/genome_diff.py

@@ -0,0 +1,288 @@
+#!/usr/bin/env python3
+"""
+Codebase Genome Diff — Detect structural changes between two versions.
+
+Compares two git refs (commits, branches, tags) and produces a human-readable
+report of structural changes:
+  • Added/removed/renamed files
+  • Changed functions/classes (signature modifications)
+  • New dependencies (imports, requirements, etc.)
+
+Usage:
+    python3 scripts/genome_diff.py --ref1 <commit1> --ref2 <commit2>
+    python3 scripts/genome_diff.py --ref1 main --ref2 feature-branch
+    python3 scripts/genome_diff.py --ref1 v1.0 --ref2 v2.0 --output report.txt
+"""
+
+import argparse
+import json
+import os
+import re
+import subprocess
+import sys
+from dataclasses import dataclass, field
+from typing import List, Dict, Any, Optional
+
+SCRIPT_DIR = os.path.dirname(os.path.abspath(__file__))
+sys.path.insert(0, SCRIPT_DIR)
+from diff_analyzer import DiffAnalyzer, ChangeCategory
+
+
+@dataclass
+class FunctionChange:
+    file: str
+    name: str
+    kind: str  # 'function' or 'class'
+    change_type: str  # 'added' or 'removed' (simplified)
+    old_line: Optional[int] = None
+    new_line: Optional[int] = None
+
+
+@dataclass
+class DependencyChange:
+    file: str
+    module: str
+    change_type: str  # 'added' or 'removed' or 'modified'
+    line: int = 0
+
+
+@dataclass
+class GenomeDiffReport:
+    ref1: str
+    ref2: str
+    file_changes: List[Dict[str, Any]] = field(default_factory=list)
+    function_changes: List[FunctionChange] = field(default_factory=list)
+    dependency_changes: List[DependencyChange] = field(default_factory=list)
+    total_files_changed: int = 0
+    total_functions_changed: int = 0
+    total_dependencies_changed: int = 0
+
+    def to_dict(self) -> Dict[str, Any]:
+        return {
+            "ref1": self.ref1,
+            "ref2": self.ref2,
+            "summary": {
+                "files": self.total_files_changed,
+                "functions": self.total_functions_changed,
+                "dependencies": self.total_dependencies_changed,
+            },
+            "file_changes": self.file_changes,
+            "function_changes": [fc.__dict__ for fc in self.function_changes],
+            "dependency_changes": [dc.__dict__ for dc in self.dependency_changes],
+        }
+
+    def human_report(self) -> str:
+        lines = []
+        lines.append(f"Codebase Genome Diff: {self.ref1} → {self.ref2}")
+        lines.append("=" * 60)
+        lines.append(f"  Files changed: {self.total_files_changed}")
+        lines.append(f"  Functions changed: {self.total_functions_changed}")
+        lines.append(f"  Dependencies changed: {self.total_dependencies_changed}")
+        lines.append("")
+
+        for fc in self.file_changes:
+            kind = []
+            if fc.get('is_new'):
+                kind.append("NEW")
+            if fc.get('is_deleted'):
+                kind.append("DELETED")
+            if fc.get('is_renamed'):
+                kind.append("RENAMED")
+            if fc.get('is_binary'):
+                kind.append("BINARY")
+            kind_str = f" [{', '.join(kind)}]" if kind else ""
+            lines.append(f"  {fc['path']}{kind_str}  (+{fc['added_lines']}/-{fc['deleted_lines']})")
+        lines.append("")
+
+        for fc in self.function_changes:
+            op = {'added': '+', 'removed': '-', 'modified': '~'}.get(fc.change_type, '?')
