Added Adjustment node
This commit is contained in:
Fillip
@@ -68,6 +68,7 @@ from .nodes.FL_PDFImageExtractor import FL_PDFImageExtractor
|
||||
from .nodes.FL_BulkPDFLoader import FL_BulkPDFLoader
|
||||
from .nodes.FL_SaveAndDisplayImage import FL_SaveAndDisplayImage
|
||||
from .nodes.FL_OllamaCaptioner import FL_OllamaCaptioner
|
||||
from .nodes.FL_ImageAdjuster import FL_ImageAdjuster
|
||||
|
||||
|
||||
|
||||
@@ -143,6 +144,7 @@ NODE_CLASS_MAPPINGS = {
|
||||
"FL_BulkPDFLoader": FL_BulkPDFLoader,
|
||||
"FL_SaveAndDisplayImage": FL_SaveAndDisplayImage,
|
||||
"FL_OllamaCaptioner": FL_OllamaCaptioner,
|
||||
"FL_ImageAdjuster": FL_ImageAdjuster,
|
||||
|
||||
}
|
||||
|
||||
@@ -217,6 +219,7 @@ NODE_DISPLAY_NAME_MAPPINGS = {
|
||||
"FL_BulkPDFLoader": "FL Bulk PDF Loader",
|
||||
"FL_SaveAndDisplayImage": "FL Save And Display Image",
|
||||
"FL_OllamaCaptioner": "FL Ollama Captioner by Cosmic",
|
||||
"FL_ImageAdjuster": "FL_ImageAdjuster",
|
||||
|
||||
}
|
||||
|
||||
|
||||
+9
-24
@@ -7,6 +7,7 @@ import sys
|
||||
from tqdm import tqdm
|
||||
import base64
|
||||
|
||||
# removed api key from input for safer use
|
||||
class FL_GPT_Vision:
|
||||
@classmethod
|
||||
def INPUT_TYPES(cls):
|
||||
@@ -17,13 +18,12 @@ class FL_GPT_Vision:
|
||||
"default": "You are a helpful assistant that describes images accurately and concisely.",
|
||||
"multiline": True}),
|
||||
"request_prompt": ("STRING", {"default": "Describe this image in detail.", "multiline": True}),
|
||||
"output_directory": ("STRING", {"default": ""}),
|
||||
"overwrite": ("BOOLEAN", {"default": False}),
|
||||
"max_tokens": ("INT", {"default": 300, "min": 1, "max": 4096}),
|
||||
"temperature": ("FLOAT", {"default": 0.7, "min": 0.0, "max": 2.0, "step": 0.1}),
|
||||
"detail": (["auto", "low", "high"],),
|
||||
"batch_size": ("INT", {"default": 5, "min": 1, "max": 20}),
|
||||
"upload_resolution": (["512", "768", "1024"],),
|
||||
"output_directory": ("STRING", {"default": ""}),
|
||||
},
|
||||
"optional": {
|
||||
"images": ("IMAGE",),
|
||||
@@ -35,21 +35,9 @@ class FL_GPT_Vision:
|
||||
RETURN_NAMES = ("message", "output_directory")
|
||||
FUNCTION = "generate_captions"
|
||||
CATEGORY = "🏵️Fill Nodes/GPT"
|
||||
OUTPUT_NODE = True
|
||||
|
||||
def resize_image(self, img, target_size):
|
||||
width, height = img.size
|
||||
aspect_ratio = width / height
|
||||
if width > height:
|
||||
new_width = target_size
|
||||
new_height = int(target_size / aspect_ratio)
|
||||
else:
|
||||
new_height = target_size
|
||||
new_width = int(target_size * aspect_ratio)
|
||||
return img.resize((new_width, new_height), Image.LANCZOS)
|
||||
|
||||
async def process_image(self, session, img, img_filename, output_directory, overwrite, api_key, model,
|
||||
system_prompt, request_prompt, max_tokens, temperature, detail, upload_resolution):
|
||||
system_prompt, request_prompt, max_tokens, temperature, detail):
|
||||
caption_filename = os.path.splitext(img_filename)[0] + ".txt"
|
||||
img_path = os.path.join(output_directory, img_filename)
|
||||
