[CI] Ignore ruff D209, rejoin the docstrings it split, and silence incompatible-rule warnings (#57212)

Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
This commit is contained in:
Harry Mellor
2026-09-16 10:57:24 -07:00
committed by GitHub
co-authored by Claude Opus 5
parent 1085b64425
commit 651a88c09a
693 changed files with 2088 additions and 4171 deletions
@@ -55,20 +55,17 @@ class BaseFuser(ABC):
self, module: nn.Module, prefix: str, vllm_config: "VllmConfig"
) -> nn.Module:
"""Apply the fusion to an already-validated `module`, returning the
module to install in its place (mutated in place, or freshly built).
"""
module to install in its place (mutated in place, or freshly built)."""
def orig_to_new_stacked(self, prefix: str) -> dict[str, tuple[str, ShardId]]:
"""`WeightsMapper.orig_to_new_stacked` entries this fuser contributes
(none unless it stacks weights).
"""
(none unless it stacks weights)."""
return {}
@property
def packed_modules_mapping(self) -> dict[str, list[str]]:
"""`packed_modules_mapping` entries this fuser contributes (none unless
it stacks weights).
"""
it stacks weights)."""
return {}
@@ -122,8 +119,7 @@ class RewriteFuser(BaseFuser):
) -> nn.Module:
"""Fuse an already-validated `module` in place (see `Fusers.__getitem__`).
Builds the merged submodule and binds the compiled forward.
"""
Builds the merged submodule and binds the compiled forward."""
self.update_attrs(module, prefix, vllm_config)
module.forward = types.MethodType(self.fused_forward, module)
return module
@@ -132,8 +128,7 @@ class RewriteFuser(BaseFuser):
@dataclass
class StackedFuser(RewriteFuser):
"""A fuser that merges sibling projections into one stacked linear and
rewrites the forward to call it.
"""
rewrites the forward to call it."""
merged_name: ClassVar[str]
"""Attribute name of the merged module created by `update_attrs`."""
@@ -152,16 +147,14 @@ class StackedFuser(RewriteFuser):
def shards(self) -> list[tuple[str, ShardId]]:
"""Each projection's original name and its shard id in the merged module.
Source for both `orig_to_new_stacked` and `packed_modules_mapping`.
"""
Source for both `orig_to_new_stacked` and `packed_modules_mapping`."""
def orig_to_new_stacked(self, prefix: str) -> dict[str, tuple[str, ShardId]]:
"""`WeightsMapper.orig_to_new_stacked` entries for one fused instance.
Maps each checkpoint name to `(merged_name, shard_id)`, keyed by qualname
so only this exact layer is remapped, never a same-named projection
elsewhere (e.g. an unfused MoE expert's `gate_proj`).
"""
elsewhere (e.g. an unfused MoE expert's `gate_proj`)."""
merged = maybe_prefix(prefix, self.merged_name)
return {
maybe_prefix(prefix, name): (merged, shard) for name, shard in self.shards
@@ -170,8 +163,7 @@ class StackedFuser(RewriteFuser):
@property
def packed_modules_mapping(self) -> dict[str, list[str]]:
"""`{merged_name: [projection names]}` so quantization can unpack the
fused layer into its per-shard configs.
"""
fused layer into its per-shard configs."""
return {self.merged_name: [name for name, _ in self.shards]}
def _unguarded_calls(
@@ -182,8 +174,7 @@ class StackedFuser(RewriteFuser):
`update_attrs` deletes the projections it merges, so any reference to one beyond
its call site must be a guard on its existence; folding those to their constant
value keeps the rewritten forward off a name that no longer exists. A reference
that is not such a guard raises, so fusion is skipped.
"""
that is not such a guard raises, so fusion is skipped."""
calls = []
for name in names:
call, refs = self_call_and_refs(funcdef, name)
@@ -232,8 +232,7 @@ class MLAFuser(StackedFuser):
def update_forward(self, module: nn.Module) -> None:
"""Merge `q_a_proj` and `kv_a_proj` into one fused down proj then split.
Bypass the KV expansion method so the compressed latent reaches the `vllm_mla`
attention interface unexpanded.
"""
attention interface unexpanded."""
funcdef, fn = recover_forward(type(module))
if self.has_q_lora:
# q_a_proj is usually inside the `else` of `if self.q_lora_rank is None`.
@@ -80,8 +80,7 @@ def _forced_dtype(nodes: Iterable[fx.Node]) -> torch.dtype | None:
Computations that must run in higher precision say so in their forward (e.g.
`hidden_states.type(torch.float32)`), so the dtype is readable from the graph
even when the config does not name it.
"""
even when the config does not name it."""
for node in nodes:
if node.op not in ("call_method", "call_function"):
continue
@@ -126,8 +125,7 @@ class SharedExpertMLP(nn.Module):
def _moe_block_forward(self: nn.Module, hidden_states: torch.Tensor) -> torch.Tensor:
"""Standard MoE block forward.
Routing and any shared experts are handled inside `self.experts: MoERunner`.
"""
Routing and any shared experts are handled inside `self.experts: MoERunner`."""
orig_shape = hidden_states.shape
hidden_states = hidden_states.reshape(-1, orig_shape[-1])
num_tokens = hidden_states.shape[0]
@@ -154,8 +152,7 @@ class MoEBlockFuser:
def _match_router(gate: nn.Module) -> tuple[str, torch.dtype | None] | None:
"""Matches `topk(score(linear(x)))`, `score` being `softmax`/`sigmoid`.
Returns the scoring function and the dtype the router computes in.
"""
Returns the scoring function and the dtype the router computes in."""
state = {name for name, _ in named_state(gate)}
if "weight" not in state or state - {"weight", "e_score_correction_bias"}:
return None
@@ -283,8 +280,7 @@ class MoEBlockFuser:
self, moe_block: nn.Module, prefix: str
) -> SharedExpertMLP | None:
"""Build the HF shared expert (and its optional gate)
as a `SharedExpertMLP` for vLLM's fused MoE.
"""
as a `SharedExpertMLP` for vLLM's fused MoE."""
if self.shared_name is None:
return None
shared_experts = getattr(moe_block, self.shared_name)