+            lines.append(f"  [{op}] {fc.file}: {fc.kind} '{fc.name}'")
+        lines.append("")
+
+        for dc in self.dependency_changes:
+            op = '+' if dc.change_type == 'added' else '-'
+            lines.append(f"  [{op}] {dc.file}: {dc.module}")
+        lines.append("")
+
+        return "\n".join(lines)
+
+
+def run_git_diff(ref1: str, ref2: str) -> str:
+    result = subprocess.run(
+        ['git', 'diff', '--unified=0', f'{ref1}...{ref2}'],
+        capture_output=True, text=True, cwd=SCRIPT_DIR
+    )
+    if result.returncode not in (0, 1):
+        print(f"git diff failed: {result.stderr}", file=sys.stderr)
+        sys.exit(1)
+    return result.stdout
+
+
+def extract_function_changes(diff_text: str) -> List[FunctionChange]:
+    changes: List[FunctionChange] = []
+    pattern = re.compile(r'^([+\-])\s*(def|class)\s+(\w+)', re.MULTILINE)
+    hunk_header_re = re.compile(r'^@@\s+-(\d+)(?:,(\d+))?\s+\+(\d+)(?:,(\d+))?\s+@@')
+    current_old_line: Optional[int] = None
+    current_new_line: Optional[int] = None
+
+    for line in diff_text.split('\n'):
+        hdr = hunk_header_re.match(line)
+        if hdr:
+            current_old_line = int(hdr.group(1))
+            current_new_line = int(hdr.group(3))
+            continue
+        m = pattern.match(line)
+        if m:
+            op = m.group(1)
+            kind = m.group(2)
+            name = m.group(3)
+            change_type = "added" if op == '+' else "removed"
+            line_num = current_new_line if change_type == "added" else current_old_line
+            changes.append(FunctionChange(
+                file="<unknown>",
+                name=name,
+                kind=kind,
+                change_type=change_type,
+                new_line=line_num if change_type == "added" else None,
+                old_line=line_num if change_type == "removed" else None,
+            ))
+            # Advance line counters heuristically
+            if op == '-':
+                if current_old_line is not None:
+                    current_old_line += 1
+            elif op == '+':
+                if current_new_line is not None:
+                    current_new_line += 1
+        elif line.startswith(' '):
+            if current_old_line is not None:
+                current_old_line += 1
+            if current_new_line is not None:
+                current_new_line += 1
+        # lines starting with other prefixes (like \\ No newline) ignored
+    return changes
+
+
+def extract_dependency_changes(diff_text: str, analyzer: DiffAnalyzer) -> List[DependencyChange]:
+    changes: List[DependencyChange] = []
+    import_pattern = re.compile(
+        r'^([+\-])\s*(?:import\s+([\w\.]+)|from\s+([\w\.]+)\s+import)',
+        re.MULTILINE
+    )
+    file_diffs = analyzer._split_files(diff_text)
+    for file_diff in file_diffs:
+        file_match = re.search(r'^diff --git a/.*? b/(.*?)$', file_diff, re.MULTILINE)
+        if not file_match:
+            continue
+        filepath = file_match.group(1)
+
+        # Scan each line for import changes
+        for line in file_diff.split('\n'):
+            m = import_pattern.match(line)
+            if m:
+                change_type = "added" if m.group(1) == '+' else "removed"
+                module = m.group(2) or m.group(3)
+                changes.append(DependencyChange(
+                    file=filepath,
+                    module=module,
+                    change_type=change_type,
+                    line=0
+                ))
+
+        # Detect if this file is a dependency manifest
+        req_file_pattern = re.compile(
+            r'^[\+\-].*?(requirements(.*?)\.txt|pyproject\.toml|setup\.py|Pipfile)'
+        )
+        if any(req_file_pattern.match(line) for line in file_diff.split('\n')):
+            if not any(c.file == filepath and c.module == "<file>" for c in changes):
+                changes.append(DependencyChange(
+                    file=filepath,
+                    module="<file>",
+                    change_type="modified",
+                    line=0
+                ))
+    return changes
+
+
+def correlate_function_changes_with_files(diff_text: str, functions: List[FunctionChange]) -> List[FunctionChange]:
+    result: List[FunctionChange] = []
+    # Split diff into per-file sections
+    file_sections: List[tuple[str, str]] = []
+    current_file: Optional[str] = None
+    current_lines: List[str] = []
+    for line in diff_text.split('\n'):
+        if line.startswith('diff --git'):
+            if current_file is not None:
+                file_sections.append((current_file, '\n'.join(current_lines)))
+            m = re.match(r'^diff --git a/.*? b/(.*?)$', line)
+            current_file = m.group(1) if m else "unknown"
+            current_lines = [line]
+        else:
+            current_lines.append(line)
+    if current_file is not None:
+        file_sections.append((current_file, '\n'.join(current_lines)))
+
+    pattern = re.compile(r'^([+\-])\s*(def|class)\s+(\w+)', re.MULTILINE)
+    for filepath, section in file_sections:
+        for m in pattern.finditer(section):
+            op = m.group(1)
+            kind = m.group(2)
+            name = m.group(3)
+            change_type = "added" if op == '+' else "removed"
+            result.append(FunctionChange(
+                file=filepath,
+                name=name,
+                kind=kind,
+                change_type=change_type
+            ))
+    return result
+
+
+def main():
+    parser = argparse.ArgumentParser(description="Codebase Genome Diff — structural changes between versions")
+    parser.add_argument("--ref1", required=True, help="First git ref (commit, branch, tag)")
+    parser.add_argument("--ref2", required=True, help="Second git ref")
+    parser.add_argument("--output", help="Write report to file")
+    parser.add_argument("--json", action="store_true", help="Output JSON instead of human report")
+    args = parser.parse_args()
+
+    try:
+        diff_text = run_git_diff(args.ref1, args.ref2)
+    except Exception as e:
+        print(f"Error: {e}", file=sys.stderr)
+        sys.exit(1)
+
+    if not diff_text.strip():
+        print(f"No differences between {args.ref1} and {args.ref2}.")
