app.py

import os
import json
import copy
import time
import random
import logging
import numpy as np
from typing import Any, Dict, List, Optional, Union
import gradio as gr
import logging
import torch
from PIL import Image
import spaces
from diffusers import DiffusionPipeline, AutoencoderTiny, AutoencoderKL, AutoPipelineForImage2Image
from diffusers.utils import load_image
from huggingface_hub import hf_hub_download, HfFileSystem, ModelCard, snapshot_download
import requests
import pandas as pd

from transformers.utils import move_cache
move_cache()

from diffusers import (
    DiffusionPipeline,
    AutoencoderTiny,
    AutoencoderKL,
    AutoPipelineForImage2Image,
    FluxPipeline,
    FlowMatchEulerDiscreteScheduler)

from huggingface_hub import (
    hf_hub_download,
    HfFileSystem,
    ModelCard,
    snapshot_download)

from diffusers.utils import load_image

from huggingface_hub import HfApi
token = os.getenv("HF_TOKEN")


#---if workspace = local or colab---

# Authenticate with Hugging Face
# from huggingface_hub import login

# Log in to Hugging Face using the provided token
# hf_token = 'hf-token-authentication'
# login(hf_token)

def calculate_shift(
    image_seq_len,
    base_seq_len: int = 256,
    max_seq_len: int = 4096,
    base_shift: float = 0.5,
    max_shift: float = 1.16,
):
    m = (max_shift - base_shift) / (max_seq_len - base_seq_len)
    b = base_shift - m * base_seq_len
    mu = image_seq_len * m + b
    return mu

def retrieve_timesteps(
    scheduler,
    num_inference_steps: Optional[int] = None,
    device: Optional[Union[str, torch.device]] = None,
    timesteps: Optional[List[int]] = None,
    sigmas: Optional[List[float]] = None,
    **kwargs,
):
    if timesteps is not None and sigmas is not None:
        raise ValueError("Apenas um entre `timesteps` ou `sigmas` pode ser passado. Por favor, escolha um para definir valores personalizados")
        scheduler.set_timesteps(timesteps=timesteps, device=device, **kwargs)
        timesteps = scheduler.timesteps
        num_inference_steps = len(timesteps)
    elif sigmas is not None:
        scheduler.set_timesteps(sigmas=sigmas, device=device, **kwargs)
        timesteps = scheduler.timesteps
        num_inference_steps = len(timesteps)
    else:
        scheduler.set_timesteps(num_inference_steps, device=device, **kwargs)
        timesteps = scheduler.timesteps
    return timesteps, num_inference_steps

# FLUX pipeline
@torch.inference_mode()
def flux_pipe_call_that_returns_an_iterable_of_images(
    self,
    prompt: Union[str, List[str]] = None,
    prompt_2: Optional[Union[str, List[str]]] = None,
    height: Optional[int] = None,
    width: Optional[int] = None,
    num_inference_steps: int = 28,
    timesteps: List[int] = None,
    guidance_scale: float = 3.5,
    num_images_per_prompt: Optional[int] = 1,
    generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
    latents: Optional[torch.FloatTensor] = None,
    prompt_embeds: Optional[torch.FloatTensor] = None,
    pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
    output_type: Optional[str] = "pil",
    return_dict: bool = True,
    joint_attention_kwargs: Optional[Dict[str, Any]] = None,
    max_sequence_length: int = 512,
    good_vae: Optional[Any] = None,
):
    height = height or self.default_sample_size * self.vae_scale_factor
    width = width or self.default_sample_size * self.vae_scale_factor
    
    self.check_inputs(
        prompt,
        prompt_2,
        height,
        width,
        prompt_embeds=prompt_embeds,
        pooled_prompt_embeds=pooled_prompt_embeds,
        max_sequence_length=max_sequence_length,
    )

    self._guidance_scale = guidance_scale
    self._joint_attention_kwargs = joint_attention_kwargs
    self._interrupt = False

