refactor: rebuild KV cache with three-layer separation architecture

- Replace CacheView/ContiguousCache/PageCache with SGLang-inspired design: KVStorage (flat token-level NHD buffers [n_layers, size, H, D]), ReqToTokenPool (index table [req_idx, pos] -> token_slot), Allocator + PrefixCache (slot allocation with LRU and prefix sharing)
- Add KVCache as pure dataclass passed to model: k_buffer, v_buffer, req_to_token, req_pool_indices, seq_lens, out_cache_loc
- PagePool orchestrates all three layers, supports contiguous mode (pre-allocated per-request blocks, default) and paged mode (page_size=1 or >1 with dynamic allocation and prefix caching)
- Attention layers now do raw buffer indexing instead of opaque write/gather method calls on CacheView objects
- Update executor.bind_tasks signature: seq_lens list + start_pos
- Rename paged_cache -> kv_cache throughout model/ and inference/

astrai/inference/__init__.py

@@ -30,21 +30,16 @@ from astrai.inference.api.openai import OpenAIResponseBuilder
 from astrai.inference.core import (
     STOP,
     Allocator,
-    CacheView,
-    ContiguousCache,
-    ContiguousCacheView,
     Executor,
     InferenceScheduler,
     KVCache,
-    PageCache,
-    PageCacheView,
+    KVStorage,
     PagePool,
     PrefixCache,
-    Storage,
+    ReqToTokenPool,
     Task,
     TaskManager,
     TaskStatus,
-    TaskTable,
     page_hash,
 )
 from astrai.inference.engine import GenerationRequest, InferenceEngine
@@ -68,16 +63,11 @@ __all__ = [
     "TaskManager",
     "TaskStatus",
     "Allocator",
-    "CacheView",
     "KVCache",
-    "ContiguousCache",
-    "ContiguousCacheView",
-    "PageCache",
-    "PageCacheView",
+    "KVStorage",
     "PagePool",
     "PrefixCache",
-    "Storage",
-    "TaskTable",
+    "ReqToTokenPool",
     "page_hash",
     "sample",
     "BaseSamplingStrategy",