feat: add clickhouse micro-batch storage

src/core/worker/batch.py

@@ -0,0 +1,73 @@
+import time
+from dataclasses import dataclass
+
+from core.domain.models import NormalizedCDR
+
+
+@dataclass(frozen=True, slots=True)
+class KafkaPosition:
+    topic: str
+    partition: int
+    offset: int
+
+
+@dataclass(frozen=True, slots=True)
+class BufferedCDR:
+    cdr: NormalizedCDR
+    position: KafkaPosition
+
+
+class MicroBatch:
+    def __init__(
+        self,
+        max_size: int,
+        flush_interval_seconds: float,
+    ) -> None:
+        self._max_size = max_size
+        self._flush_interval_seconds = flush_interval_seconds
+
+        self._items: list[BufferedCDR] = []
+
+        self._started_at: float | None = None
+
+    def add(
+        self,
+        item: BufferedCDR,
+    ) -> None:
+        if not self._items:
+            self._started_at = time.monotonic()
+
+        self._items.append(item)
+
+    @property
+    def size(self) -> int:
+        return len(self._items)
+
+    @property
+    def is_empty(self) -> bool:
+        return not self._items
+
+    @property
+    def is_full(self) -> bool:
+        return len(self._items) >= self._max_size
+
+    def is_due(self) -> bool:
+        if not self._items:
+            return False
+
+        if self._started_at is None:
+            return False
+
+        elapsed = (
+            time.monotonic()
+            - self._started_at
+        )
+
+        return elapsed >= self._flush_interval_seconds
+
+    def snapshot(self) -> list[BufferedCDR]:
+        return list(self._items)
+
+    def clear(self) -> None:
+        self._items.clear()
+        self._started_at = None