Add pydocstyle to the ruff rules (#52136)

Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Signed-off-by: Harry Mellor <hej.mellor@gmail.com>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
This commit is contained in:
Harry Mellor
2026-09-16 16:44:29 +00:00
committed by GitHub
co-authored by Claude Opus 5
parent 804e5377aa
commit 8c1557a79c
2002 changed files with 15024 additions and 14269 deletions
-2
View File
@@ -28,7 +28,6 @@ def compute_ple_shard_overlap(
tp_end: int,
) -> PLEShardOverlap | None:
"""Compute the overlap of a checkpoint shard and one TP vocabulary range."""
if checkpoint_start < 0 or checkpoint_rows < 0:
raise ValueError("checkpoint shard bounds must be non-negative")
if tp_start < 0 or tp_end < tp_start:
@@ -54,7 +53,6 @@ def copy_ple_embedding_shard_(
tp_end: int,
) -> int:
"""Copy the overlapping rows of a PLE checkpoint shard into a TP table."""
if destination.ndim == 0 or loaded_weight.ndim != destination.ndim:
raise ValueError("destination and loaded weight must have matching ranks")
if destination.shape[1:] != loaded_weight.shape[1:]:
@@ -47,7 +47,6 @@ from vllm.v1.kv_cache_interface import (
def canonical_qsa_rope_positions(positions: torch.Tensor) -> torch.Tensor:
"""Return exact per-token positions as ``[tokens, 1, 3]`` int64 rows."""
if positions.ndim == 1:
positions = positions.unsqueeze(0).expand(3, -1)
elif positions.ndim != 2 or positions.shape[0] not in (1, 3):
@@ -115,7 +114,6 @@ def circular_qsa_slot_mapping(
out: torch.Tensor | None = None,
) -> torch.Tensor:
"""Map each request to its fixed physical block as a circular token ring."""
if compressor_state_size <= 0:
raise ValueError("QSA circular buffer size must be positive")
if block_table.ndim != 2:
@@ -167,7 +165,6 @@ def compressed_qsa_slot_mapping(
out: torch.Tensor | None = None,
) -> torch.Tensor:
"""Build boundary-only slots for an ``MLAAttentionSpec`` QSA cache."""
if storage_block_size <= 0 or compress_ratio <= 0:
raise ValueError("QSA block size and compression ratio must be positive")
compressed_positions = torch.div(