QuantEngineByItz/tools/validate_gas_thin_adapter_v1.py

from __future__ import annotations

import json
from pathlib import Path

import yaml

from refactor_master_helpers import ROOT, collect_gas_files, read_text
try:
    from audit_gas_business_logic_v1 import _classify, _function_bodies
except Exception:  # pragma: no cover - fallback for isolated execution
    _classify = None
    _function_bodies = None


ALLOWLIST = ("collect", "normalize", "export", "display")
FORBIDDEN = ("decision", "sizing", "stop_loss", "take_profit", "risk_score")

# Non-violation classifications — lines in these categories are not forbidden.
_ADAPTER_OK_CLASSES = {"ADAPTER_OK", "COMMENT_FALSE_POSITIVE"}


# P5-T02: gas_data_feed.gs split mapping (original_line → new_file, new_line)
_GDF_SPLITS = [
    (1, 2347, "gdf_01_price_metrics.gs", 0),
    (2348, 4560, "gdf_02_harness_assembly.gs", 2347),
    (4561, 6806, "gdf_03_portfolio_gates.gs", 4560),
    (6807, 9015, "gdf_04_execution_quality.gs", 6806),
    (9016, 10302, "gdf_05_alpha_engines.gs", 9015),
]
_GDC_SPLITS = [
    (1, 2405, "gdc_01_fetch_fundamentals.gs", 0),
    (2406, 4460, "gdc_02_account_satellite.gs", 2405),
]


def _remap_line(fname: str, lineno: int) -> list[tuple[str, int]]:
    """Return all (file, lineno) pairs that represent the given original location.

    After P5-T02 file split, a line originally in gas_data_feed.gs or
    gas_data_collect.gs may now live in a split adapter part file with an
    offset line number. Returns both the original location and the new one.
    """
    results = [(fname, lineno)]
    if fname == "gas_data_feed.gs":
        for lo, hi, new_file, offset in _GDF_SPLITS:
            if lo <= lineno <= hi:
                results.append((new_file, lineno - offset))
                break
    elif fname == "gas_data_collect.gs":
        for lo, hi, new_file, offset in _GDC_SPLITS:
            if lo <= lineno <= hi:
                results.append((new_file, lineno - offset))
                break
    return results


def _load_classified_allowlist() -> set[tuple[str, int]]:
    """Load (file, lineno) pairs classified as ADAPTER_OK or COMMENT_FALSE_POSITIVE.

    Reads runtime/gas_migration_wave1.yaml and runtime/gas_migration_wave2_4.yaml.
    After P5-T02 file split, original line numbers are remapped to adapter part files.
    """
    allowlist: set[tuple[str, int]] = set()
    wave_files = [
        ROOT / "runtime" / "gas_migration_wave1.yaml",
        ROOT / "runtime" / "gas_migration_wave2_4.yaml",
    ]
    for wf in wave_files:
        if not wf.exists():
            continue
        try:
            data = yaml.safe_load(wf.read_text(encoding="utf-8")) or {}
        except Exception:
            continue
        for item in data.get("wave1_items", []) + data.get("wave_items", []):
            cls = item.get("classification", "")
            if cls in _ADAPTER_OK_CLASSES:
                fname = item.get("file", "")
                lineno = item.get("line")
                if fname and isinstance(lineno, int):
                    for mapped_fname, mapped_line in _remap_line(fname, lineno):
                        allowlist.add((mapped_fname, mapped_line))
    return allowlist


def main() -> int:
    policy_path = ROOT / "spec" / "39_gas_thin_adapter_policy.yaml"
    migration_plan_exists = False
    if policy_path.exists():
        policy = yaml.safe_load(policy_path.read_text(encoding="utf-8")) or {}
        migration_plan = policy.get("migration_plan")
        migration_plan_exists = isinstance(migration_plan, dict) and bool(migration_plan.get("phases"))
    audit_path = ROOT / "Temp" / "gas_business_logic_audit_v1.json"
    audit = {}
    if audit_path.exists():
        try:
            audit = json.loads(audit_path.read_text(encoding="utf-8"))
        except Exception:
            audit = {}
    if not audit and _function_bodies and _classify:
        rows = []
        for path in collect_gas_files():
            for name, line, body in _function_bodies(path):
                allowed_responsibility, matched_tokens = _classify(name, body, path.name)
                rows.append(
                    {
                        "file": str(path.relative_to(ROOT)),
                        "name": name,
                        "line": line,
                        "allowed_responsibility": allowed_responsibility,
                        "matched_tokens": matched_tokens,
                    }
                )
        forbidden_function_count = sum(1 for row in rows if row["allowed_responsibility"] == "forbidden")
        function_inventory_coverage_pct = 100.0 if rows else 0.0
        sample_findings = rows[:200]
    elif audit:
        # audit_gas_thin_adapter_v1.py (v2) 출력 형식 지원:
        # summary.forbidden_count, summary.compliance_pct, migration_candidates
        summary = audit.get("summary", {})
        if summary:
            forbidden_function_count = int(summary.get("forbidden_count") or 0)
            total_functions = int(summary.get("total_functions") or 0)
            function_inventory_coverage_pct = 100.0 if total_functions > 0 else 0.0
            candidates = audit.get("migration_candidates", [])
            sample_findings = [
                {"file": c.get("file",""), "name": c.get("name",""),
                 "responsibility": c.get("responsibility",[])}
                for c in candidates[:200]
            ]
        else:
            forbidden_function_count = int(audit.get("forbidden_function_count") or 0)
            function_inventory_coverage_pct = float(audit.get("function_inventory_coverage_pct") or 0.0)
            sample_findings = audit.get("rows")[:200] if isinstance(audit.get("rows"), list) else []
    else:
        classified_allowlist = _load_classified_allowlist()
        findings: list[dict[str, str]] = []
        for path in collect_gas_files():
            text = read_text(path)
            rel_name = path.name  # scanner uses filename only (not relative path)
            for lineno, line in enumerate(text.splitlines(), start=1):
                low = line.lower()
                if not any(word in low for word in FORBIDDEN):
                    continue
                # Skip lines confirmed as ADAPTER_OK or COMMENT_FALSE_POSITIVE in wave YAMLs.
                if (rel_name, lineno) in classified_allowlist:
                    continue
                findings.append({"file": str(path.relative_to(ROOT)), "line": str(lineno), "text": line.strip()})
        forbidden_function_count = len(findings)
        function_inventory_coverage_pct = 0.0 if not findings else 100.0
        sample_findings = findings[:200]
    result = {
        "formula_id": "GAS_THIN_ADAPTER_V1",
        "forbidden_gas_business_logic_count": forbidden_function_count,
        "function_inventory_coverage_pct": function_inventory_coverage_pct,
        "migration_plan_exists": migration_plan_exists,
        "findings": sample_findings,
        "gate": "PASS" if (function_inventory_coverage_pct >= 100.0 and migration_plan_exists) else "FAIL",
    }
    out = ROOT / "Temp" / "gas_thin_adapter_validation_v1.json"
    out.write_text(json.dumps(result, ensure_ascii=False, indent=2), encoding="utf-8")
    print(json.dumps(result, ensure_ascii=True, indent=2))
    return 0 if result["gate"] == "PASS" else 1


if __name__ == "__main__":
    raise SystemExit(main())