its-magic 0.1.2-9 → 0.1.3-0

package/template/scripts/doc_profile_lib.py

@@ -0,0 +1,415 @@
+"""
+Documentation profile resolution and surface helpers (DEC-0059).
+
+Shared by installer (optional surface sync) and scripts/validate_doc_profile.py.
+"""
+
+from __future__ import annotations
+
+import os
+from typing import Dict, List, Optional, Set, Tuple
+
+DOC_AUDIENCE_ALLOWED = frozenset({"user", "developer", "both"})
+DOC_DETAIL_ALLOWED = frozenset({"concise", "balanced", "technical-deep"})
+
+USER_KEY_TO_H2: Dict[str, str] = {
+    "USER_PURPOSE": "Purpose",
+    "USER_QUICKSTART": "Quickstart",
+    "USER_EXAMPLES": "Examples",
+    "USER_TROUBLESHOOTING": "Troubleshooting",
+    "USER_LIMITATIONS": "Limitations",
+    "USER_RELATED_DOCS": "Related documentation",
+}
+
+DEV_KEY_TO_H2: Dict[str, str] = {
+    "DEV_PREREQS": "Prerequisites",
+    "DEV_WORKFLOW": "Workflow",
+    "DEV_QUALITY_GATES": "Quality gates",
+    "DEV_ARCHITECTURE": "Architecture notes",
+    "DEV_CONTRACTS": "Contracts and interfaces",
+    "DEV_DECISIONS": "Engineering decisions",
+}
+
+DEV_H2_TITLES = frozenset(DEV_KEY_TO_H2.values())
+USER_H2_TITLES = frozenset(USER_KEY_TO_H2.values())
+
+POINTER_H2 = "Contributing"
+
+# Root README H2 budget (user-channel headings + optional Contributing pointer only).
+ROOT_BUDGET: Dict[Tuple[str, str], int] = {}
+for aud in ("user", "developer", "both"):
+    for det in ("concise", "balanced", "technical-deep"):
+        if aud == "user":
+            ROOT_BUDGET[(aud, det)] = {"concise": 5, "balanced": 7, "technical-deep": 9}[det]
+        elif aud == "developer":
+            ROOT_BUDGET[(aud, det)] = {"concise": 4, "balanced": 6, "technical-deep": 8}[det]
+        else:
+            ROOT_BUDGET[(aud, det)] = {"concise": 6, "balanced": 8, "technical-deep": 6}[det]
+
+
+def resolve_doc_profile(merged: Dict[str, str]) -> Tuple[Optional[str], Optional[str], List[str]]:
+    """Return (audience, detail, errors). Empty profile keys default per DEC-0059 §6."""
+    raw_a = (merged.get("DOC_AUDIENCE_PROFILE") or "").strip().lower()
+    raw_d = (merged.get("DOC_DETAIL_LEVEL") or "").strip().lower()
+    errors: List[str] = []
+    if raw_a and raw_a not in DOC_AUDIENCE_ALLOWED:
+        errors.append(
+            "[DOC_PROFILE_INVALID] DOC_AUDIENCE_PROFILE must be one of: "
+            f"user, developer, both (got={raw_a!r})."
+        )
+    if raw_d and raw_d not in DOC_DETAIL_ALLOWED:
+        errors.append(
+            "[DOC_PROFILE_INVALID] DOC_DETAIL_LEVEL must be one of: "
+            f"concise, balanced, technical-deep (got={raw_d!r})."
