refactor: remove inference redundancy and fix cache leaks

- drop Executor unused tokenizer field, _head_dim, stale metrics docstring
- unify greedy sampling via SamplingPipeline.sample, drop top-level duplicate
- drop Task.flush_remaining no-op and unreachable prompt-length branch
- drop ProtocolHandler redundant chunks list (reuse body)
- fix page_size=1 token-slot leak on task_free
- clear _task_pages/_task_slots on alloc-failure paths
- reset _bind_state on task_free to avoid stale steady-state reuse
- remove unreachable contiguous branches in paged-only helpers

astrai/inference/core/metrics.py

@@ -12,7 +12,7 @@ class TaskTiming:
     """Timestamp snapshots and computed metrics for one generation task.
 
     Created by :class:`MetricsCollector` at task-registration time;
-    updated via ``prefill_scope`` / ``mark_finished``.
+    updated via ``record`` / ``mark_finished``.
     """
 
     task_id: str
@@ -27,8 +27,6 @@ class TaskTiming:
 
     # derived metrics
 
-    # derived metrics
-
     @property
     def queue_wait_ms(self) -> Optional[float]:
         if self.prefill_start_time is not None:
@@ -116,7 +114,7 @@ class MetricsCollector:
         metrics = MetricsCollector()
         metrics.register(task_id, arrival_time)
 
-        with metrics.prefill_scope(task_ids):
+        with metrics.record(task_ids, "prefill"):
             run_prefill(...)
 
         metrics.mark_finished(task_id, input_tokens, output_tokens)