Update to latest ComfyUI

comfy/app/frontend_management.py

@@ -20,9 +20,15 @@ from ..cmd.folder_paths import add_model_folder_path  # pylint: disable=import-e
 
 REQUEST_TIMEOUT = 10  # seconds
 
+
 def check_frontend_version():
     return None
 
+
+def frontend_install_warning_message() -> str:
+    return ""
+
+
 class Asset(TypedDict):
     url: str
 
@@ -141,8 +147,6 @@ class FrontendManager:
 
 comfyui-workflow-templates is not installed.
 
-{frontend_install_warning_message()}
-
 ********** ERROR ***********
 """.strip()
             )

comfy/cli_args.py

@@ -131,7 +131,7 @@ def _create_parser() -> EnhancedConfigArgParser:
     parser.add_argument("--deterministic", action="store_true",
                         help="Make pytorch use slower deterministic algorithms when it can. Note that this might not make images deterministic in all cases.")
 
-    parser.add_argument("--fast", nargs="*", type=PerformanceFeature, help="Enable some untested and potentially quality deteriorating optimizations. Pass a list specific optimizations if you only want to enable specific ones. Current valid optimizations: fp16_accumulation fp8_matrix_mult cublas_ops")
+    parser.add_argument("--fast", nargs="*", type=PerformanceFeature, help="Enable some untested and potentially quality deteriorating optimizations. Pass a list specific optimizations if you only want to enable specific ones. Current valid optimizations: fp16_accumulation fp8_matrix_mult cublas_ops", default=set())
 
     parser.add_argument("--dont-print-server", action="store_true", help="Don't print server output.")
     parser.add_argument("--quick-test-for-ci", action="store_true", help="Quick test for CI. Raises an error if nodes cannot be imported,")

comfy/client/embedded_comfy_client.py

@@ -71,7 +71,8 @@ async def __execute_prompt(
             args.update(configuration)
 
         with tracer.start_as_current_span("Initialize Prompt Executor", context=span_context):
-            prompt_executor = PromptExecutor(progress_handler, lru_size=configuration.cache_lru if configuration is not None else 0)
+            # todo: deal with new caching features
+            prompt_executor = PromptExecutor(progress_handler)
             prompt_executor.raise_exceptions = True
             _prompt_executor.executor = prompt_executor
 

comfy/clip_vision.py

@@ -1,5 +1,6 @@
 import json
 import logging
+from typing import Optional
 
 import torch
 
@@ -116,7 +117,8 @@ def convert_to_transformers(sd, prefix):
     return sd
 
 
-def load_clipvision_from_sd(sd, prefix="", convert_keys=False):
+def load_clipvision_from_sd(sd, prefix="", convert_keys=False) -> Optional[ClipVisionModel]:
+    json_config: dict = {}
     if convert_keys:
         sd = convert_to_transformers(sd, prefix)
     if "vision_model.encoder.layers.47.layer_norm1.weight" in sd:

comfy/cmd/execution.py

@@ -745,7 +745,7 @@ def validate_inputs(prompt, item, validated: typing.Dict[str, ValidateInputsTupl
             r = get_nodes().NODE_CLASS_MAPPINGS[o_class_type].RETURN_TYPES
             received_type = r[val[1]]
             received_types[x] = received_type
-            any_enum = received_type == [] and (isinstance(type_input, list) or isinstance(type_input, tuple))
+            any_enum = received_type == [] and (isinstance(input_type, list) or isinstance(input_type, tuple))
 
             if 'input_types' not in validate_function_inputs and not validate_node_input(received_type, input_type) and not any_enum:
                 details = f"{x}, {received_type} != {input_type}"
@@ -857,8 +857,8 @@ def validate_inputs(prompt, item, validated: typing.Dict[str, ValidateInputsTupl
                     if "\\" in val:
                         # try to normalize paths for comparison purposes
                         val = canonicalize_path(val)
-                    if all(isinstance(item, (str, PathLike)) for item in type_input):
+                    if all(isinstance(item, (str, PathLike)) for item in combo_options):
-                        type_input = [canonicalize_path(item) for item in type_input]
+                        combo_options = [canonicalize_path(item) for item in combo_options]
                     if val not in combo_options:
                         input_config = info
                         list_info = ""

comfy/model_downloader.py

@@ -36,6 +36,10 @@ _session = Session()
 _hf_fs = HfFileSystem()
 