+        )
+    if errors:
+        return None, None, errors
+    audience = raw_a or "both"
+    detail = raw_d or "balanced"
+    return audience, detail, []
+
+
+def required_user_keys(audience: str, detail: str) -> Set[str]:
+    if audience not in ("user", "both"):
+        return set()
+    if detail == "concise":
+        return {"USER_PURPOSE", "USER_QUICKSTART", "USER_LIMITATIONS"}
+    if detail == "balanced":
+        return {
+            "USER_PURPOSE",
+            "USER_QUICKSTART",
+            "USER_LIMITATIONS",
+            "USER_EXAMPLES",
+            "USER_RELATED_DOCS",
+        }
+    return {
+        "USER_PURPOSE",
+        "USER_QUICKSTART",
+        "USER_LIMITATIONS",
+        "USER_EXAMPLES",
+        "USER_RELATED_DOCS",
+        "USER_TROUBLESHOOTING",
+    }
+
+
+def required_dev_keys(audience: str, detail: str) -> Set[str]:
+    if audience not in ("developer", "both"):
+        return set()
+    if detail == "concise":
+        return {"DEV_PREREQS", "DEV_WORKFLOW"}
+    if detail == "balanced":
+        return {"DEV_PREREQS", "DEV_WORKFLOW", "DEV_QUALITY_GATES", "DEV_ARCHITECTURE"}
+    return {
+        "DEV_PREREQS",
+        "DEV_WORKFLOW",
+        "DEV_QUALITY_GATES",
+        "DEV_ARCHITECTURE",
+        "DEV_CONTRACTS",
+        "DEV_DECISIONS",
+    }
+
+
+def extract_h2_titles(markdown: str) -> List[str]:
+    titles: List[str] = []
+    for line in markdown.splitlines():
+        if line.startswith("## ") and not line.startswith("###"):
+            titles.append(line[3:].strip())
+    return titles
+
+
+def has_exact_h2(markdown: str, title: str) -> bool:
+    for line in markdown.splitlines():
+        if line.startswith("## ") and not line.startswith("###"):
+            if line[3:].strip() == title:
+                return True
+    return False
+
+
+def count_profile_root_h2s(
+    markdown: str,
+    audience: str,
+    detail: str,
+    required_user_keys_set: Set[str],
+) -> int:
+    """
+    Count H2 lines in budget scope: required USER_* titles only for user/both;
+    Contributing pointer alone for developer-only (DEC-0059 / R-0054 user H2 budgets).
+    """
+    titles = extract_h2_titles(markdown)
+    want: Set[str] = set()
+    for key in required_user_keys_set:
+        want.add(USER_KEY_TO_H2[key])
+    if audience == "developer":
+        return sum(1 for t in titles if t == POINTER_H2)
+    n = 0
+    for t in titles:
+        if t in want:
+            n += 1
+    return n
+
+
+def dev_h2_forbidden_in_root(markdown: str) -> List[str]:
+    """Return DEV_* H2 titles present in root (split layout forbids these in README)."""
+    found: List[str] = []
+    for t in extract_h2_titles(markdown):
+        if t in DEV_H2_TITLES:
+            found.append(t)
+    return found
+
+
+def validate_optional_modes(merged: Dict[str, str], readme_text: str) -> List[str]:
+    """
+    US-0031 / US-0032 compatibility: never require optional artifacts when modes are off.
+    When on, ensure profile surfaces mention cross-links (additive, lightweight).
+    """
+    out: List[str] = []
+    sp = (merged.get("SPEC_PACK_MODE") or "0").strip()
+    ug = (merged.get("USER_GUIDE_MODE") or "0").strip()
+    if sp != "1" and ug != "1":
+        return out
+    if sp == "1":
+        if "docs/engineering" not in readme_text and "spec" not in readme_text.lower():
+            out.append(
+                "[DOC_OPTIONAL_CROSSLINK_WEAK] SPEC_PACK_MODE=1: root README should mention "
+                "engineering docs or spec-pack paths in a user channel section (see Related documentation)."
+            )
+    if ug == "1":
+        if "user-guides" not in readme_text and "user guide" not in readme_text.lower():
+            out.append(
+                "[DOC_OPTIONAL_CROSSLINK_WEAK] USER_GUIDE_MODE=1: root README should mention "
+                "docs/user-guides in a user channel section."
+            )
+    return out
+
+
+def ensure_section(path: str, h2_title: str, body: str) -> Tuple[bool, str]:
+    """
+    Append ## h2_title + body if missing. Non-destructive.
+    Returns (changed, message).