caption_path = os.path.join(output_directory, caption_filename)
|
||||
@@ -57,15 +45,12 @@ class FL_GPT_Vision:
|
||||
if not overwrite and os.path.exists(caption_path):
|
||||
return None
|
||||
|
||||
# Save the original image
|
||||
# Save the image
|
||||
img.save(img_path)
|
||||
|
||||
# Resize image for API upload
|
||||
resized_img = self.resize_image(img, int(upload_resolution))
|
||||
|
||||
# Encode resized image to base64
|
||||
# Encode image to base64
|
||||
buffered = io.BytesIO()
|
||||
resized_img.save(buffered, format="PNG")
|
||||
img.save(buffered, format="PNG")
|
||||
img_str = base64.b64encode(buffered.getvalue()).decode()
|
||||
|
||||
payload = {
|
||||
@@ -116,8 +101,8 @@ class FL_GPT_Vision:
|
||||
return await asyncio.gather(*tasks)
|
||||
|
||||
def generate_captions(self, model, system_prompt, request_prompt, output_directory, overwrite, max_tokens,
|
||||
temperature, detail, batch_size, upload_resolution, images=None, input_directory=None):
|
||||
api_key = os.getenv("OPENAI_API_KEY")
|
||||
temperature, detail, batch_size, images=None, input_directory=None):
|
||||
api_key = os.getenv("OPENAI_API_KEY") #looks for api key in env variable
|
||||
try:
|
||||
if not api_key:
|
||||
raise ValueError("API key is not set as an environment variable")
|
||||
@@ -155,7 +140,7 @@ class FL_GPT_Vision:
|
||||
for batch in tqdm(batches, desc="Processing batches", file=sys.stdout):
|
||||
batch_captions = await self.process_batch(batch, session, output_directory, overwrite, api_key,
|
||||
model, system_prompt, request_prompt, max_tokens,
|
||||
temperature, detail, upload_resolution)
|
||||
temperature, detail)
|
||||
all_captions.extend(batch_captions)
|
||||
return all_captions
|
||||
|
||||
|
||||
@@ -0,0 +1,83 @@
|
||||
import torch
|
||||
import numpy as np
|
||||
from PIL import Image, ImageEnhance, ImageFilter
|
||||
import base64
|
||||
import io
|
||||
from server import PromptServer
|
||||
from .utils import tensor_to_pil, pil_to_tensor
|
||||
|
||||
|
||||
class FL_ImageAdjuster:
|
||||
@classmethod
|
||||
def INPUT_TYPES(s):
|
||||
return {
|
||||
"required": {
|
||||
"image": ("IMAGE",),
|
||||
"hue": ("FLOAT", {"default": 0, "min": -180, "max": 180, "step": 1}),
|
||||
"saturation": ("FLOAT", {"default": 0, "min": -100, "max": 100, "step": 1}),
|
||||
"brightness": ("FLOAT", {"default": 0, "min": -100, "max": 100, "step": 1}),
|
||||
"contrast": ("FLOAT", {"default": 0, "min": -100, "max": 100, "step": 1}),
|
||||
"sharpness": ("FLOAT", {"default": 0, "min": 0, "max": 100, "step": 1}),
|
||||
},
|
||||
}
|
||||
|
||||
RETURN_TYPES = ("IMAGE",)
|
||||
FUNCTION = "adjust_image"
|
||||
CATEGORY = "🏵️Fill Nodes/Image"
|
||||
OUTPUT_NODE = True
|
||||
|
||||
def adjust_image(self, image, hue, saturation, brightness, contrast, sharpness):
|
||||
# Convert tensor to PIL Image
|
||||
pil_image = tensor_to_pil(image)
|
||||
|
||||
# Apply adjustments
|
||||
adjusted_image = self.apply_adjustments(pil_image, hue, saturation, brightness, contrast, sharpness)
|
||||
|
||||
# Convert back to tensor
|