    batch_size = 1 if isinstance(prompt, str) else len(prompt)
    device = self._execution_device

    lora_scale = joint_attention_kwargs.get("scale", None) if joint_attention_kwargs is not None else None
    prompt_embeds, pooled_prompt_embeds, text_ids = self.encode_prompt(
        prompt=prompt,
        prompt_2=prompt_2,
        prompt_embeds=prompt_embeds,
        pooled_prompt_embeds=pooled_prompt_embeds,
        device=device,
        num_images_per_prompt=num_images_per_prompt,
        max_sequence_length=max_sequence_length,
        lora_scale=lora_scale,
    )
    
    num_channels_latents = self.transformer.config.in_channels // 4
    latents, latent_image_ids = self.prepare_latents(
        batch_size * num_images_per_prompt,
        num_channels_latents,
        height,
        width,
        prompt_embeds.dtype,
        device,
        generator,
        latents,
    )
    
    sigmas = np.linspace(1.0, 1 / num_inference_steps, num_inference_steps)
    image_seq_len = latents.shape[1]
    mu = calculate_shift(
        image_seq_len,
        self.scheduler.config.base_image_seq_len,
        self.scheduler.config.max_image_seq_len,
        self.scheduler.config.base_shift,
        self.scheduler.config.max_shift,
    )
    timesteps, num_inference_steps = retrieve_timesteps(
        self.scheduler,
        num_inference_steps,
        device,
        timesteps,
        sigmas,
        mu=mu,
    )
    self._num_timesteps = len(timesteps)

    guidance = torch.full([1], guidance_scale, device=device, dtype=torch.float32).expand(latents.shape[0]) if self.transformer.config.guidance_embeds else None

    for i, t in enumerate(timesteps):
        if self.interrupt:
            continue

        timestep = t.expand(latents.shape[0]).to(latents.dtype)

        noise_pred = self.transformer(
            hidden_states=latents,
            timestep=timestep / 1000,
            guidance=guidance,
            pooled_projections=pooled_prompt_embeds,
            encoder_hidden_states=prompt_embeds,
            txt_ids=text_ids,
            img_ids=latent_image_ids,
            joint_attention_kwargs=self.joint_attention_kwargs,
            return_dict=False,
        )[0]

        latents_for_image = self._unpack_latents(latents, height, width, self.vae_scale_factor)
        latents_for_image = (latents_for_image / self.vae.config.scaling_factor) + self.vae.config.shift_factor
        image = self.vae.decode(latents_for_image, return_dict=False)[0]
        yield self.image_processor.postprocess(image, output_type=output_type)[0]
        latents = self.scheduler.step(noise_pred, t, latents, return_dict=False)[0]
        torch.cuda.empty_cache()
        
    latents = self._unpack_latents(latents, height, width, self.vae_scale_factor)
    latents = (latents / good_vae.config.scaling_factor) + good_vae.config.shift_factor
    image = good_vae.decode(latents, return_dict=False)[0]
    self.maybe_free_model_hooks()
    torch.cuda.empty_cache()
    yield self.image_processor.postprocess(image, output_type=output_type)[0]

#------------------------------------------------------------------------------------------------------------------------------------------------------------#
loras = [
    #Super-Realism
    {
        "image": "https://huggingface.co/Collos/Jalves/resolve/main/images/jose.webp",
        "title": "Jose Alves",
        "repo": "Collos/Jalves",
        "weights": "Jalves.safetensors",
        "trigger_word": "José Alves"            
    },
    {
        "image": "https://huggingface.co/Collos/JulioCesar/resolve/main/images/WhatsApp%20Image%202024-12-10%20at%2009.33.50.jpeg",
        "title": "Júlio César",
        "repo": "Collos/JulioCesar",
        "weights": "julio.safetensorss",
        "trigger_word": "Júlio"            
    },
     {
        "image": "https://huggingface.co/Collos/PedroJr/resolve/main/images/WhatsApp%20Image%202024-12-10%20at%2009.34.01.jpeg",
        "title": "Pedro Jr.",
        "repo": "Collos/PedroJr",
        "weights": "pedrojr.safetensors",
        "trigger_word": "Pedro"            
    },
    {
        "image": "https://huggingface.co/Collos/JoseClovis/resolve/main/images/WhatsApp%20Image%202024-12-10%20at%2009.38.50.jpeg",
        "title": "José Clóvis",
        "repo": "Collos/JoseClovis",
        "weights": "clovis.safetensors",
        "trigger_word": "Clóvis"            
    }
    