 
+def get_filename_list(folder_name: str) -> list[str]:
+    return get_filename_list_with_downloadable(folder_name)
+
+
 def get_filename_list_with_downloadable(folder_name: str, known_files: Optional[List[Downloadable] | KnownDownloadables] = None) -> List[str]:
     if known_files is None:
         known_files = _get_known_models_for_folder_name(folder_name)
@@ -45,6 +49,13 @@ def get_filename_list_with_downloadable(folder_name: str, known_files: Optional[
     return list(map(canonicalize_path, sorted(list(existing | downloadable))))
 
 
+def get_full_path_or_raise(folder_name: str, filename: str) -> str:
+    res = get_or_download(folder_name, filename)
+    if res is None:
+        raise FileNotFoundError(f"{folder_name} does not contain {filename}")
+    return res
+
+
 def get_or_download(folder_name: str, filename: str, known_files: Optional[List[Downloadable] | KnownDownloadables] = None) -> Optional[str]:
     if known_files is None:
         known_files = _get_known_models_for_folder_name(folder_name)

comfy/node_helpers.py

@@ -102,6 +102,7 @@ def string_to_torch_dtype(string):
 
 
 def image_alpha_fix(destination, source):
+    import torch
     if destination.shape[-1] < source.shape[-1]:
         source = source[..., :destination.shape[-1]]
     elif destination.shape[-1] > source.shape[-1]:

comfy/nodes/base_nodes.py

@@ -1714,8 +1714,9 @@ class LoadImage:
                     mask = np.array(i.getchannel('A')).astype(np.float32) / 255.0
                     mask = 1. - torch.from_numpy(mask)
                 elif i.mode == 'P' and 'transparency' in i.info:
-                mask = np.array(i.convert('RGBA').getchannel('A')).astype(np.float32) / 255.0
+                    mask = np.array(i.convert('RGBA').getchannel('A')).astype(np.float32) / 255.0
-                mask = 1. - torch.from_numpy(mask)else:
+                    mask = 1. - torch.from_numpy(mask)
+                else:
                     mask = torch.zeros((64,64), dtype=torch.float32, device="cpu")
                 output_images.append(image)
                 output_masks.append(mask.unsqueeze(0))

comfy/sampler_helpers.py

@@ -1,6 +1,5 @@
 import uuid
 
-from . import conds
 from . import model_management
 from . import patcher_extension
 from . import utils
@@ -113,9 +112,9 @@ def cleanup_additional_models(models):
 
 
 def prepare_sampling(model: ModelPatcher, noise_shape, conds, model_options=None):
-    executor = comfy.patcher_extension.WrapperExecutor.new_executor(
+    executor = patcher_extension.WrapperExecutor.new_executor(
         _prepare_sampling,
-        comfy.patcher_extension.get_all_wrappers(comfy.patcher_extension.WrappersMP.PREPARE_SAMPLING, model_options, is_model_options=True)
+        patcher_extension.get_all_wrappers(patcher_extension.WrappersMP.PREPARE_SAMPLING, model_options, is_model_options=True)
     )
     return executor.execute(model, noise_shape, conds, model_options=model_options)
 

comfy/text_encoders/aura_t5.py

@@ -16,6 +16,7 @@ class PT5XlModel(sd1_clip.SDClipModel):
 class PT5XlTokenizer(sd1_clip.SDTokenizer):
     def __init__(self, embedding_directory=None, **kwargs):
         tokenizer_path = resources.files("comfy.text_encoders.t5_pile_tokenizer") / "tokenizer.model"
+        tokenizer_data = kwargs.pop("tokenizer_data", {})
         super().__init__(tokenizer_path, pad_with_end=False, embedding_size=2048, embedding_key='pile_t5xl', tokenizer_class=SPieceTokenizer, has_start_token=False, pad_to_max_length=False, max_length=99999999, min_length=256, pad_token=1, tokenizer_data=tokenizer_data)
 
 

comfy/text_encoders/hydit.py

@@ -21,6 +21,7 @@ class HyditBertModel(sd1_clip.SDClipModel):
 class HyditBertTokenizer(sd1_clip.SDTokenizer):
     def __init__(self, **kwargs):
         tokenizer_path = get_package_as_path(f"{__package__}.hydit_clip_tokenizer")
+        tokenizer_data = kwargs.pop("tokenizer_data", {})
         super().__init__(tokenizer_path, pad_with_end=False, embedding_size=1024, embedding_key='chinese_roberta', tokenizer_class=BertTokenizer, pad_to_max_length=False, max_length=512, min_length=77, tokenizer_data=tokenizer_data)
 
