import math, time, os import numpy as np from PIL import Image from modules.ui import plaintext_to_html import modules.shared as shared from .helpers import ( fix_env_Path_ffprobe, closest_upper_divisible_by_eight, load_model_from_setting, do_upscaleImg,value_to_bool ) from .sd_helpers import renderImg2Img, renderTxt2Img from .image import shrink_and_paste_on_blank, open_image, apply_alpha_mask, draw_gradient_ellipse, resize_and_crop_image from .video import write_video def create_zoom( prompts_array, negative_prompt, num_outpainting_steps, guidance_scale, num_inference_steps, custom_init_image, custom_exit_image, video_frame_rate, video_zoom_mode, video_start_frame_dupe_amount, video_last_frame_dupe_amount, inpainting_denoising_strength, inpainting_mask_blur, inpainting_fill_mode, inpainting_full_res, inpainting_padding, zoom_speed, seed, outputsizeW, outputsizeH, batchcount, sampler, upscale_do, upscaler_name, upscale_by, progress=None, ): for i in range(batchcount): print(f"Batch {i+1}/{batchcount}") result = create_zoom_single( prompts_array, negative_prompt, num_outpainting_steps, guidance_scale, num_inference_steps, custom_init_image, custom_exit_image, video_frame_rate, video_zoom_mode, video_start_frame_dupe_amount, video_last_frame_dupe_amount, inpainting_denoising_strength, inpainting_mask_blur, inpainting_fill_mode, inpainting_full_res, inpainting_padding, zoom_speed, seed, outputsizeW, outputsizeH, sampler, upscale_do, upscaler_name, upscale_by, progress, ) return result def create_zoom_single( prompts_array, negative_prompt, num_outpainting_steps, guidance_scale, num_inference_steps, custom_init_image, custom_exit_image, video_frame_rate, video_zoom_mode, video_start_frame_dupe_amount, video_last_frame_dupe_amount, inpainting_denoising_strength, inpainting_mask_blur, inpainting_fill_mode, inpainting_full_res, inpainting_padding, zoom_speed, seed, outputsizeW, outputsizeH, sampler, upscale_do, upscaler_name, upscale_by, progress=None, ): # try: # if gr.Progress() is not None: # progress = gr.Progress() # progress(0, desc="Preparing Initial Image") # except Exception: # pass fix_env_Path_ffprobe() prompts = {} prompt_images = {} prompt_alpha_mask_images = {} prompt_image_is_keyframe = {} for x in prompts_array: try: key = int(x[0]) value = str(x[1]) file_loc = str(x[2]) alpha_mask_loc = str(x[3]) is_keyframe = bool(x[4]) prompts[key] = value prompt_images[key] = file_loc prompt_alpha_mask_images[key] = alpha_mask_loc prompt_image_is_keyframe[key] = value_to_bool(is_keyframe) except ValueError: pass assert len(prompts_array) > 0, "prompts is empty" print(str(len(prompts)) + " prompts found") print(str(len(prompt_images)) + " prompts Images found") width = closest_upper_divisible_by_eight(outputsizeW) height = closest_upper_divisible_by_eight(outputsizeH) current_image = Image.new(mode="RGBA", size=(width, height)) mask_image = np.array(current_image)[:, :, 3] mask_image = Image.fromarray(255 - mask_image).convert("RGB") current_image = current_image.convert("RGB") current_seed = seed extra_frames = 0 if custom_init_image: current_image = resize_and_crop_image(custom_init_image, width, height) print("using Custom Initial Image") else: if prompt_images[min(k for k in prompt_images.keys() if k >= 0)] == "": load_model_from_setting( "infzoom_txt2img_model", progress, "Loading Model for txt2img: " ) processed, current_seed = renderTxt2Img( prompts[min(k for k in prompts.keys() if k >= 0)], negative_prompt, sampler, num_inference_steps, guidance_scale, current_seed, width, height, ) current_image = processed.images[0] else: current_image = open_image(prompt_images[min(k for k in prompt_images.keys() if k >= 0)]) current_image = resize_and_crop_image(current_image, width, height) mask_width = math.trunc(width / 4) # was initially 512px => 128px mask_height = math.trunc(height / 4) # was initially 512px => 128px num_interpol_frames = round(video_frame_rate * zoom_speed) all_frames = [] if upscale_do and progress: progress(0, desc="upscaling inital image") all_frames.append( do_upscaleImg(current_image, upscale_do, upscaler_name, upscale_by) if upscale_do else current_image ) load_model_from_setting("infzoom_inpainting_model", progress, "Loading Model for inpainting/img2img: " ) if custom_exit_image: extra_frames += 2 for i in range(num_outpainting_steps + extra_frames): print_out = ( "Outpaint step: " + str(i + 1) + " / " + str(num_outpainting_steps + extra_frames) + " Seed: " + str(current_seed) ) print(print_out) if progress: progress(((i + 1) / num_outpainting_steps), desc=print_out) # apply available alpha mask of previous image if prompt_alpha_mask_images[max(k for k in prompt_alpha_mask_images.keys() if k <= (i + 1))] != "": current_image = apply_alpha_mask(current_image, open_image(prompt_alpha_mask_images[max(k