    #add new
]

# Initialize the base model
use_auth_token=True
dtype = torch.bfloat16
device = "cuda" if torch.cuda.is_available() else "cpu"
base_model = "black-forest-labs/FLUX.1-dev"


taef1 = AutoencoderTiny.from_pretrained("madebyollin/taef1", torch_dtype=dtype).to(device)
good_vae = AutoencoderKL.from_pretrained(base_model, subfolder="vae", torch_dtype=dtype).to(device)
pipe = DiffusionPipeline.from_pretrained(base_model, torch_dtype=dtype, vae=taef1).to(device)
pipe_i2i = AutoPipelineForImage2Image.from_pretrained(
    base_model,
    vae=good_vae,
    transformer=pipe.transformer,
    text_encoder=pipe.text_encoder,
    tokenizer=pipe.tokenizer,
    text_encoder_2=pipe.text_encoder_2,
    tokenizer_2=pipe.tokenizer_2,
    torch_dtype=dtype
)


#TAEF1 is very tiny autoencoder which uses the same "latent API" as FLUX.1's VAE. FLUX.1 is useful for real-time previewing of the FLUX.1 generation process.#
taef1 = AutoencoderTiny.from_pretrained("madebyollin/taef1", torch_dtype=dtype).to(device)
good_vae = AutoencoderKL.from_pretrained(base_model, subfolder="vae", torch_dtype=dtype).to(device)
pipe = DiffusionPipeline.from_pretrained(base_model, torch_dtype=dtype, vae=taef1).to(device)
pipe_i2i = AutoPipelineForImage2Image.from_pretrained(base_model,
                                                      vae=good_vae,
                                                      transformer=pipe.transformer,
                                                      text_encoder=pipe.text_encoder,
                                                      tokenizer=pipe.tokenizer,
                                                      text_encoder_2=pipe.text_encoder_2,
                                                      tokenizer_2=pipe.tokenizer_2,
                                                      torch_dtype=dtype
                                                     )

MAX_SEED = 2**32-1

pipe.flux_pipe_call_that_returns_an_iterable_of_images = flux_pipe_call_that_returns_an_iterable_of_images.__get__(pipe)

class calculateDuration:
    def __init__(self, activity_name=""):
        self.activity_name = activity_name

    def __enter__(self):
        self.start_time = time.time()
        return self
    
    def __exit__(self, exc_type, exc_value, traceback):
        self.end_time = time.time()
        self.elapsed_time = self.end_time - self.start_time
        if self.activity_name:
            print(f"tempo passado para {self.activity_name}: {self.elapsed_time:.6f} segundos")
        else:
            print(f"tempo passado: {self.elapsed_time:.6f} segundos")

def update_selection(evt: gr.SelectData, width, height):
    selected_lora = loras[evt.index]
    new_placeholder = f"Digite o prompt para {selected_lora['title']}"
    lora_repo = selected_lora["repo"]
    updated_text = f"### Selecionado: [{lora_repo}](https://huggingface.co/{lora_repo}) ✅"
    if "aspect" in selected_lora:
        if selected_lora["aspect"] == "retrato":
            width = 768
            height = 1024
        elif selected_lora["aspect"] == "paisagem":
            width = 1024
            height = 768
        else:
            width = 1024
            height = 1024
    return (
        gr.update(placeholder=new_placeholder),
        updated_text,
        evt.index,
        width,
        height,
    )