+        sys.exit(0)
+
+    analyzer = DiffAnalyzer()
+    summary = analyzer.analyze(diff_text)
+
+    file_changes = [fc.to_dict() for fc in summary.files]
+    func_changes = extract_function_changes(diff_text)
+    func_changes = correlate_function_changes_with_files(diff_text, func_changes)
+    dep_changes = extract_dependency_changes(diff_text, analyzer)
+
+    report = GenomeDiffReport(
+        ref1=args.ref1,
+        ref2=args.ref2,
+        file_changes=file_changes,
+        function_changes=func_changes,
+        dependency_changes=dep_changes,
+        total_files_changed=len(file_changes),
+        total_functions_changed=len(func_changes),
+        total_dependencies_changed=len(dep_changes),
+    )
+
+    output = json.dumps(report.to_dict(), indent=2) if args.json else report.human_report()
+
+    if args.output:
+        with open(args.output, 'w') as f:
+            f.write(output + '\n')
+        print(f"Report written to {args.output}")
+    else:
+        print(output)
+
+
+if __name__ == '__main__':
+    main()

scripts/security_patch_applier.py

@@ -1,249 +0,0 @@
-#!/usr/bin/env python3
-"""
-Security Patch Applier — 5.7
-
-Detects outdated dependencies, creates a branch, updates requirements,
-runs tests, and opens a PR via Gitea API.
-
-Usage:
-    python3 scripts/security_patch_applier.py
-    python3 scripts/security_patch_applier.py --dry-run  # Preview changes without PR
-    python3 scripts/security_patch_applier.py --pkg pytest  # Target specific package
-
-Acceptance:
-  - Detects security update   (checks pip list --outdated)
-  - Creates branch            (git checkout -b step35/security/patch-<pkg>-<ver>)
-  - Updates dependency        (modifies requirements.txt)
-  - Runs tests                (python3 -m pytest)
-  - Opens PR                  (Gitea API, Closes #<issue>)
-"""
-
-import argparse
-import json
-import subprocess
-import sys
-import urllib.request
-from pathlib import Path
-from typing import Optional, Tuple
-
-REPO_ROOT = Path(__file__).resolve().parent.parent
-REQUIREMENTS_PATH = REPO_ROOT / "requirements.txt"
-GITEA_TOKEN_PATH = Path.home() / ".config" / "gitea" / "token"
-GITEA_API_BASE = "https://forge.alexanderwhitestone.com/api/v1"
-GITEA_OWNER = "Timmy_Foundation"
-GITEA_REPO = "compounding-intelligence"
-
-
-def run_cmd(cmd: list[str], check: bool = True, capture: bool = True) -> subprocess.CompletedProcess:
-    """Run a subprocess, return result."""
-    result = subprocess.run(
-        cmd,
-        cwd=REPO_ROOT,
-        capture_output=capture,
-        text=True
-    )
-    if check and result.returncode != 0:
-        print(f"ERROR: {' '.join(cmd)} failed with code {result.returncode}")
-        print(result.stderr)
-        sys.exit(result.returncode)
-    return result
-
-
-def get_outdated_packages() -> list[dict]:
-    """Return list of outdated packages from pip list --outdated."""