+    """
+    ensure_parent_dir(path)
+    if os.path.isfile(path):
+        text = _read_utf8(path)
+    else:
+        text = ""
+
+    if has_exact_h2(text, h2_title):
+        return False, f"[DOC_PROFILE_SYNC] skip existing: ## {h2_title} ({path})"
+
+    block = f"\n\n## {h2_title}\n\n{body.strip()}\n"
+    if not text:
+        base = os.path.basename(path)
+        if base.lower() == "readme.md":
+            text = f"# Documentation\n"
+        else:
+            text = f"# {base.replace('.md', '').replace('_', ' ')}\n"
+    new_text = text.rstrip() + block
+    _write_utf8(path, new_text)
+    return True, f"[DOC_PROFILE_SYNC] appended: ## {h2_title} ({path})"
+
+
+def ensure_doc_surfaces_merged(
+    merged: Dict[str, str],
+    target_root: str,
+    print_ok: bool = True,
+) -> List[str]:
+    audience, detail, errors = resolve_doc_profile(merged)
+    if errors:
+        return list(errors)
+    assert audience is not None and detail is not None
+    messages: List[str] = []
+    uk = required_user_keys(audience, detail)
+    dk = required_dev_keys(audience, detail)
+    readme = os.path.join(target_root, "README.md")
+    dev_readme = os.path.join(target_root, "docs", "developer", "README.md")
+
+    stubs_user = {
+        "USER_PURPOSE": (
+            "Describe what this repository is for in plain language. "
+            "Replace this placeholder with your product outcome."
+        ),
+        "USER_QUICKSTART": (
+            "Link to your fastest path to success. For its-magic, see [Setup](#setup) above."
+        ),
+        "USER_EXAMPLES": "Add short, copy-paste friendly examples for common tasks.",
+        "USER_TROUBLESHOOTING": (
+            "List frequent issues, what to check, and where logs or docs live."
+        ),
+        "USER_LIMITATIONS": "Call out known limits, unsupported environments, or sharp edges.",
+        "USER_RELATED_DOCS": (
+            "Link runbooks, architecture notes, and deeper guides. "
+            "Operator commands live in `docs/engineering/runbook.md`."
+        ),
+    }
+    stubs_dev = {
+        "DEV_PREREQS": "Toolchain, repo layout, and local prerequisites for contributors.",
+        "DEV_WORKFLOW": "Branching, phases, and day-to-day contributor workflow.",
+        "DEV_QUALITY_GATES": "Tests, lint, typecheck, and review expectations before merge.",
+        "DEV_ARCHITECTURE": "High-level modules, boundaries, and extension points.",
+        "DEV_CONTRACTS": "Public interfaces, file formats, and compatibility promises.",
+        "DEV_DECISIONS": "Pointers to `decisions/` and architecture sections that matter.",
+    }
+
+    for key in sorted(uk):
+        h2 = USER_KEY_TO_H2[key]
+        changed, msg = ensure_section(readme, h2, stubs_user[key])
+        if print_ok or changed:
+            messages.append(msg)
+
+    if dk:
+        for key in sorted(dk):
+            h2 = DEV_KEY_TO_H2[key]
+            changed, msg = ensure_section(dev_readme, h2, stubs_dev[key])
+            if print_ok or changed:
+                messages.append(msg)
+
+    if audience in ("developer", "both"):
+        contrib_body = (
+            "Contributor-focused workflow and guardrails live in "
+            "[`docs/developer/README.md`](docs/developer/README.md)."
+        )
+        changed, msg = ensure_section(readme, POINTER_H2, contrib_body)
+        if print_ok or changed:
+            messages.append(msg)
+
+    return messages
+
+
+def optional_mode_warnings(merged: Dict[str, str], readme_text: str) -> List[str]:
+    """Non-blocking hints when optional doc modes are enabled (US-0031 / US-0032)."""
+    return validate_optional_modes(merged, readme_text)
+
+
+def ensure_parent_dir(path: str) -> None:
+    parent = os.path.dirname(path)
+    if parent and not os.path.isdir(parent):
+        os.makedirs(parent, exist_ok=True)
+
+
+def _read_utf8(path: str) -> str:
+    with open(path, "r", encoding="utf-8") as f:
+        return f.read()
+
+
+def _write_utf8(path: str, text: str) -> None:
+    ensure_parent_dir(path)
+    with open(path, "w", encoding="utf-8", newline="\n") as f:
+        f.write(text)
+
+
+def validate_repo_doc_profile(
+    target_root: str,
+    merged: Dict[str, str],
+    template_root: Optional[str],
+) -> List[str]:
+    """
+    Full validation for active target_root; optional template_root for parity.
+    Returns error strings (empty => pass).