@spaces.GPU(duration=100)
def generate_image(prompt_mash, steps, seed, cfg_scale, width, height, lora_scale, progress):
    pipe.to("cuda")
    generator = torch.Generator(device="cuda").manual_seed(seed)
    with calculateDuration("Generating image"):
        # Generate image
        for img in pipe.flux_pipe_call_that_returns_an_iterable_of_images(
            prompt=prompt_mash,
            num_inference_steps=steps,
            guidance_scale=cfg_scale,
            width=width,
            height=height,
            generator=generator,
            joint_attention_kwargs={"scale": lora_scale},
            output_type="pil",
            good_vae=good_vae,
        ):
            yield img

def generate_image_to_image(prompt_mash, image_input_path, image_strength, steps, cfg_scale, width, height, lora_scale, seed):
    generator = torch.Generator(device="cuda").manual_seed(seed)
    pipe_i2i.to("cuda")
    image_input = load_image(image_input_path)
    final_image = pipe_i2i(
        prompt=prompt_mash,
        image=image_input,
        strength=image_strength,
        num_inference_steps=steps,
        guidance_scale=cfg_scale,
        width=width,
        height=height,
        generator=generator,
        joint_attention_kwargs={"scale": lora_scale},
        output_type="pil",
    ).images[0]
    return final_image 

@spaces.GPU(duration=100)
def run_lora(prompt, image_input, image_strength, cfg_scale, steps, selected_index, randomize_seed, seed, width, height, lora_scale, progress=gr.Progress(track_tqdm=True)):
    if selected_index is None:
        raise gr.Error("Selecione um modelo para continuar.")
    selected_lora = loras[selected_index]
    lora_path = selected_lora["repo"]
    trigger_word = selected_lora["trigger_word"]
    if(trigger_word):
        if "trigger_position" in selected_lora:
            if selected_lora["trigger_position"] == "prepend":
                prompt_mash = f"{trigger_word} {prompt}"
            else:
                prompt_mash = f"{prompt} {trigger_word}"
        else:
            prompt_mash = f"{trigger_word} {prompt}"
    else:
        prompt_mash = prompt

    with calculateDuration("Carregando Modelo"):
        pipe.unload_lora_weights()
        pipe_i2i.unload_lora_weights()
        
    #LoRA weights flow
    with calculateDuration(f"Carregando modelo para {selected_lora['title']}"):
        pipe_to_use = pipe_i2i if image_input is not None else pipe
        weight_name = selected_lora.get("Pesos", None)
        
        pipe_to_use.load_lora_weights(
            lora_path, 
            weight_name=weight_name, 
            low_cpu_mem_usage=True
        )
            
    with calculateDuration("Gerando fontes"):
        if randomize_seed:
            seed = random.randint(0, MAX_SEED)
            
    if(image_input is not None):
        
        final_image = generate_image_to_image(prompt_mash, image_input, image_strength, steps, cfg_scale, width, height, lora_scale, seed)
        yield final_image, seed, gr.update(visible=False)
    else:
        image_generator = generate_image(prompt_mash, steps, seed, cfg_scale, width, height, lora_scale, progress)
    
        final_image = None
        step_counter = 0
        for image in image_generator:
            step_counter+=1
            final_image = image
            progress_bar = f'<div class="progress-container"><div class="progress-bar" style="--current: {step_counter}; --total: {steps};"></div></div>'
            yield image, seed, gr.update(value=progress_bar, visible=True)
            
        yield final_image, seed, gr.update(value=progress_bar, visible=False)
        
def get_huggingface_safetensors(link):
  split_link = link.split("/")
  if(len(split_link) == 2):
            model_card = ModelCard.load(link)
            base_model = model_card.data.get("base_model")
            print(base_model)
      
            #Allows Both
            if((base_model != "black-forest-labs/FLUX.1-dev") and (base_model != "black-forest-labs/FLUX.1-schnell")):
                raise Exception("Flux LoRA Not Found!")
                
