cmshoppe/app/product_status.py

"""Shared product-status rules for the collection, generation, and apply flows."""

from __future__ import annotations

import hashlib
import json
import re


STATUS_NORMAL = "normal"
STATUS_UNLISTED = "unlisted"
STATUS_REVIEWING = "reviewing"
STATUS_UNKNOWN = "unknown"

VALID_PRODUCT_STATUSES = frozenset(
    {
        STATUS_NORMAL,
        STATUS_UNLISTED,
        STATUS_REVIEWING,
        STATUS_UNKNOWN,
    }
)

PRODUCT_STATUS_LABELS = {
    STATUS_NORMAL: "正常",
    STATUS_UNLISTED: "未上架",
    STATUS_REVIEWING: "审核中",
    STATUS_UNKNOWN: "状态未知",
}

SCOPE_NORMAL_ONLY = "normal_only"
SCOPE_ALL = "all"

# Kept as explicit collection aliases while generation and apply adopt the
# shared scope values in their own tasks.
COLLECT_SCOPE_NORMAL_ONLY = SCOPE_NORMAL_ONLY
COLLECT_SCOPE_ALL = SCOPE_ALL

_WHITESPACE_RE = re.compile(r"\s+")
_NOTE_LIMIT = 2000


def normalize_status(value) -> str:
    """Return a stable status code; legacy and invalid values are unknown."""

    normalized = str(value or "").strip().lower()
    if normalized in VALID_PRODUCT_STATUSES:
        return normalized
    return STATUS_UNKNOWN


def status_label(value) -> str:
    return PRODUCT_STATUS_LABELS[normalize_status(value)]


def is_normal(value) -> bool:
    return normalize_status(value) == STATUS_NORMAL


def is_known_abnormal(value) -> bool:
    return normalize_status(value) in {STATUS_UNLISTED, STATUS_REVIEWING}


def normalize_text(value) -> str:
    return _WHITESPACE_RE.sub(" ", str(value or "").strip())


def classify_alerts(alerts) -> dict:
    """Classify normalized EDS warning-alert data without retaining page HTML."""

    if not isinstance(alerts, list):
        return _snapshot(STATUS_UNKNOWN, "")
    if not alerts:
        return _snapshot(STATUS_NORMAL, "")

    first_note = ""
    valid_alert_seen = False
    for alert in alerts:
        if not isinstance(alert, dict):
            continue
        valid_alert_seen = True
        title = normalize_text(alert.get("title"))
        description = normalize_text(alert.get("description"))
        note = _alert_note(title, description)
        if not first_note:
            first_note = note
        if title in {"審核中", "审核中"}:
            return _snapshot(STATUS_REVIEWING, note)
        if "您的商品未上架" in title:
            return _snapshot(STATUS_UNLISTED, note)

    if not valid_alert_seen:
        return _snapshot(STATUS_UNKNOWN, "")
    return _snapshot(STATUS_UNKNOWN, first_note)


def partition_tasks(tasks) -> dict:
    """Return tasks grouped by normalized status for callers' run planners."""

    grouped = {status: [] for status in VALID_PRODUCT_STATUSES}
    for task in tasks:
        grouped[normalize_status(getattr(task, "product_status", None))].append(task)
    return grouped


def build_generation_plan(tasks, generate_mode, scope=SCOPE_NORMAL_ONLY) -> dict:
    """Build a frozen, status-aware generation plan without mutating tasks."""

    base_candidates = _deduplicate_tasks(tasks)
    status_counts = {status: 0 for status in VALID_PRODUCT_STATUSES}
    for task in base_candidates:
        status_counts[_task_status(task)] += 1

    normalized_scope = normalize_scope(scope)
    if normalized_scope == SCOPE_ALL:
        execution_tasks = list(base_candidates)
    else:
        execution_tasks = [
            task for task in base_candidates if is_normal(_task_status(task))
        ]

    return {
        "base_candidates": base_candidates,
        "status_counts": status_counts,
        "execution_tasks": execution_tasks,
        "scope": normalized_scope,
        "scope_excluded": len(base_candidates) - len(execution_tasks),
        "fingerprint": _task_plan_fingerprint(base_candidates, generate_mode),
    }


def build_apply_plan(tasks, update_mode) -> dict:
    """Return a status-first, content-aware plan for Shopee update actions."""

