sansan0/TrendRadar · error · InvalidParameterError

日期太久远: {date.strftime('%Y-%m-%d')} ({days_ago}天前)

Error message

日期太久远: {date.strftime('%Y-%m-%d')} ({days_ago}天前)

What it means

Thrown by DateParser.validate_date_not_too_old when the parsed date is more than max_days (default 365) before today. Together with the not-future check it bounds queries to a rolling one-year window that matches the data retention period.

Source

Thrown at mcp_server/utils/date_parser.py:325

                f"不能查询未来的日期: {date.strftime('%Y-%m-%d')}",
                suggestion="请使用今天或过去的日期"
            )

    @staticmethod
    def validate_date_not_too_old(date: datetime, max_days: int = 365) -> None:
        """
        验证日期不太久远

        Args:
            date: 待验证的日期
            max_days: 最大天数

        Raises:
            InvalidParameterError: 日期太久远
        """
        days_ago = (datetime.now().date() - date.date()).days
        if days_ago > max_days:
            raise InvalidParameterError(
                f"日期太久远: {date.strftime('%Y-%m-%d')} ({days_ago}天前)",
                suggestion=f"请查询{max_days}天内的数据"
            )

    @staticmethod
    def resolve_date_range_expression(expression: str) -> Dict:
        """
        将自然语言日期表达式解析为标准日期范围

        这是专门为 MCP 工具设计的方法，用于在服务器端解析日期表达式，
        避免 AI 模型自己计算日期导致的不一致问题。

        Args:
            expression: 自然语言日期表达式，支持：
                - 单日: "今天", "昨天", "today", "yesterday"
                - 本周/上周: "本周", "上周", "this week", "last week"
                - 本月/上月: "本月", "上月", "this month", "last month"
                - 最近N天: "最近7天", "最近30天", "last 7 days", "last 30 days"

View on GitHub (pinned to 8ee26026ba)

Solutions

1. Query dates within the last 365 days
2. If you control the call and retention allows, pass a larger max_days explicitly
3. Split older ranges out or accept they are outside the dataset's coverage

Example fix

# before
DateParser.validate_date_not_too_old(datetime(2020, 1, 1))  # raises, default 365

# after
DateParser.validate_date_not_too_old(datetime(2020, 1, 1), max_days=3650)  # if retention allows
# or query within the last year:
DateParser.validate_date_not_too_old(datetime.now() - timedelta(days=200))

from datetime import datetime, timedelta

def within_retention(d: datetime, max_days: int = 365) -> bool:
    return (datetime.now().date() - d.date()).days <= max_days

def is_within_max_days(d, max_days=365) -> bool:
    from datetime import datetime
    return (datetime.now().date() - d.date()).days <= max_days

try:
    DateParser.validate_date_not_too_old(d)
except InvalidParameterError as e:
    if "太久远" in str(e):
        d = datetime.now() - timedelta(days=365)  # clamp to retention window
    else:
        raise

When it happens

Trigger: Calling the validator (directly or via a tool that enforces the window) with a date more than 365 days in the past; long-horizon historical research queries; archived dates from previous years.

Common situations: Backfill scripts targeting older data; users asking for '去年今天' near the boundary; forgetting the default max_days is 365 when the tool only stores a year.

Related errors

• INVALID_PARAMETER
• 不能查询未来的日期: {date.strftime('%Y-%m-%d')}
• 日期表达式不能为空
• 日期格式错误: {date_str}
• 日期格式错误: {stripped}

AI-assisted analysis of sansan0/TrendRadar@8ee26026ba (2026-08-15). Data as JSON: /api/errors/e2c9616b596c4ec8. Report an issue: GitHub.