            # Only allow "black-forest-labs/FLUX.1-dev"
            #if base_model != "black-forest-labs/FLUX.1-dev":
                #raise Exception("Only FLUX.1-dev is supported, other LoRA models are not allowed!")
                
            image_path = model_card.data.get("widget", [{}])[0].get("output", {}).get("url", None)
            trigger_word = model_card.data.get("instance_prompt", "")
            image_url = f"https://huggingface.co/{link}/resolve/main/{image_path}" if image_path else None
            fs = HfFileSystem()
            try:
                list_of_files = fs.ls(link, detail=False)
                for file in list_of_files:
                    if(file.endswith(".safetensors")):
                        safetensors_name = file.split("/")[-1]
                    if (not image_url and file.lower().endswith((".jpg", ".jpeg", ".png", ".webp"))):
                      image_elements = file.split("/")
                      image_url = f"https://huggingface.co/{link}/resolve/main/{image_elements[-1]}"
            except Exception as e:
              print(e)
              gr.Warning(f"You didn't include a link neither a valid Hugging Face repository with a *.safetensors LoRA")
              raise Exception(f"You didn't include a link neither a valid Hugging Face repository with a *.safetensors LoRA")
            return split_link[1], link, safetensors_name, trigger_word, image_url

def check_custom_model(link):
    if(link.startswith("https://")):
        if(link.startswith("https://huggingface.co") or link.startswith("https://www.huggingface.co")):
            link_split = link.split("huggingface.co/")
            return get_huggingface_safetensors(link_split[1])
    else: 
        return get_huggingface_safetensors(link)

def add_custom_lora(custom_lora):
    global loras
    if(custom_lora):
        try:
            title, repo, path, trigger_word, image = check_custom_model(custom_lora)
            print(f"Modelo Externo: {repo}")
            card = f'''
            <div class="custom_lora_card">
              <span>Loaded custom LoRA:</span>
              <div class="card_internal">
                <img src="{image}" />
                <div>
                    <h3>{title}</h3>
                    <small>{"Usando: <code><b>"+trigger_word+"</code></b> como palavra-chave" if trigger_word else "Não encontramos a palavra-chave, se tiver, coloque-a no prompt."}<br></small>
                </div>
              </div>
            </div>
            '''
            existing_item_index = next((index for (index, item) in enumerate(loras) if item['repo'] == repo), None)
            if(not existing_item_index):
                new_item = {
                    "image": image,
                    "title": title,
                    "repo": repo,
                    "weights": path,
                    "trigger_word": trigger_word
                }
                print(new_item)
                existing_item_index = len(loras)
                loras.append(new_item)
        
            return gr.update(visible=True, value=card), gr.update(visible=True), gr.Gallery(selected_index=None), f"Custom: {path}", existing_item_index, trigger_word
        except Exception as e:
            gr.Warning(f"Modelo Inválido: ou o link está errado ou não é um FLUX")
            return gr.update(visible=True, value=f"Modelo Inválido: ou o link está errado ou não é um FLUX"), gr.update(visible=False), gr.update(), "", None, ""
    else:
        return gr.update(visible=False), gr.update(visible=False), gr.update(), "", None, ""

def remove_custom_lora():
    return gr.update(visible=False), gr.update(visible=False), gr.update(), "", None, ""

run_lora.zerogpu = True

#import gradio as gr

collos = gr.themes.Soft(
    primary_hue="gray",
    secondary_hue="stone",
    neutral_hue="slate",
    radius_size=gr.themes.Size(lg="15px", md="8px", sm="6px", xl="16px", xs="4px", xxl="24px", xxs="2px")
).set(
    body_background_fill='*primary_100',
    embed_radius='*radius_lg',
    shadow_drop='0 1px 2px rgba(0, 0, 0, 0.1)',
    shadow_drop_lg='0 1px 2px rgba(0, 0, 0, 0.1)',
    shadow_inset='0 1px 2px rgba(0, 0, 0, 0.1)',
    shadow_spread='0 1px 2px rgba(0, 0, 0, 0.1)',
    shadow_spread_dark='0 1px 2px rgba(0, 0, 0, 0.1)',
    block_radius='*radius_lg',
    block_shadow='*shadow_drop',
    container_radius='*radius_lg'
)

