mpdt 0.2.2__tar.gz → 0.2.4__tar.gz

{mpdt-0.2.2 → mpdt-0.2.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mpdt
-Version: 0.2.2
+Version: 0.2.4
 Summary: 公众号运营数据本地 CLI：把后台三源数据抓回本机 SQLite，一键生成周报
 Author: ttxs69
 License: MIT

{mpdt-0.2.2 → mpdt-0.2.4}/mpdt/__init__.py

@@ -6,4 +6,4 @@
   3 凭据问题（token 刷新失败）
   4 上游接口异常（结构变更，含原始返回片段）
 """
-__version__ = '0.2.2'
+__version__ = '0.2.4'

{mpdt-0.2.2 → mpdt-0.2.4}/mpdt/cli.py

@@ -100,9 +100,21 @@ def cmd_fetch(args):
     pl = publish_list.fetch(s, count=args.count)
     for i, a in enumerate(pl['articles'], 1):
         r.progress(i, len(pl['articles']), f'入库 {a["title"][:24]}')
+        # R101：create_time/publish_date/send_hour 元数据随行入库（C20 publish_window
+        # 消费链的 L1 缺口——解析层早有、入库曾丢弃）。send_hour=群发时刻北京时间
+        # HH:MM；上游缺 create_time 时三键均存 null，不虚构。
+        ct = a.get('create_time')
+        send_hour = None
+        if isinstance(ct, int) and ct > 0:
+            send_hour = dt.datetime.fromtimestamp(
+                ct, dt.timezone(dt.timedelta(hours=8))).strftime('%H:%M')
         store.upsert(con, 'publish_list', stat_date, a['appmsgid'], a['itemidx'],
-                     a['title'], {k: a[k] for k in
-                                  ('read_num', 'share_num', 'like_num', 'old_like_num', 'comment_num')})
+                     a['title'], {**{k: a.get(k) for k in
+                                     ('read_num', 'share_num', 'like_num',
+                                      'old_like_num', 'comment_num')},
+                                  'create_time': ct if isinstance(ct, int) and ct > 0 else None,
+                                  'publish_date': a.get('publish_date') or None,
+                                  'send_hour': send_hour})
     n_total += len(pl['articles'])
     store.log_run(con, 'fetch:publish_list', True, f"{len(pl['articles'])} articles")
     r.say(f'拉取 datacube（{begin} ~ {end}，T+1 口径）…')

{mpdt-0.2.2 → mpdt-0.2.4}/mpdt/fetchers/publish_list.py

@@ -53,6 +53,11 @@ def parse_publish_list(raw: str) -> dict:
         if not isinstance(info, dict):
             continue
         ct = (info.get('publish_info') or {}).get('create_time')
+        if not (isinstance(ct, int) and ct > 0):
+            # 贴图类（type=9 群发图）publish_info 为空对象，群发时间在 sent_info.time
+            # （R70 实证：2247485215 sent_info.time=1792145145 = 2026-10-16 18:05:45+08，
+            #  与 publish-read.py 实查 create_time 一致；保持 article 侧口径=群发时刻）
+            ct = (info.get('sent_info') or {}).get('time')
         pub_date = ''
         if isinstance(ct, int) and ct > 0:
             pub_date = _dt.datetime.fromtimestamp(

{mpdt-0.2.2 → mpdt-0.2.4}/mpdt.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mpdt
-Version: 0.2.2
+Version: 0.2.4
 Summary: 公众号运营数据本地 CLI：把后台三源数据抓回本机 SQLite，一键生成周报
 Author: ttxs69
 License: MIT

{mpdt-0.2.2 → mpdt-0.2.4}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "mpdt"
-version = "0.2.2"
+version = "0.2.4"
 description = "公众号运营数据本地 CLI：把后台三源数据抓回本机 SQLite，一键生成周报"
 readme = "README.md"
 license = { text = "MIT" }