 

comfy_extras/nodes/nodes_hidream.py

@@ -1,16 +1,18 @@
-import folder_paths
-import comfy.sd
 import comfy.model_management
+import comfy.sd
+from comfy.cmd import folder_paths
+from comfy import model_downloader
 
 
 class QuadrupleCLIPLoader:
     @classmethod
     def INPUT_TYPES(s):
-        return {"required": { "clip_name1": (folder_paths.get_filename_list("text_encoders"), ),
+        return {"required": {"clip_name1": (model_downloader.get_filename_list("text_encoders"),),
-                              "clip_name2": (folder_paths.get_filename_list("text_encoders"), ),
+                             "clip_name2": (model_downloader.get_filename_list("text_encoders"),),
-                              "clip_name3": (folder_paths.get_filename_list("text_encoders"), ),
+                             "clip_name3": (model_downloader.get_filename_list("text_encoders"),),
-                              "clip_name4": (folder_paths.get_filename_list("text_encoders"), )
+                             "clip_name4": (model_downloader.get_filename_list("text_encoders"),)
                              }}
+
     RETURN_TYPES = ("CLIP",)
     FUNCTION = "load_clip"
 
@@ -19,35 +21,37 @@ class QuadrupleCLIPLoader:
     DESCRIPTION = "[Recipes]\n\nhidream: long clip-l, long clip-g, t5xxl, llama_8b_3.1_instruct"
 
     def load_clip(self, clip_name1, clip_name2, clip_name3, clip_name4):
-        clip_path1 = folder_paths.get_full_path_or_raise("text_encoders", clip_name1)
+        clip_path1 = model_downloader.get_full_path_or_raise("text_encoders", clip_name1)
-        clip_path2 = folder_paths.get_full_path_or_raise("text_encoders", clip_name2)
+        clip_path2 = model_downloader.get_full_path_or_raise("text_encoders", clip_name2)
-        clip_path3 = folder_paths.get_full_path_or_raise("text_encoders", clip_name3)
+        clip_path3 = model_downloader.get_full_path_or_raise("text_encoders", clip_name3)
-        clip_path4 = folder_paths.get_full_path_or_raise("text_encoders", clip_name4)
+        clip_path4 = model_downloader.get_full_path_or_raise("text_encoders", clip_name4)
         clip = comfy.sd.load_clip(ckpt_paths=[clip_path1, clip_path2, clip_path3, clip_path4], embedding_directory=folder_paths.get_folder_paths("embeddings"))
         return (clip,)
 
+
 class CLIPTextEncodeHiDream:
     @classmethod
     def INPUT_TYPES(s):
         return {"required": {
-            "clip": ("CLIP", ),
+            "clip": ("CLIP",),
             "clip_l": ("STRING", {"multiline": True, "dynamicPrompts": True}),
             "clip_g": ("STRING", {"multiline": True, "dynamicPrompts": True}),
             "t5xxl": ("STRING", {"multiline": True, "dynamicPrompts": True}),
             "llama": ("STRING", {"multiline": True, "dynamicPrompts": True})
-            }}
+        }}
+
     RETURN_TYPES = ("CONDITIONING",)
     FUNCTION = "encode"
 
     CATEGORY = "advanced/conditioning"
 
     def encode(self, clip, clip_l, clip_g, t5xxl, llama):
-
         tokens = clip.tokenize(clip_g)
         tokens["l"] = clip.tokenize(clip_l)["l"]
         tokens["t5xxl"] = clip.tokenize(t5xxl)["t5xxl"]
         tokens["llama"] = clip.tokenize(llama)["llama"]
-        return (clip.encode_from_tokens_scheduled(tokens), )
+        return (clip.encode_from_tokens_scheduled(tokens),)
+
 
 NODE_CLASS_MAPPINGS = {
     "QuadrupleCLIPLoader": QuadrupleCLIPLoader,

comfy_extras/nodes/nodes_hunyuan3d.py

@@ -447,6 +447,8 @@ class VoxelToMesh:
             mesh_function = voxel_to_mesh
         elif algorithm == "surface net":
             mesh_function = voxel_to_mesh_surfnet
+        else:
+            mesh_function = None
 
         for x in voxel.data:
             v, f = mesh_function(x, threshold=threshold, device=None)