with gr.Blocks(theme=collos, delete_cache=(60, 60)) as app:
    title = gr.HTML(
        """<img src="https://collos.com.br/wp-content/uploads/2024/12/collos_p.png" alt="Logo" style="display: block; margin: 0 auto; padding: 5px 0px 20px 0px; width: 200px;" />""",
        elem_id="title",
    )
    selected_index = gr.State(None)
    with gr.Row():
        with gr.Column(scale=3):
            prompt = gr.Textbox(label="Prompt", lines=1, placeholder=":/ Selecione o modelo ")
        with gr.Column(scale=1):
            generate_button = gr.Button("Gerar Imagem", variant="primary", elem_id="cta")
    with gr.Row():
        with gr.Column():
            selected_info = gr.Markdown("")
            gallery = gr.Gallery(
                [(item["image"], item["title"]) for item in loras],
                allow_preview=False,
                columns=3,
                show_share_button=False
            )
            with gr.Group():
                custom_lora = gr.Textbox(label="Selecione um Modelo Externo", placeholder="black-forest-labs/FLUX.1-dev")
                gr.Markdown("[Cheque a lista de modelos do Huggingface](https://huggingface.co/models?other=base_model:adapter:black-forest-labs/FLUX.1-dev)", elem_id="lora_list")
            custom_lora_info = gr.HTML(visible=False)
            custom_lora_button = gr.Button("Remova modelo Externo", visible=False)
        with gr.Column():
            progress_bar = gr.Markdown(elem_id="progress", visible=False)
            result = gr.Image(label="Imagem Gerada")

    with gr.Row():
        with gr.Accordion("Configurações Avançadas", open=False):
            with gr.Row():
                input_image = gr.Image(label="Insira uma Imagem", type="filepath")
                image_strength = gr.Slider(label="Remossão de ruído", info="Valores mais baixos significam maior influência da imagem.", minimum=0.1, maximum=1.0, step=0.01, value=0.75)
            with gr.Column():
                with gr.Row():
                    cfg_scale = gr.Slider(label="Aumentar Escala", minimum=1, maximum=20, step=0.5, value=3.5)
                    steps = gr.Slider(label="Passos", minimum=1, maximum=50, step=1, value=28)
                
                with gr.Row():
                    width = gr.Slider(label="Largura", minimum=256, maximum=1536, step=64, value=1024)
                    height = gr.Slider(label="Altura", minimum=256, maximum=1536, step=64, value=1024)
                
                with gr.Row():
                    randomize_seed = gr.Checkbox(True, label="Fonte Randomizada")
                    seed = gr.Slider(label="Fontes", minimum=0, maximum=MAX_SEED, step=1, value=0, randomize=True)
                    lora_scale = gr.Slider(label="Escala do Modelo", minimum=0, maximum=3, step=0.01, value=0.95)

    gallery.select(
        update_selection,
        inputs=[width, height],
        outputs=[prompt, selected_info, selected_index, width, height]
    )
    custom_lora.input(
        add_custom_lora,
        inputs=[custom_lora],
        outputs=[custom_lora_info, custom_lora_button, gallery, selected_info, selected_index, prompt]
    )
    custom_lora_button.click(
        remove_custom_lora,
        outputs=[custom_lora_info, custom_lora_button, gallery, selected_info, selected_index, custom_lora]
    )
    gr.on(
        triggers=[generate_button.click, prompt.submit],
        fn=run_lora,
        inputs=[prompt, input_image, image_strength, cfg_scale, steps, selected_index, randomize_seed, seed, width, height, lora_scale],
        outputs=[result, seed, progress_bar]
    )

app.queue()
app.launch()