New nodes

__init__.py

@@ -101,6 +101,9 @@ NODE_CONFIG = {
     #curve nodes
     "SplineEditor": {"class": SplineEditor, "name": "Spline Editor"},
     "CreateShapeMaskOnPath": {"class": CreateShapeMaskOnPath, "name": "Create Shape Mask On Path"},
+    "CreateTextOnPath": {"class": CreateTextOnPath, "name": "Create Text On Path"},
+    "CreateGradientFromCoords": {"class": CreateGradientFromCoords, "name": "Create Gradient From Coords"},
+    "GradientToFloat": {"class": GradientToFloat, "name": "Gradient To Float"},
     "WeightScheduleExtend": {"class": WeightScheduleExtend, "name": "Weight Schedule Extend"},
     "MaskOrImageToWeight": {"class": MaskOrImageToWeight, "name": "Mask Or Image To Weight"},
     "WeightScheduleConvert": {"class": WeightScheduleConvert, "name": "Weight Schedule Convert"},

nodes/curve_nodes.py

@@ -1,7 +1,7 @@
 import torch
 from torchvision import transforms
 import json
-from PIL import Image, ImageDraw, ImageFont
+from PIL import Image, ImageDraw, ImageFont, ImageColor
 import numpy as np
 from ..utility.utility import pil2tensor
 import folder_paths
@@ -388,6 +388,116 @@ Locations are center locations.
         out_images = torch.cat(image_list, dim=0).cpu().float()
         out_masks = torch.cat(mask_list, dim=0)
         return (out_images, out_masks, 1.0 - out_masks,)
+
+class CreateGradientFromCoords:
+    
+    RETURN_TYPES = ("IMAGE", )
+    RETURN_NAMES = ("image", )
+    FUNCTION = "generate"
+    CATEGORY = "KJNodes/image"
+    DESCRIPTION = """
+Creates a gradient image from coordinates.    
+"""
+
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                "coordinates": ("STRING", {"forceInput": True}),
+                "frame_width": ("INT", {"default": 512,"min": 16, "max": 4096, "step": 1}),
+                "frame_height": ("INT", {"default": 512,"min": 16, "max": 4096, "step": 1}),
+                "start_color": ("STRING", {"default": 'white'}),
+                "end_color": ("STRING", {"default": 'black'}),
+                "multiplier": ("FLOAT", {"default": 1.0, "min": 0.01, "max": 100.0, "step": 0.01}),
+        },
+    } 
+    
+    def generate(self, coordinates, frame_width, frame_height, start_color, end_color, multiplier):
+        # Parse the coordinates
+        coordinates = json.loads(coordinates.replace("'", '"'))
+
+        # Create an image
+        image = Image.new("RGB", (frame_width, frame_height))
+        draw = ImageDraw.Draw(image)
+
+        # Extract start and end points for the gradient
+        start_coord = coordinates[0]
+        end_coord = coordinates[1]
+
+        start_color = ImageColor.getrgb(start_color)
+        end_color = ImageColor.getrgb(end_color)
+
+        # Calculate the gradient direction (vector)
+        gradient_direction = (end_coord['x'] - start_coord['x'], end_coord['y'] - start_coord['y'])
+        gradient_length = (gradient_direction[0] ** 2 + gradient_direction[1] ** 2) ** 0.5
+
+        # Iterate over each pixel in the image
+        for y in range(frame_height):
+            for x in range(frame_width):
+                # Calculate the projection of the point on the gradient line
+                point_vector = (x - start_coord['x'], y - start_coord['y'])
+                projection = (point_vector[0] * gradient_direction[0] + point_vector[1] * gradient_direction[1]) / gradient_length
+                projection = max(min(projection, gradient_length), 0)  # Clamp the projection value
+
+                # Calculate the blend factor for the current pixel
+                blend = projection * multiplier / gradient_length 
+
+                # Determine the color of the current pixel
+                color = (
+                    int(start_color[0] + (end_color[0] - start_color[0]) * blend),
+                    int(start_color[1] + (end_color[1] - start_color[1]) * blend),
+                    int(start_color[2] + (end_color[2] - start_color[2]) * blend)
+                )
+
+                # Set the pixel color
+                draw.point((x, y), fill=color)
+
+        # Convert the PIL image to a tensor (assuming such a function exists in your context)
+        image_tensor = pil2tensor(image)
+
+        return (image_tensor,)
+
+class GradientToFloat:
+    
+    RETURN_TYPES = ("FLOAT", "FLOAT",)
+    RETURN_NAMES = ("float_x", "float_y", )
+    FUNCTION = "sample"
+    CATEGORY = "KJNodes/image"
+    DESCRIPTION = """
+Calculates list of floats from image.    
+"""
+
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                "image": ("IMAGE", ),
+                "steps": ("INT", {"default": 10, "min": 2, "max": 10000, "step": 1}),
+        },
+    } 
+    
+    def sample(self, image, steps):
+        # Assuming image is a tensor with shape [B, H, W, C]
+        B, H, W, C = image.shape
+
+        # Sample along the width axis (W)
+        w_intervals = torch.linspace(0, W - 1, steps=steps, dtype=torch.int64)
+        # Assuming we're sampling from the first batch and the first channel
+        w_sampled = image[0, :, w_intervals, 0]
+
+        # Sample along the height axis (H)
+        h_intervals = torch.linspace(0, H - 1, steps=steps, dtype=torch.int64)
+        # Assuming we're sampling from the first batch and the first channel
+        h_sampled = image[0, h_intervals, :, 0]
+
+        # Taking the mean across the height for width sampling, and across the width for height sampling
+        w_values = w_sampled.mean(dim=0).tolist()
+        h_values = h_sampled.mean(dim=1).tolist()
+
+        print("Sampled width axis values:", w_values)
+        print("Sampled height axis values:", h_values)
+
+        return (w_values, h_values)
     
 class MaskOrImageToWeight: