scripts/verify_project.py - master - Chaizee/ZenithCode_Incident-LLM-analytics

scripts/verify_project.py
147 строк4 KB
Chaizee
fix: fix speed
12 июн 2026, 16:28
12 июн 2026, 16:28b2745cc
#!/usr/bin/env python3
from __future__ import annotations

import ast
import compileall
import sys
from pathlib import Path

ROOT = Path(__file__).resolve().parent.parent
SKIP_DIRS = {".venv", "venv", "__pycache__", ".git"}

if str(ROOT) not in sys.path:
    sys.path.insert(0, str(ROOT))


def _py_files() -> list[Path]:
    out: list[Path] = []
    for path in ROOT.rglob("*.py"):
        if any(part in SKIP_DIRS for part in path.parts):
            continue
        out.append(path)
    return sorted(out)


def _check_future_imports(path: Path, source: str) -> list[str]:
    errors: list[str] = []
    lines = source.splitlines()
    seen_non_future = False
    for i, line in enumerate(lines, 1):
        stripped = line.strip()
        if not stripped or stripped.startswith("#"):
            continue
        if stripped.startswith("from __future__ import"):
            if seen_non_future:
                errors.append(f"{path}:{i}: from __future__ must be at file start")
            continue
        if stripped.startswith(('"""', "'''")) and not seen_non_future:
            continue
        seen_non_future = True
    return errors


def _check_legacy_files() -> list[str]:
    errors: list[str] = []
    legacy = [
        ROOT / "config.py",
        ROOT / "src" / "data.py",
        ROOT / "llm.py",
        ROOT / "src" / "pipeline.py",
        ROOT / "src" / "geo_map.py",
        ROOT / "src" / "geo_coords.py",
        ROOT / "src" / "chatbot.py",
        ROOT / "src" / "settings.py",
        ROOT / "scripts" / "strip_comments.py",
        ROOT / "requirements_hybrid.txt",
    ]
    for path in legacy:
        if path.is_file():
            errors.append(f"legacy file must be removed: {path.relative_to(ROOT)}")
    init_py = ROOT / "config" / "__init__.py"
    if init_py.is_file():
        text = init_py.read_text(encoding="utf-8")
        if text.count("from __future__ import") > 0:
            errors.append("config/__init__.py must not contain from __future__ (use config/settings.py)")
        if len(text.splitlines()) > 5:
            errors.append("config/__init__.py must only re-export config.settings")
    return errors


def _check_config_exports() -> list[str]:
    errors: list[str] = []
    try:
        from config import (  # noqa: PLC0415
            APP_ICON,
            APP_TITLE,
            EXCEL_COLUMN_MAP,
            EXCEL_HEADERS,
            ROOT as PROJECT_ROOT,
            USE_GPU,
        )
    except Exception as exc:
        return [f"config import failed: {exc}"]
    if PROJECT_ROOT != ROOT:
        errors.append(f"config.ROOT mismatch: {PROJECT_ROOT} != {ROOT}")
    for name in ("EXCEL_COLUMN_MAP", "EXCEL_HEADERS", "APP_TITLE", "APP_ICON", "USE_GPU"):
        if name not in dir():
            errors.append(f"config missing export: {name}")
    return errors


def _check_app_imports() -> list[str]:
    errors: list[str] = []
    modules = (
        "app",
        "src.data_polars",
        "src.pipeline_hybrid",
        "src.ui_results",
        "src.ml_classifier",
        "src.local_llm",
    )
    for mod in modules:
        try:
            __import__(mod)
        except ModuleNotFoundError as exc:
            optional = {
                "polars", "pandas", "streamlit", "torch", "torchvision", "loguru", "joblib",
                "numpy", "sklearn", "sentence_transformers", "xgboost",
                "llama_cpp", "geopandas", "folium", "plotly", "openpyxl",
            }
            if exc.name in optional:
                continue
            errors.append(f"import {mod}: {exc}")
        except ImportError as exc:
            errors.append(f"import {mod}: {exc}")
    return errors


def main() -> int:
    errors: list[str] = []
    errors.extend(_check_legacy_files())

    for path in _py_files():
        source = path.read_text(encoding="utf-8")
        errors.extend(_check_future_imports(path, source))
        try:
            ast.parse(source, filename=str(path))
        except SyntaxError as exc:
            errors.append(f"syntax error {path}: {exc}")

    if not compileall.compile_dir(str(ROOT), quiet=1):
        errors.append("compileall failed")

    errors.extend(_check_config_exports())
    errors.extend(_check_app_imports())

    if errors:
        print("PROJECT CHECK FAILED:")
        for err in errors:
            print(f"  - {err}")
        return 1

    print("PROJECT CHECK OK")
    return 0


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