    base_candidates = _deduplicate_tasks(tasks)
    status_counts = {status: 0 for status in VALID_PRODUCT_STATUSES}
    status_excluded = {
        STATUS_UNLISTED: [],
        STATUS_REVIEWING: [],
        STATUS_UNKNOWN: [],
        "unverified": [],
    }
    normal_candidates = []
    for task in base_candidates:
        status = _task_status(task)
        status_counts[status] += 1
        if status == STATUS_NORMAL:
            normal_candidates.append(task)
        elif status == STATUS_UNKNOWN and not _has_status_snapshot(task):
            status_excluded["unverified"].append(task)
        else:
            status_excluded[status].append(task)

    mode = str(update_mode or "").strip().lower()
    requires_title = mode in {"title", "title_cover"}
    requires_cover = mode in {"cover", "title_cover"}
    missing_title = [
        task
        for task in normal_candidates
        if requires_title and not str(_task_value(task, "new_title") or "").strip()
    ]
    missing_cover = [
        task
        for task in normal_candidates
        if requires_cover and not str(_task_value(task, "new_cover_path") or "").strip()
    ]
    content_excluded_ids = {id(task) for task in missing_title + missing_cover}
    executable_tasks = [
        task for task in normal_candidates if id(task) not in content_excluded_ids
    ]
    status_excluded_count = sum(len(rows) for rows in status_excluded.values())

    return {
        "base_candidates": base_candidates,
        "status_counts": status_counts,
        "executable": executable_tasks,
        "unlisted": status_excluded[STATUS_UNLISTED],
        "reviewing": status_excluded[STATUS_REVIEWING],
        "unknown": status_excluded[STATUS_UNKNOWN],
        "unverified": status_excluded["unverified"],
        "missing_title": missing_title,
        "missing_cover": missing_cover,
        "status_scope_excluded": status_excluded_count,
        "content_scope_excluded": len(content_excluded_ids),
        "scope_excluded": status_excluded_count + len(content_excluded_ids),
        "fingerprint": _task_plan_fingerprint(base_candidates, update_mode),
    }


def normalize_collect_scope(value) -> str:
    return normalize_scope(value)


def normalize_scope(value) -> str:
    value = str(value or "").strip().lower()
    return SCOPE_ALL if value == SCOPE_ALL else SCOPE_NORMAL_ONLY


def should_collect_content(status, scope) -> bool:
    return normalize_collect_scope(scope) == COLLECT_SCOPE_ALL or is_normal(status)


def collect_skip_reason(status) -> str:
    return f"{status_label(status)}，按本轮范围略过"


def _deduplicate_tasks(tasks) -> list:
    seen = set()
    unique = []
    for task in list(tasks or []):
        task_id = _task_value(task, "id")
        key = ("id", str(task_id)) if task_id is not None else ("object", id(task))
        if key in seen:
            continue
        seen.add(key)
        unique.append(task)
    return unique


def _task_plan_fingerprint(tasks, mode) -> str:
    snapshots = [
        {
            "task_id": _task_value(task, "id"),
            "updated_at": _task_value(task, "updated_at"),
            "product_status": _task_status(task),
            "new_title": _task_value(task, "new_title"),
            "new_cover_path": _task_value(task, "new_cover_path"),
            "mode": str(mode or ""),
        }
        for task in tasks
    ]
    encoded = json.dumps(
        snapshots,
        ensure_ascii=False,
        sort_keys=True,
        separators=(",", ":"),
        default=str,
    ).encode("utf-8")
    return hashlib.sha256(encoded).hexdigest()


def _task_status(task) -> str:
    return normalize_status(_task_value(task, "product_status"))


def _has_status_snapshot(task) -> bool:
    return bool(str(_task_value(task, "product_status") or "").strip())


def _task_value(task, name, default=None):
    if isinstance(task, dict):
        return task.get(name, default)
    return getattr(task, name, default)


def _alert_note(title: str, description: str) -> str:
    parts = []
    if title:
        parts.append(f"标题：{title}")
    if description:
        parts.append(f"说明：{description}")
    return "；".join(parts)[:_NOTE_LIMIT]


def _snapshot(status: str, note: str) -> dict:
    return {
        "product_status": normalize_status(status),
        "product_status_note": str(note or "")[:_NOTE_LIMIT] or None,
    }