for k in prompt_alpha_mask_images.keys() if k <= (i + 1))])) else: #generate automatic alpha mask current_image_gradient_ratio = 0.615 #max((min(current_image.width/current_image.height,current_image.height/current_image.width) * 0.89),0.1) current_image = apply_alpha_mask(current_image, draw_gradient_ellipse(current_image.width, current_image.height, current_image_gradient_ratio, 0.0, 3.0).convert("RGB")) prev_image_fix = current_image prev_image = shrink_and_paste_on_blank(current_image, mask_width, mask_height) current_image = prev_image # create mask (black image with white mask_width width edges) mask_image = np.array(current_image)[:, :, 3] mask_image = Image.fromarray(255 - mask_image).convert("RGB") # inpainting step current_image = current_image.convert("RGB") paste_previous_image = not prompt_image_is_keyframe[max(k for k in prompt_image_is_keyframe.keys() if k <= (i + 1))] # Custom and specified images work like keyframes if custom_exit_image and (i + 1) >= (num_outpainting_steps + extra_frames): current_image = resize_and_crop_image(custom_exit_image, width, height) print("using Custom Exit Image") else: if prompt_images[max(k for k in prompt_images.keys() if k <= (i + 1))] == "": processed, current_seed = renderImg2Img( prompts[max(k for k in prompts.keys() if k <= (i + 1))], negative_prompt, sampler, num_inference_steps, guidance_scale, current_seed, width, height, current_image, mask_image, inpainting_denoising_strength, inpainting_mask_blur, inpainting_fill_mode, inpainting_full_res, inpainting_padding, ) current_image = processed.images[0] # only paste previous image when generating a new image #current_image.paste(prev_image, mask=prev_image) paste_previous_image = True else: # use prerendered image, known as keyframe. Resize to target size current_image = open_image(prompt_images[max(k for k in prompt_images.keys() if k <= (i + 1))]) current_image = resize_and_crop_image(current_image, width, height) # apply available alpha mask of previous image #if prompt_alpha_mask_images[max(k for k in prompt_alpha_mask_images.keys() if k <= (i + 1))] != "": # current_image = apply_alpha_mask(current_image, open_image(prompt_alpha_mask_images[max(k for k in prompt_alpha_mask_images.keys() if k <= (i + 1))])) #else: # current_image_gradient_ratio = 0.65 #max((min(current_image.width/current_image.height,current_image.height/current_image.width) * 0.925),0.1) # current_image = draw_gradient_ellipse(current_image.width, current_image.height, current_image_gradient_ratio, 0.0, 1.8).convert("RGB") # paste previous image on current image if paste_previous_image: current_image.paste(prev_image, mask=prev_image) # interpolation steps between 2 inpainted images (=sequential zoom and crop) for j in range(num_interpol_frames - 1): interpol_image = current_image interpol_width = round( ( 1 - (1 - 2 * mask_width / width) ** (1 - (j + 1) / num_interpol_frames) ) * width / 2 ) interpol_height = round( ( 1 - (1 - 2 * mask_height / height) ** (1 - (j + 1) / num_interpol_frames) ) * height / 2 ) interpol_image = interpol_image.crop( ( interpol_width, interpol_height, width - interpol_width, height - interpol_height, ) ) interpol_image = interpol_image.resize((width, height)) # paste the higher resolution previous image in the middle to avoid drop in quality caused by zooming interpol_width2 = round( (1 - (width - 2 * mask_width) / (width - 2 * interpol_width)) / 2 * width ) interpol_height2 = round( (1 - (height - 2 * mask_height) / (height - 2 * interpol_height)) / 2 * height ) prev_image_fix_crop = shrink_and_paste_on_blank( prev_image_fix, interpol_width2, interpol_height2 ) interpol_image.paste(prev_image_fix_crop, mask=prev_image_fix_crop) if upscale_do and progress: progress(((i + 1) / num_outpainting_steps), desc="upscaling interpol") all_frames.append( do_upscaleImg(interpol_image, upscale_do, upscaler_name, upscale_by) if upscale_do else interpol_image ) if upscale_do and progress: progress(((i + 1) / num_outpainting_steps), desc="upscaling current") all_frames.append( do_upscaleImg(current_image, upscale_do, upscaler_name, upscale_by) if upscale_do else current_image ) video_file_name = "infinite_zoom_" + str(int(time.time())) + ".mp4" output_path = shared.opts.data.get( "infzoom_outpath", shared.opts.data.get("outdir_img2img_samples") ) save_path = os.path.join( output_path, shared.opts.data.get("infzoom_outSUBpath", "infinite-zooms") ) print("save to: " + save_path) if not os.path.exists(save_path): os.makedirs(save_path) out = os.path.join(save_path, video_file_name) write_video( out, all_frames, video_frame_rate, video_zoom_mode, int(video_start_frame_dupe_amount), int(video_last_frame_dupe_amount), ) return ( out, processed.images, processed.js(), plaintext_to_html(processed.info), plaintext_to_html(""), )