||||
tensor_image = pil_to_tensor(adjusted_image)
|
||||
|
||||
# Prepare image for display
|
||||
display_image = self.prepare_image_for_display(adjusted_image)
|
||||
|
||||
# Send the image to the frontend
|
||||
PromptServer.instance.send_sync("fl_image_adjuster", {"image": display_image})
|
||||
|
||||
return (tensor_image,)
|
||||
|
||||
def apply_adjustments(self, image, hue, saturation, brightness, contrast, sharpness):
|
||||
# Convert to HSV for hue and saturation adjustments
|
||||
hsv_image = image.convert('HSV')
|
||||
h, s, v = hsv_image.split()
|
||||
|
||||
# Hue adjustment
|
||||
h = h.point(lambda x: (x + hue) % 256)
|
||||
|
||||
# Saturation adjustment
|
||||
s = s.point(lambda x: max(0, min(255, x + saturation * 255 / 100)))
|
||||
|
||||
# Merge channels
|
||||
hsv_image = Image.merge('HSV', (h, s, v))
|
||||
rgb_image = hsv_image.convert('RGB')
|
||||
|
||||
# Brightness adjustment
|
||||
enhancer = ImageEnhance.Brightness(rgb_image)
|
||||
rgb_image = enhancer.enhance(1 + brightness / 100)
|
||||
|
||||
# Contrast adjustment
|
||||
enhancer = ImageEnhance.Contrast(rgb_image)
|
||||
rgb_image = enhancer.enhance(1 + contrast / 100)
|
||||
|
||||
# Sharpness adjustment
|
||||
if sharpness > 0:
|
||||
# Convert sharpness to an integer percentage between 100 and 200
|
||||
sharpness_percent = int(100 + sharpness)
|
||||
rgb_image = rgb_image.filter(ImageFilter.UnsharpMask(radius=2, percent=sharpness_percent, threshold=3))
|
||||
|
||||
return rgb_image
|
||||
|
||||
def prepare_image_for_display(self, pil_image):
|
||||
# Convert PIL Image to base64 string
|
||||
buffered = io.BytesIO()
|
||||
pil_image.save(buffered, format="PNG")
|
||||
img_str = base64.b64encode(buffered.getvalue()).decode()
|
||||
return f"data:image/png;base64,{img_str}"
|
||||
@@ -16,5 +16,3 @@ PyMuPDF
|
||||
reportlab
|
||||
PyPDF2
|
||||
ollama
|
||||
opencv-python
|
||||
kornia
|
||||
@@ -0,0 +1,131 @@
|
||||
import { app } from "../../../scripts/app.js";
|
||||
import { api } from "../../../scripts/api.js";
|
||||
|
||||
function hideWidgetForGood(node, widget) {
|
||||
if (!widget) return;
|
||||
widget.origType = widget.type;
|
||||
widget.origComputeSize = widget.computeSize;
|
||||
widget.origSerializeValue = widget.serializeValue;
|
||||
widget.computeSize = () => [0, -4];
|
||||
widget.type = "converted-widget";
|
||||
}
|
||||
|
||||
app.registerExtension({
|
||||
name: "Comfy.FL_ImageAdjuster",
|
||||
async nodeCreated(node) {
|
||||
if (node.comfyClass === "FL_ImageAdjuster") {
|
||||
addImageAdjusterUI(node);
|
||||
}
|
||||
}
|
||||
});
|
||||
|
||||
function addImageAdjusterUI(node) {
|
||||
const MIN_WIDTH = 200;
|
||||
const MIN_HEIGHT = 300;
|
||||
const SLIDER_HEIGHT = 25;
|
||||
const PADDING = 0;
|
||||
|
||||
// Default values for sliders
|
||||
const DEFAULT_VALUES = {
|
||||
hue: 0,
|
||||
saturation: 0,
|
||||
brightness: 0,
|
||||
contrast: 0,
|
||||
sharpness: 0
|
||||
};
|
||||
|
||||
// Find and hide the original widgets
|
||||
const originalWidgets = {};
|
||||