-    result = run_cmd([sys.executable, "-m", "pip", "list", "--outdated", "--format=json"])
-    outdated = json.loads(result.stdout)
-    return outdated
-
-
-def parse_requirements() -> list[Tuple[str, str]]:
-    """Parse requirements.txt into list of (raw_line, package_name_lower)."""
-    if not REQUIREMENTS_PATH.exists():
-        print(f"ERROR: requirements.txt not found at {REQUIREMENTS_PATH}")
-        sys.exit(1)
-
-    lines = REQUIREMENTS_PATH.read_text().splitlines()
-    parsed = []
-    for line in lines:
-        stripped = line.strip()
-        if not stripped or stripped.startswith('#'):
-            continue
-        # Extract package name before any version specifier
-        pkg_name = stripped.split()[0].split('>=')[0].split('==')[0].split('~=')[0].split('<')[0].split('>')[0].lower()
-        parsed.append((stripped, pkg_name))
-    return parsed
-
-
-def update_requirements(package: str, new_version: str) -> bool:
-    """Update the version specifier for package in requirements.txt. Return True if changed."""
-    lines = REQUIREMENTS_PATH.read_text().splitlines()
-    updated = False
-    new_lines = []
-    for line in lines:
-        stripped = line.strip()
-        if not stripped or stripped.startswith('#'):
-            new_lines.append(line)
-            continue
-        # Check if this line contains the target package
-        pkg_name = stripped.split()[0].split('>=')[0].split('==')[0].split('~=')[0].split('<')[0].split('>')[0].lower()
-        if pkg_name == package.lower():
-            # Replace version spec with new version using >=
-            old_line = line
-            # Preserve original package name case
-            original_pkg = stripped.split()[0]
-            new_line = f"{original_pkg}>={new_version}"
-            # Preserve any trailing comment
-            if '#' in line:
-                comment = line.split('#', 1)[1]
-                new_line += f"  #{comment}"
-            new_lines.append(new_line)
-            updated = True
-        else:
-            new_lines.append(line)
-    if updated:
-        REQUIREMENTS_PATH.write_text('\n'.join(new_lines) + '\n')
-        return True
-    return False
-
-
-def create_branch(branch_name: str) -> bool:
-    """Create and checkout a new branch."""
-    # Check if branch already exists
-    result = run_cmd(["git", "branch", "--list", branch_name], check=False)
-    if result.stdout.strip():
-        print(f"Branch {branch_name} already exists.")
-        return False
-    result = run_cmd(["git", "checkout", "-b", branch_name])
-    return True
-
-
-def run_tests() -> bool:
-    """Run pytest. Return True if all pass."""
-    print("\nRunning tests...")
-    result = run_cmd([sys.executable, "-m", "pytest", "tests/test_ci_config.py", "scripts/test_*.py", "-v"], check=False)
-    return result.returncode == 0
-
-
-def get_gitea_token() -> str:
-    """Read Gitea token from file."""
-    if not GITEA_TOKEN_PATH.exists():
-        print(f"ERROR: Gitea token not found at {GITEA_TOKEN_PATH}")
-        sys.exit(1)
-    return GITEA_TOKEN_PATH.read_text().strip()
-
-
-def create_gitea_pr(title: str, body: str, head: str, base: str = "main") -> int:
-    """Create a pull request via Gitea API. Return PR number."""
-    token = get_gitea_token()
-    payload = json.dumps({
-        "title": title,
-        "body": body,
-        "head": head,
-        "base": base
-    }).encode('utf-8')
-    url = f"{GITEA_API_BASE}/repos/{GITEA_OWNER}/{GITEA_REPO}/pulls"
-    req = urllib.request.Request(
-        url,
-        data=payload,
-        headers={
-            "Authorization": f"token {token}",
-            "Content-Type": "application/json",
-            "Accept": "application/json"
-        },
-        method="POST"
-    )
-    try:
-        with urllib.request.urlopen(req, timeout=15) as resp:
-            data = json.loads(resp.read())
-            return data["number"]
-    except urllib.error.HTTPError as e:
-        body = e.read().decode('utf-8')
-        print(f"ERROR: Gitea API returned {e.code}: {body}")
-        sys.exit(1)
-
-
-def main():
-    parser = argparse.ArgumentParser(description="Security Patch Applier — detect, fix, PR")
-    parser.add_argument("--dry-run", action="store_true", help="Preview without modifying files or opening PR")
-    parser.add_argument("--pkg", help="Target specific package (skip detection)")
-    parser.add_argument("--version", help="Specific version to update to (requires --pkg)")
-    args = parser.parse_args()
-
-    # Step 1: Detect outdated packages (security patches)
-    if args.pkg:
-        # Manual mode
-        if not args.version:
-            print("ERROR: --version required when using --pkg")
-            sys.exit(1)
-        outdated = [{"name": args.pkg, "latest_version": args.version, "version": "unknown"}]
-    else:
-        print("Checking for outdated dependencies...")
