daily-robots/tests/test_news_fetch_window.py

"""Tests for news time window filtering."""

from __future__ import annotations

import os
import unittest
from datetime import datetime, timezone, timedelta
from unittest.mock import patch
from zoneinfo import ZoneInfo

from daily.news.fetch import _cutoff_datetime, _parse_datetime, _within_window


class NewsWindowTests(unittest.TestCase):
    def test_cutoff_floor_today_excludes_yesterday_even_within_24h(self):
        tz = ZoneInfo("Asia/Shanghai")
        # 2026-07-09 09:00 CST = 2026-07-09 01:00 UTC
        fixed = datetime(2026, 7, 9, 1, 0, tzinfo=timezone.utc)
        with patch("daily.news.fetch._now_utc", return_value=fixed):
            with patch.dict(os.environ, {"DAILY_AI_NEWS_HOURS": "24"}, clear=False):
                cutoff = _cutoff_datetime(floor_today=True)
        start_today_cst = datetime(2026, 7, 9, 0, 0, tzinfo=tz).astimezone(timezone.utc)
        self.assertEqual(cutoff, start_today_cst)
        yesterday = datetime(2026, 7, 8, 20, 0, tzinfo=tz).astimezone(timezone.utc)
        self.assertFalse(_within_window({"published": yesterday.isoformat()}, cutoff))

    def test_cutoff_rolling_only_includes_last_24h(self):
        fixed = datetime(2026, 7, 9, 12, 0, tzinfo=timezone.utc)
        with patch("daily.news.fetch._now_utc", return_value=fixed):
            with patch.dict(os.environ, {"DAILY_AI_NEWS_HOURS": "24"}, clear=False):
                cutoff = _cutoff_datetime(floor_today=False)
        self.assertEqual(cutoff, fixed - timedelta(hours=24))

    def test_within_window_rejects_missing_datetime(self):
        cutoff = datetime(2026, 7, 9, 0, 0, tzinfo=timezone.utc)
        self.assertFalse(_within_window({"title": "x", "link": "https://a.com"}, cutoff))

    def test_parse_date_only_uses_local_noon(self):
        with patch.dict(os.environ, {"DAILY_AI_NEWS_TZ": "Asia/Shanghai"}, clear=False):
            dt = _parse_datetime("2026-07-09")
        self.assertIsNotNone(dt)
        assert dt is not None
        local = dt.astimezone(ZoneInfo("Asia/Shanghai"))
        self.assertEqual(local.hour, 12)


class NewsFormatTests(unittest.TestCase):
    def test_ai_news_lines_use_desc_as_link_text(self):
        from daily.format_wecom import _ai_news_lines

        lines = _ai_news_lines(
            [
                {
                    "title": "English Title",
                    "link": "https://example.com/a",
                    "source_name": "Src",
                    "published_fmt": "07-11",
                    "desc_short": "中文摘要一句",
                }
            ]
        )
        self.assertEqual(len(lines), 1)
        self.assertIn("[中文摘要一句](https://example.com/a)", lines[0])
        self.assertNotIn("English Title", lines[0])
        self.assertNotIn("> ", lines[0])

    def test_ai_news_lines_fallback_to_title(self):
        from daily.format_wecom import _ai_news_lines

        lines = _ai_news_lines(
            [
                {
                    "title": "仅标题",
                    "link": "https://example.com/b",
                    "source_name": "Src",
                    "published_fmt": "",
                    "desc_short": "",
                }
            ]
        )
        self.assertIn("[仅标题](https://example.com/b)", lines[0])


class NewsSummaryTests(unittest.TestCase):
    def test_brief_news_summary_no_ellipsis(self):
        from daily.news.fetch import brief_news_summary

        text = (
            "Meta told Dylan Byers, of Puck News, that the company removed "
            "the controversial AI feature after user backlash on Instagram."
        )
        out = brief_news_summary(text, limit=72)
        self.assertNotIn("...", out)
        self.assertLessEqual(len(out), 72)
        self.assertTrue(out.startswith("Meta told"))

    def test_brief_news_summary_filters_junk(self):
        from daily.news.fetch import brief_news_summary

        self.assertEqual(brief_news_summary("点击查看原文>"), "")
        self.assertEqual(brief_news_summary("Article URL: https://example.com"), "")

    def test_sync_wecom_news_rows_after_localize(self):
        from daily.news.fetch import _to_wecom_news_row, sync_wecom_news_rows

        row = _to_wecom_news_row(
            {
                "title": "t",
                "link": "https://a.com/x",
                "source_name": "s",
                "published_fmt": "07-11",
                "summary": "Short english stub that was truncated early...",
            }
        )
        flat = [
            {
                "link": "https://a.com/x",
                "summary": "苹果指控 OpenAI 窃取硬件商业机密，诉讼称 misconduct 涉及多名前员工。",
            }
        ]
        sync_wecom_news_rows([row], flat)
        self.assertNotIn("...", row["desc_short"])
        self.assertIn("苹果", row["desc_short"])


    def test_finalize_wecom_news_forces_chinese(self):
        from daily.news.fetch import finalize_wecom_news_items

        items = [
            {
                "link": "https://a.com/1",
                "desc_short": "Meta removed the feature after backlash.",
                "summary_plain": "Meta removed the feature after backlash.",
            }
        ]
        with patch(
            "daily.localize.localize_brief_descriptions",
            return_value={"wecom-news:https://a.com/1": "Meta 在舆论压力下移除了该功能"},
        ):
            finalize_wecom_news_items(items, force_chinese=True)
        self.assertIn("Meta", items[0]["desc_short"])
        self.assertNotIn("backlash", items[0]["desc_short"])


class NewsPickTests(unittest.TestCase):
    def test_pick_and_backfill_to_limit(self):
        from daily.news.fetch import _fill_picked_to_limit, _pick_news_items

        flat = [
            {"link": f"https://a.com/{i}", "title": f"t{i}", "category_id": "media", "summary": "s"}
            for i in range(12)
        ]
        picked = _pick_news_items(flat, 10, ("media",))
        self.assertEqual(len(picked), 10)
        picked = _fill_picked_to_limit(picked[:3], [flat], 10)
        self.assertEqual(len(picked), 10)