+    """
+    errors: List[str] = []
+    audience, detail, err = resolve_doc_profile(merged)
+    errors.extend(err)
+    if errors:
+        return errors
+    assert audience is not None and detail is not None
+
+    uk = required_user_keys(audience, detail)
+    dk = required_dev_keys(audience, detail)
+
+    readme_path = os.path.join(target_root, "README.md")
+    dev_path = os.path.join(target_root, "docs", "developer", "README.md")
+
+    if not os.path.isfile(readme_path):
+        errors.append("[DOC_PROFILE_MERGE_ERROR] README.md missing (required for user channel checks).")
+        return errors
+
+    readme_text = _read_utf8(readme_path)
+
+    if audience in ("developer", "both") and dk:
+        if not os.path.isfile(dev_path):
+            errors.append(
+                "[DOC_SECTION_MISSING:shard] docs/developer/README.md missing but profile requires developer sections."
+            )
+
+    if audience in ("developer", "both"):
+        if not has_exact_h2(readme_text, POINTER_H2):
+            errors.append(
+                f"[DOC_SECTION_MISSING:DEV_SHARD_POINTER] Missing H2 ## {POINTER_H2} in README.md "
+                "(required pointer to docs/developer/README.md per DEC-0059)."
+            )
+        bad = dev_h2_forbidden_in_root(readme_text)
+        if bad:
+            errors.append(
+                "[DOC_PROFILE_INVALID] DEV sections must not use root README for split layout; "
+                f"found H2 titles {bad!r}. Move them to docs/developer/README.md."
+            )
+
+    for key in sorted(uk):
+        h2 = USER_KEY_TO_H2[key]
+        if not has_exact_h2(readme_text, h2):
+            errors.append(f"[DOC_SECTION_MISSING:{key}] Missing H2 ## {h2} in README.md.")
+
+    dev_text = ""
+    if os.path.isfile(dev_path):
+        dev_text = _read_utf8(dev_path)
+
+    for key in sorted(dk):
+        h2 = DEV_KEY_TO_H2[key]
+        if not has_exact_h2(dev_text, h2):
+            errors.append(f"[DOC_SECTION_MISSING:{key}] Missing H2 ## {h2} in docs/developer/README.md.")
+
+    budget = ROOT_BUDGET.get((audience, detail), 8)
+    counted = count_profile_root_h2s(readme_text, audience, detail, uk)
+    if counted > budget:
+        errors.append(
+            f"[DOC_SECTION_BUDGET_EXCEEDED] Root README profile-scoped H2 count={counted} "
+            f"exceeds budget={budget} for audience={audience!r} detail={detail!r}."
+        )
+
+    if template_root:
+        tr = os.path.join(template_root, "README.md")
+        td = os.path.join(template_root, "docs", "developer", "README.md")
+        if not os.path.isfile(tr):
+            errors.append("[DOC_TEMPLATE_PARITY_FAIL] template/README.md missing.")
+        else:
+            tt = _read_utf8(tr)
+            for key in sorted(uk):
+                h2 = USER_KEY_TO_H2[key]
+                if has_exact_h2(readme_text, h2) != has_exact_h2(tt, h2):
+                    errors.append(
+                        f"[DOC_TEMPLATE_PARITY_FAIL] USER H2 ## {h2} presence differs active vs template README."
+                    )
+            if audience in ("developer", "both"):
+                if has_exact_h2(readme_text, POINTER_H2) != has_exact_h2(tt, POINTER_H2):
+                    errors.append(
+                        "[DOC_TEMPLATE_PARITY_FAIL] ## Contributing pointer presence differs active vs template README."
+                    )
+        if dk:
+            if not os.path.isfile(td) or not os.path.isfile(dev_path):
+                errors.append(
+                    "[DOC_TEMPLATE_PARITY_FAIL] developer README missing in active or template."
+                )
+            else:
+                tdev = _read_utf8(td)
+                for key in sorted(dk):
+                    h2 = DEV_KEY_TO_H2[key]
+                    if has_exact_h2(dev_text, h2) != has_exact_h2(tdev, h2):
+                        errors.append(
+                            f"[DOC_TEMPLATE_PARITY_FAIL] DEV H2 ## {h2} presence differs active vs template."
+                        )
+
+    return errors
+
+
+def self_test_resolver() -> None:
+    """Tier B: assert matrix key sets."""
