Schedule Value Threshold

Documentation

Class name: SaltThresholdSchedule
Category: SALT/Scheduling
Output node: False

The SaltThresholdSchedule node is designed to apply a fading effect to a schedule list based on specified parameters, such as fade type and strength. This node enables dynamic adjustment of schedule values to create smooth transitions or emphasize certain parts of the schedule through fading in, fading out, or both.

Input types

Required

float_schedule
- The list of schedule values to be modified with a fading effect. This parameter is crucial for determining the base values that the fading effect will be applied to.
- Comfy dtype: LIST
- Python dtype: List[float]
max_frames
- Specifies the maximum number of frames for the schedule. This parameter limits the length of the output schedule list.
- Comfy dtype: INT
- Python dtype: int
output_mode
- Defines the format of the output schedule list, such as 'prompt_schedule', 'float_list', 'int_list', or 'raw'. This parameter determines how the modified schedule values are presented.
- Comfy dtype: COMBO[STRING]
- Python dtype: str
schedule_values
- A string detailing specific schedule values and their corresponding descriptions or settings. This parameter allows for the customization of schedule values based on textual descriptions.
- Comfy dtype: STRING
- Python dtype: str

Output types

value_schedule_list
- Comfy dtype: *
- The modified schedule list after applying the specified fading effect, showcasing the dynamic transitions between values.
- Python dtype: List[float]

Usage tips

Infra type: CPU
Common nodes: unknown

Source code

class SaltThresholdSchedule:
    @classmethod
    def INPUT_TYPES(cls):
        return {
            "required": {
                "float_schedule": ("LIST", ),
                "max_frames": ("INT", {"min": 1, "max": 4096, "default": 16}),
                "output_mode": (["prompt_schedule", "float_list", "int_list", "raw"],),
                "schedule_values": ("STRING", {"multiline": True, "default": '''"0.0": "A beautiful forest, (green:1.2) color scheme",
"0.5": "A beautiful forest, (autumn:1.2) color scheme",
"1.0": "A beautiful forest, (winter:1.2) color scheme"'''}),
            }
        }

    RETURN_TYPES = (WILDCARD,)
    RETURN_NAMES = ("value_schedule_list",)

    FUNCTION = "generate_sequence"
    CATEGORY = "SALT/Scheduling"

    def generate_sequence(self, float_schedule, schedule_values, max_frames, output_mode):
        try:
            text_dict = json.loads("{" + schedule_values + "}")
        except json.JSONDecodeError as e:
            raise ValueError("Unable to decode prompt schedule:", e)

        prompt_sequence = []
        prompt_dict = dict(sorted({float(key): value for key, value in text_dict.items()}.items(), key=lambda x: x[0]))
        default_prompt = next(iter(prompt_dict.values())) if prompt_dict else None

        adjusted_float_schedule = [float_schedule[i] if i < len(float_schedule) else 0 for i in range(max_frames)]

        for float_val in adjusted_float_schedule:
            closest_prompt = default_prompt
            min_diff = float('inf')
            
            for prompt_key, prompt_val in prompt_dict.items():
                diff = abs(float_val - prompt_key)
                
                if diff < min_diff:
                    closest_prompt = prompt_val
                    min_diff = diff

            prompt_sequence.append(closest_prompt)

        if output_mode == "prompt_schedule":
            output = ", ".join(f'"{i}": "{prompt}"' for i, prompt in enumerate(prompt_sequence))
        elif output_mode == "float_list":
            output = [float(value) for value in prompt_sequence]
        else:
            output = [int(value) for value in prompt_sequence]

        return (output, )

3.9 KiB Raw Permalink Blame History