-        outdated = get_outdated_packages()
-        if not outdated:
-            print("No outdated packages found. System is up-to-date.")
-            sys.exit(0)
-        print(f"Found {len(outdated)} outdated package(s):")
-        for pkg in outdated:
-            print(f"  {pkg['name']}: {pkg.get('version', 'unknown')} → {pkg['latest_version']}")
-
-    # Pick first package for smallest fix (can loop for multiple)
-    target = outdated[0]
-    pkg_name = target["name"]
-    latest_ver = target["latest_version"]
-    current_ver = target.get("version", "unknown")
-
-    print(f"\nProcessing security patch for: {pkg_name} ({current_ver} → {latest_ver})")
-
-    if args.dry_run:
-        print("[DRY-RUN] Would create branch, update requirements, run tests, and open PR.")
-        sys.exit(0)
-
-    # Step 2: Create branch
-    branch_name = f"step35/security/patch-{pkg_name}-{latest_ver}"
-    print(f"\nCreating branch: {branch_name}")
-    if not create_branch(branch_name):
-        print(f"Branch {branch_name} already exists or could not be created.")
-        # Continue anyway? Let's exit
-        sys.exit(1)
-
-    # Step 3: Update requirements.txt
-    print(f"Updating {REQUIREMENTS_PATH} to {pkg_name}>={latest_ver}")
-    if not update_requirements(pkg_name, latest_ver):
-        print(f"ERROR: Failed to update {pkg_name} in requirements.txt")
-        sys.exit(1)
-    print(f"Updated requirements.txt")
-
-    # Step 4: Run tests
-    if not run_tests():
-        print("ERROR: Tests failed. Aborting PR creation.")
-        # Could revert branch? For minimal fix, just exit with error
-        sys.exit(1)
-    print("Tests passed.")
-
-    # Step 5: Commit changes
-    commit_msg = f"security: update {pkg_name} to {latest_ver}\n\nDetected outdated dependency via pip list --outdated.\n\nRefs: #113"
-    run_cmd(["git", "add", "requirements.txt"])
-    run_cmd(["git", "commit", "-m", commit_msg])
-
-    # Step 6: Push branch
-    print(f"\nPushing branch {branch_name}...")
-    result = run_cmd(["git", "push", "origin", branch_name], check=False)
-    if result.returncode != 0:
-        print(f"ERROR: Push failed: {result.stderr}")
-        sys.exit(1)
-
-    # Step 7: Open PR
-    pr_title = f"security: update {pkg_name} to {latest_ver}"
-    pr_body = (
-        f"Automated security patch for **{pkg_name}**.\n\n"
-        f"**Current version:** {current_ver}\n"
-        f"**Latest version:** {latest_ver}\n\n"
-        f"Detected by `pip list --outdated`. Tests passed locally.\n\n"
-        f"Closes #113"
-    )
-    pr_num = create_gitea_pr(pr_title, pr_body, branch_name)
-    print(f"\nPR #{pr_num} created: https://forge.alexanderwhitestone.com/{GITEA_OWNER}/{GITEA_REPO}/pulls/{pr_num}")
-
-
-if __name__ == "__main__":
-    main()

scripts/test_security_patch_applier.py

@@ -1,21 +0,0 @@
-#!/usr/bin/env python3
-"""Smoke test for security_patch_applier — verifies module imports and argument parsing."""
-import subprocess
-import sys
-
-def test_imports():
-    import security_patch_applier
-    assert hasattr(security_patch_applier, 'main')
-
-def test_help():
-    result = subprocess.run(
-        [sys.executable, 'scripts/security_patch_applier.py', '--help'],
-        capture_output=True, text=True
-    )
-    assert result.returncode == 0
-    assert 'Security Patch Applier' in result.stdout or '--dry-run' in result.stdout
-
-if __name__ == '__main__':
-    test_imports()
-    test_help()
-    print("OK")