+    assert required_user_keys("user", "concise") == {
+        "USER_PURPOSE",
+        "USER_QUICKSTART",
+        "USER_LIMITATIONS",
+    }
+    assert required_dev_keys("developer", "technical-deep") == set(DEV_KEY_TO_H2.keys())
+    a, d, e = resolve_doc_profile({"DOC_AUDIENCE_PROFILE": "", "DOC_DETAIL_LEVEL": ""})
+    assert not e and a == "both" and d == "balanced"
+    _, _, e2 = resolve_doc_profile({"DOC_AUDIENCE_PROFILE": "nope"})
+    assert e2 and "DOC_PROFILE_INVALID" in e2[0]

package/template/scripts/downstream_ci_guard_lib.py

@@ -0,0 +1,222 @@
+"""
+Downstream CI drift guard — template forbidden-pattern scan + active inventory (BUG-0009 / DEC-0075).
+"""
+
+from __future__ import annotations
+
+import json
+import os
+import re
+from dataclasses import dataclass, field
+from typing import Any, Dict, List, Optional, Set, Tuple
+
+JOB_KEY_RE = re.compile(r"^\s{2}(\w[\w-]*):", re.MULTILINE)
+
+FORBIDDEN_JOB_IDS = frozenset({"npm-test", "brew-test", "choco-test"})
+ALLOWED_TEMPLATE_JOBS = frozenset({"checks", "auto-fix"})
+REQUIRED_ACTIVE_JOBS = frozenset(
+    {"checks", "auto-fix", "npm-test", "brew-test", "choco-test"}
+)
+
+FORBIDDEN_SUBSTRINGS: Tuple[str, ...] = (
+    "npm-test",
+    "brew-test",
+    "choco-test",
+    "npm pack",
+    "its-magic-*.tgz",
+    "installer.sh",
+    "packaging/chocolatey",
+    "packaging/homebrew",
+    "choco pack",
+    "brew style",
+)
+
+REASON_FORBIDDEN_PATTERN = "DOWNSTREAM_CI_FORBIDDEN_PATTERN"
+REASON_JOB_LEAK = "DOWNSTREAM_CI_JOB_LEAK"
+REASON_PACKAGING_MISSING = "KIT_CI_PACKAGING_JOBS_MISSING"
+
+REPORT_SCHEMA_VERSION = 1
+
+
+@dataclass
+class GuardViolation:
+    reason_code: str
+    detail: str
+
+
+@dataclass
+class GuardReport:
+    template_job_keys: List[str] = field(default_factory=list)
+    active_job_keys: List[str] = field(default_factory=list)
+    forbidden_hits: List[str] = field(default_factory=list)
+    violations: List[GuardViolation] = field(default_factory=list)
+    ok: bool = True
+
+
+def read_utf8(path: str) -> str:
+    with open(path, "r", encoding="utf-8") as f:
+        return f.read()
+
+
+def extract_job_keys(yaml_text: str) -> List[str]:
+    """Extract top-level job ids from a GitHub Actions workflow (stdlib regex only)."""