["hue", "saturation", "brightness", "contrast", "sharpness"].forEach(name => {
|
||||
const widget = node.widgets.find(w => w.name === name);
|
||||
if (widget) {
|
||||
originalWidgets[name] = widget;
|
||||
hideWidgetForGood(node, widget);
|
||||
}
|
||||
});
|
||||
|
||||
// Create custom sliders
|
||||
const sliders = [
|
||||
createSlider("Hue", -180, 180, originalWidgets.hue),
|
||||
createSlider("Saturation", -100, 100, originalWidgets.saturation),
|
||||
createSlider("Brightness", -100, 100, originalWidgets.brightness),
|
||||
createSlider("Contrast", -100, 100, originalWidgets.contrast),
|
||||
createSlider("Sharpness", 0, 100, originalWidgets.sharpness)
|
||||
];
|
||||
|
||||
function createSlider(name, min, max, originalWidget) {
|
||||
const slider = node.addWidget("slider", name, originalWidget.value, (v) => {
|
||||
originalWidget.value = v;
|
||||
node.setDirtyCanvas(true);
|
||||
}, { min: min, max: max, step: 1 });
|
||||
return slider;
|
||||
}
|
||||
|
||||
// Add reset button
|
||||
const resetButton = node.addWidget("button", "Reset", null, () => {
|
||||
resetSliders();
|
||||
});
|
||||
|
||||
function resetSliders() {
|
||||
sliders.forEach(slider => {
|
||||
const defaultValue = DEFAULT_VALUES[slider.name.toLowerCase()];
|
||||
slider.value = defaultValue;
|
||||
slider.callback(defaultValue);
|
||||
});
|
||||
|
||||
node.setDirtyCanvas(true);
|
||||
node.triggerSlot(0);
|
||||
}
|
||||
|
||||
// Add image preview
|
||||
const img = new Image();
|
||||
img.onload = () => node.setDirtyCanvas(true);
|
||||
|
||||
node.onDrawBackground = function(ctx) {
|
||||
if (!this.flags.collapsed) {
|
||||
const [w, h] = this.size;
|
||||
|
||||
// Calculate the Y position of the last widget
|
||||
const lastWidget = node.widgets[node.widgets.length - 1];
|
||||
const lastWidgetY = lastWidget.last_y || 0;
|
||||
|
||||
// Set the image Y offset to be just below the last widget
|
||||
const IMAGE_Y_OFFSET = lastWidgetY + SLIDER_HEIGHT + PADDING;
|
||||
|
||||
const imageArea = h - IMAGE_Y_OFFSET - PADDING;
|
||||
|
||||
// Draw image
|
||||
if (img.src) {
|
||||
const aspectRatio = img.width / img.height;
|
||||
let drawWidth = w - 2 * PADDING;
|
||||
let drawHeight = imageArea;
|
||||
|
||||
if (drawWidth / drawHeight > aspectRatio) {
|
||||
drawWidth = drawHeight * aspectRatio;
|
||||
} else {
|
||||
drawHeight = drawWidth / aspectRatio;
|
||||
}
|
||||
|
||||
const x = PADDING + (w - 2 * PADDING - drawWidth) / 2;
|
||||
const y = IMAGE_Y_OFFSET;
|
||||
|
||||
ctx.drawImage(img, x, y, drawWidth, drawHeight);
|
||||
}
|
||||
}
|
||||
};
|
||||
|
||||
// Listen for the image from the backend
|
||||
api.addEventListener("fl_image_adjuster", (event) => {
|
||||
if (event.detail.image) {
|
||||
img.src = event.detail.image;
|
||||
}
|
||||
});
|
||||
|
||||
function updateNodeSize() {
|
||||
node.size[0] = Math.max(MIN_WIDTH, node.size[0]);
|
||||
node.size[1] = Math.max(MIN_HEIGHT, node.size[1]);
|
||||
}
|
||||
|
||||
node.onResize = updateNodeSize;
|
||||
updateNodeSize();
|
||||
}
|
||||
Reference in New Issue
Block a user