{mpdt-0.2.2 → mpdt-0.2.4}/tests/test_cli.py

@@ -285,6 +285,45 @@ class TestFetchStubbed(CliBase):
         con.close()
         self.assertEqual(n, 2)
 
+    def test_fetch_persists_create_time_and_send_hour(self):
+        """R101：publish_list 入库须带 create_time/publish_date/send_hour（元数据）。
+
+        根因=C20 external_rules.publish_window 结构性不可消费——解析层提取了
+        create_time 但 cli 入库只存 5 数值指标，发布时刻在 L1 就被丢弃。
+        send_hour 口径=群发时刻（publish_info.create_time，贴图回退 sent_info.time）
+        的北京时间 HH:MM。
+        """
+        art = self.pl_article(create_time=1791508200, publish_date='2026-10-09')  # 10-09 09:10+08（parse 层从 create_time 派生 publish_date，恒一致）
+        pl = _pl_stub([art])
+        dc = _dc_stub([])
+        rc, out, err = self.run_cli(['--json', 'fetch'], patch_fetch=(pl, dc))
+        self.assertEqual(rc, 0)
+        con = real_store.connect(config.DEFAULTS['db_path'])
+        row = con.execute(
+            "SELECT metrics FROM snapshots WHERE stat_date='2026-10-17'"
+            " AND source='publish_list' AND appmsgid='mA'").fetchone()
+        con.close()
+        self.assertIsNotNone(row, 'publish_list 行未入库')
+        m = json.loads(row[0])
+        self.assertEqual(m.get("create_time"), 1791508200)
+        self.assertEqual(m.get('publish_date'), '2026-10-09')
+        self.assertEqual(m.get('send_hour'), '09:10')
+
+    def test_fetch_persists_send_hour_none_when_no_create_time(self):
+        art = self.pl_article(create_time=None)  # 老上游缺字段：不虚构，存 null
+        pl = _pl_stub([art])
+        dc = _dc_stub([])
+        rc, out, err = self.run_cli(['--json', 'fetch'], patch_fetch=(pl, dc))
+        self.assertEqual(rc, 0)
+        con = real_store.connect(config.DEFAULTS['db_path'])
+        row = con.execute(
+            "SELECT metrics FROM snapshots WHERE stat_date='2026-10-17'"
+            " AND source='publish_list' AND appmsgid='mA'").fetchone()
+        con.close()
+        m = json.loads(row[0])
+        self.assertIsNone(m.get('send_hour'))
+        self.assertIsNone(m.get('create_time'))
+
     def test_fetch_from_to_dates(self):
         pl = _pl_stub([])
         dc = _dc_stub([])

{mpdt-0.2.2 → mpdt-0.2.4}/tests/test_fetchers.py

@@ -28,6 +28,20 @@ class TestPublishList(unittest.TestCase):
         for k in ('appmsgid', 'itemidx', 'title', 'read_num', 'share_num', 'like_num'):
             self.assertIn(k, a)
 
+    def test_pic_type_gets_publish_date_from_sent_info(self):
+        # R70：贴图类 publish_info 为空对象，时间戳在 sent_info.time（群发时间）。
+        # 实证：2247485215 sent_info.time=1792145145 → 2026-10-16 18:05:45+08，
+        # 与 PIC-rea 实查 t0（publish-read.py create_time）完全一致。
+        out = parse_publish_list(self.raw)
+        by_id = {a['appmsgid']: a for a in out['articles']}
+        a = by_id.get(2247485215)
+        self.assertIsNotNone(a, 'fixture 应含贴图样本 2247485215')
+        self.assertEqual(a['publish_date'], '2026-10-16')
+        self.assertEqual(a['create_time'], 1792145145)
+        b = by_id.get(2247485156)
+        self.assertIsNotNone(b, 'fixture 应含贴图样本 2247485156')
+        self.assertEqual(b['publish_date'], '2026-10-16')
+
     def test_structure_change_raises_upstream_4(self):
         with self.assertRaises(UpstreamError):
             parse_publish_list('{"base_resp":{"ret":0},"publish_page":"not json"}')