GTI_GERENCIADOR_FISCAL/fiscal/central/app/installation_identity.py

"""Stable installation identity and safe operational metadata helpers."""

from __future__ import annotations

import hashlib
import re
import uuid
from datetime import datetime, timezone
from pathlib import Path

from .config import settings

SECRET_KEYS = {"password", "fb_password", "token", "agent_token", "api_key", "secret", "authorization", "connection_string"}


def utcnow() -> datetime:
    return datetime.now(timezone.utc)


def ensure_installation_id(state_path: str, *, new_installation: bool = False) -> str:
    """Persist a UUID beside the local Agent state without using hostname/IP."""
    path = Path(state_path)
    identity_path = path.with_suffix(path.suffix + ".installation_id")
    if identity_path.exists() and not new_installation:
        value = identity_path.read_text(encoding="utf-8").strip()
        try:
            return str(uuid.UUID(value))
        except ValueError:
            pass
    value = str(uuid.uuid4())
    identity_path.parent.mkdir(parents=True, exist_ok=True)
    temporary = identity_path.with_suffix(identity_path.suffix + ".tmp")
    temporary.write_text(value + "\n", encoding="utf-8")
    temporary.replace(identity_path)
    return value


def mask_cnpj(value: str | None) -> str | None:
    digits = re.sub(r"\D", "", value or "")
    if len(digits) != 14:
        return None if not value else "***"
    return f"{digits[:2]}.***.***/{digits[8:12]}-{digits[12:]}"


def sanitize_database_path(value: str | None) -> str | None:
    if not value:
        return None
    normalized = str(value).replace("\\", "/")
    return normalized.rsplit("/", 1)[-1][:255]


def _safe_value(key: str, value):
    if key.lower() in SECRET_KEYS:
        return "[REDACTED]"
    if isinstance(value, dict):
        return {str(k): _safe_value(str(k), v) for k, v in value.items() if str(k).lower() not in SECRET_KEYS}
    if isinstance(value, list):
        return [_safe_value(key, item) for item in value[:100]]
    return value


def sanitize_registration(payload: dict) -> dict:
    allowed = {
        "installation_id", "display_name", "customer_code", "notes", "hostname", "windows_version", "machine_fingerprint",
        "company_code", "company_name", "company_cnpj", "company_uf", "company_crt", "company_contribuinte_ipi", "firebird_version", "firebird_ods",
        "database_identifier", "database_path", "capabilities", "fiscal_engine_version", "dataset_versions", "started_at",
        "status", "schema_fingerprint", "agent_version", "last_job_id", "coverage", "health", "event_type", "message", "context",
    }
    result = {key: _safe_value(key, value) for key, value in payload.items() if key in allowed}
    if "company_cnpj" in result:
        result["company_cnpj_masked"] = mask_cnpj(result.pop("company_cnpj"))
    if "database_path" in result:
        result["database_identifier"] = sanitize_database_path(result.pop("database_path"))
    if result.get("installation_id"):
        result["installation_id"] = str(uuid.UUID(str(result["installation_id"])))
    return result


def compute_agent_health(snapshot, now: datetime | None = None, *, supplemental=None) -> tuple[str, str]:
    now = now or utcnow()
    last_seen = getattr(snapshot, "last_seen_at", None)
    if last_seen is None:
        return "UNKNOWN", "no heartbeat received"
    if last_seen.tzinfo is None:
        last_seen = last_seen.replace(tzinfo=timezone.utc)
    source = supplemental or snapshot
    if any(getattr(candidate, "agent_health", None) == "FAILED" or getattr(candidate, "last_status", None) == "FAILED" for candidate in (snapshot, supplemental) if candidate is not None):
        return "FAILED", "persistent technical failure"
    offline_after = settings.agent_offline_after_seconds or settings.agent_heartbeat_interval_seconds * 3
    if (now - last_seen).total_seconds() > offline_after:
        return "OFFLINE", f"last heartbeat exceeds {offline_after}s"
    if getattr(source, "failed_products", 0) or getattr(source, "last_job_status", "") == "FAILED":
        return "DEGRADED", "recent technical or job failure"
    return "ONLINE", "recent heartbeat without critical technical failure"


_VERSION_RE = re.compile(r"^v?(\d+(?:\.\d+)*)(?:[-_.]?(rc|beta|alpha|b|a)[-_.]?(\d*))?$", re.IGNORECASE)
_PRE_RANK = {"alpha": -3, "a": -3, "beta": -2, "b": -2, "rc": -1}


def parse_version(value: str | None):
    """Comparable tuple for '1.1.7', '1.0.0rc1', ... (a pre-release sorts below its final); None if unparsable."""
    match = _VERSION_RE.match((value or "").strip())
    if not match:
        return None
    parts = [int(x) for x in match.group(1).split(".")]
    while len(parts) < 4:
        parts.append(0)
    pre = (_PRE_RANK[match.group(2).lower()], int(match.group(3) or 0)) if match.group(2) else (0, 0)
    return (*parts, *pre)


def reference_agent_version(session) -> str:
    """The version an Agent should be running: the highest PUBLISHED Windows Agent release (REVOKED/DRAFT are
    never a target). Falls back to the configured server version when no release is published."""
    from sqlalchemy import select

    from .models import AgentRelease

    best, best_key = None, None
    for version in session.scalars(select(AgentRelease.version).where(AgentRelease.status == "PUBLISHED", AgentRelease.platform == "windows")):
        key = parse_version(version)
        if key is not None and (best_key is None or key > best_key):
            best, best_key = version, key
    return best or settings.agent_server_version


def compute_version_drift(agent_version: str | None, reference: str | None = None) -> str:
    """CURRENT when the Agent is at (or ahead of) the reference release, OUTDATED when behind it."""
    if not agent_version:
        return "UNKNOWN"
    reference = reference or settings.agent_server_version
    have, want = parse_version(agent_version), parse_version(reference)
    if have is None or want is None:
        return "CURRENT" if agent_version == reference else "OUTDATED"
    return "CURRENT" if have >= want else "OUTDATED"


def outdated_versions(session, reference: str) -> list[str]:
    """Distinct Agent versions in use that are behind the reference (used to keep SQL filters exact)."""
    from sqlalchemy import select

    from .models import Agent

    used = [v for v in session.scalars(select(Agent.agent_version).where(Agent.agent_version.is_not(None)).distinct())]
    return [v for v in used if compute_version_drift(v, reference) == "OUTDATED"]