feat(state): persist last_update_id and rolling processed_ids

src/journal_bot/state.py

@@ -0,0 +1,46 @@
+import json
+from collections import deque
+from pathlib import Path
+
+
+class State:
+    """Persistent state: last_update_id + rolling processed_ids."""
+
+    def __init__(self, state_dir: Path, processed_ids_max: int = 500):
+        self.state_dir = state_dir
+        self.state_dir.mkdir(parents=True, exist_ok=True)
+        self._last_path = state_dir / "last_update_id.json"
+        self._processed_path = state_dir / "processed_ids.json"
+        self._max = processed_ids_max
+        self._processed: deque[int] = deque(maxlen=processed_ids_max)
+        self._load()
+
+    def _load(self) -> None:
+        if self._last_path.exists():
+            self._last_update_id = json.loads(self._last_path.read_text())["value"]
+        else:
+            self._last_update_id = 0
+        if self._processed_path.exists():
+            ids = json.loads(self._processed_path.read_text())["ids"]
+            self._processed = deque(ids[-self._max:], maxlen=self._max)
+
+    @property
+    def last_update_id(self) -> int:
+        return self._last_update_id
+
+    def set_last_update_id(self, value: int) -> None:
+        self._last_update_id = value
+        tmp = self._last_path.with_suffix(".tmp")
+        tmp.write_text(json.dumps({"value": value}))
+        tmp.replace(self._last_path)
+
+    def has_processed(self, update_id: int) -> bool:
+        return update_id in self._processed
+
+    def mark_processed(self, update_id: int) -> None:
+        if update_id in self._processed:
+            return
+        self._processed.append(update_id)
+        tmp = self._processed_path.with_suffix(".tmp")
+        tmp.write_text(json.dumps({"ids": list(self._processed)}))
+        tmp.replace(self._processed_path)

tests/test_state.py

@@ -0,0 +1,34 @@
+from pathlib import Path
+from journal_bot.state import State
+
+
+def test_state_initial_values(tmp_path):
+    s = State(tmp_path)
+    assert s.last_update_id == 0
+    assert not s.has_processed(123)
+
+
+def test_state_persists_last_update_id(tmp_path):
+    s = State(tmp_path)
+    s.set_last_update_id(456)
+    s2 = State(tmp_path)
+    assert s2.last_update_id == 456
+
+
+def test_state_processed_ids_rolling_window(tmp_path):
+    s = State(tmp_path, processed_ids_max=3)
+    s.mark_processed(1)
+    s.mark_processed(2)
+    s.mark_processed(3)
+    s.mark_processed(4)
+    assert not s.has_processed(1)  # dropped
+    assert s.has_processed(2)
+    assert s.has_processed(3)
+    assert s.has_processed(4)
+
+
+def test_state_processed_ids_persist(tmp_path):
+    s = State(tmp_path)
+    s.mark_processed(42)
+    s2 = State(tmp_path)
+    assert s2.has_processed(42)