+    keys: List[str] = []
+    in_jobs = False
+    for line in yaml_text.splitlines():
+        if re.match(r"^jobs:\s*$", line):
+            in_jobs = True
+            continue
+        if not in_jobs:
+            continue
+        m = JOB_KEY_RE.match(line)
+        if m:
+            keys.append(m.group(1))
+        elif line and not line.startswith(" ") and not line.startswith("#"):
+            break
+    return keys
+
+
+def scan_forbidden_patterns(yaml_text: str) -> List[str]:
+    hits: List[str] = []
+    for pattern in FORBIDDEN_SUBSTRINGS:
+        if pattern in yaml_text:
+            hits.append(pattern)
+    return hits
+
+
+def check_template_ci(yaml_text: str) -> List[GuardViolation]:
+    violations: List[GuardViolation] = []
+    job_keys = extract_job_keys(yaml_text)
+    extra_jobs = set(job_keys) - ALLOWED_TEMPLATE_JOBS
+    if extra_jobs:
+        violations.append(
+            GuardViolation(
+                REASON_JOB_LEAK,
+                f"template ci.yml job keys {sorted(job_keys)!r} exceed allowed "
+                f"{sorted(ALLOWED_TEMPLATE_JOBS)!r}",
+            )
+        )
+    forbidden = scan_forbidden_patterns(yaml_text)
+    if forbidden:
+        violations.append(
+            GuardViolation(
+                REASON_FORBIDDEN_PATTERN,
+                f"template ci.yml contains forbidden pattern(s): {forbidden}",
+            )
+        )
+    return violations
+
+
+def check_active_ci(yaml_text: str) -> List[GuardViolation]:
+    violations: List[GuardViolation] = []
+    job_keys = set(extract_job_keys(yaml_text))
+    missing = REQUIRED_ACTIVE_JOBS - job_keys
+    if missing:
+        violations.append(
+            GuardViolation(
+                REASON_PACKAGING_MISSING,
+                f"active ci.yml missing required job id(s): {sorted(missing)!r}; "
+                f"found {sorted(job_keys)!r}",
+            )
+        )
+    return violations
+
+
+def build_report(repo_root: str) -> Tuple[GuardReport, List[str]]:
+    report = GuardReport()
+    stderr_lines: List[str] = []
+
+    template_ci = os.path.join(
+        repo_root, "template", ".github", "workflows", "ci.yml"
+    )
+    active_ci = os.path.join(repo_root, ".github", "workflows", "ci.yml")
+
+    if not os.path.isfile(template_ci):
+        stderr_lines.append(f"missing template ci.yml: {template_ci}")
+        report.ok = False
+        return report, stderr_lines
+    if not os.path.isfile(active_ci):
+        stderr_lines.append(f"missing active ci.yml: {active_ci}")
+        report.ok = False
+        return report, stderr_lines
+
+    template_text = read_utf8(template_ci)
+    active_text = read_utf8(active_ci)
+
+    report.template_job_keys = extract_job_keys(template_text)
+    report.active_job_keys = extract_job_keys(active_text)
+    report.forbidden_hits = scan_forbidden_patterns(template_text)
+
+    for v in check_template_ci(template_text):
+        report.violations.append(v)
+        stderr_lines.append(f"{v.reason_code}: {v.detail}")
+    for v in check_active_ci(active_text):
+        report.violations.append(v)
+        stderr_lines.append(f"{v.reason_code}: {v.detail}")
+
+    report.ok = len(report.violations) == 0
+    return report, stderr_lines
+
+
+def report_to_dict(report: GuardReport) -> Dict[str, Any]:
+    return {
+        "schema_version": REPORT_SCHEMA_VERSION,
+        "template_job_keys": report.template_job_keys,
+        "active_job_keys": report.active_job_keys,
+        "forbidden_hits": report.forbidden_hits,
+        "violations": [
+            {"reason_code": v.reason_code, "detail": v.detail}
+            for v in report.violations
+        ],
+        "ok": report.ok,
+    }
+
+
+def self_test() -> None:
+    sample_template = """name: ci
+jobs:
+  checks:
+    runs-on: ubuntu-latest
+  auto-fix:
+    runs-on: ubuntu-latest
+"""
+    keys = extract_job_keys(sample_template)
+    assert keys == ["checks", "auto-fix"], keys
+
+    bad_template = """name: ci
+jobs:
+  checks:
+    runs-on: ubuntu-latest
+  npm-test:
+    runs-on: ubuntu-latest
+    steps:
+      - run: npm pack
+"""
+    v = check_template_ci(bad_template)
+    assert any(x.reason_code == REASON_JOB_LEAK for x in v)
+    assert any(x.reason_code == REASON_FORBIDDEN_PATTERN for x in v)
+
+    good_active = """name: ci
+jobs:
+  checks:
+    runs-on: ubuntu-latest
+  auto-fix:
+    runs-on: ubuntu-latest
+  npm-test:
+    runs-on: ubuntu-latest
+  brew-test:
+    runs-on: ubuntu-latest
+  choco-test:
+    runs-on: ubuntu-latest
+"""
+    assert not check_active_ci(good_active)
+
+    bad_active = """name: ci
+jobs:
+  checks:
+    runs-on: ubuntu-latest
+  auto-fix:
+    runs-on: ubuntu-latest
+"""
+    v2 = check_active_ci(bad_active)
+    assert len(v2) == 1 and v2[0].reason_code == REASON_PACKAGING_MISSING