PromptGallery-stable-diffus.../paste_this_to_webui_scripts.../prompt_gallery.py

612 lines
22 KiB
Python

import copy
import os
import random
import sys
import traceback
import shlex
import yaml
import platform
import subprocess as sp
import shutil
import json
import tempfile
import gradio as gr
import csv
import typing
import base64
import io
from PIL import Image
import mimetypes
mimetypes.init()
mimetypes.add_type('application/javascript', '.js')
import modules.generation_parameters_copypaste as parameters_copypaste
from modules.generation_parameters_copypaste import image_from_url_text
import modules.scripts as scripts
from modules.processing import Processed, process_images
from modules.shared import opts, cmd_opts, state
import modules.shared as shared
if '__file__' in locals().keys():
root_path = os.path.dirname(os.path.dirname(os.path.realpath(__file__)))
root_path = os.path.abspath(os.path.join(root_path, os.path.pardir))
else:
root_path = os.path.abspath(shared.script_path)
try:
with open("./extensions/prompt_gallery_name.json") as fd:
extension_name = json.load(fd)['name']
except:
extension_name = "Prompt Gallery"
OUTPATH_SAMPLES = os.path.join(root_path, 'extensions', extension_name, 'assets', 'preview')
OUTPATH_GRIDS = os.path.join(root_path, 'extensions', extension_name, 'assets', 'grid')
BATCH_SIZE = 2
N_ITER = 2
STEPS = 30
CFG_SCALE = 11.5
WIDTH = 512
HEIGHT = 768
SAMPLER_INDEX = 1
RESTORE_FACE = 'true'
TILING = 'false'
DO_NOT_SAVE_GRID = 'false'
EXCLUDED_TAGS = ['']
global SKIP_EXISTS
SKIP_EXISTS = True
OUTPUTS_DICT = list()
OUTPUTS = {}
rawDict = {}
qc_dict = {}
trg_img = ''
current_folder = ''
map_sampler_to_idx = {
'Euler a': 0,
'Euler': 1,
'LMS': 2,
'Heun': 3,
'DPM2': 4,
'DPM2 a': 5,
'DPM fast': 6,
'DPM adaptive': 7,
'LMS Karras': 8,
'DPM2 Karras': 9,
'DPM2 a Karras': 10,
'DDIM': 11,
'PLMS': 12}
map_keys = {
"value": "prompt",
"negative": "negative_prompt"}
map_param = {
"sd_model": "sd_model",
"outpath_samples": "outpath_samples",
"outpath_grids": "outpath_grids",
"prompt_for_display": "prompt_for_display",
"styles": "styles",
"Seed": "seed",
"Variation seed strength": "subseed_strength",
"Variation seed": "subseed",
"seed_resize_from_h": "seed_resize_from_h",
"seed_resize_from_w": "seed_resize_from_w",
"Sampler": "sampler_index",
"batch_size": "batch_size",
"n_iter": "n_iter",
"Steps": "steps",
"CFG scale": "cfg_scale",
"width": "width",
"height": "height",
"restore_faces": "restore_faces",
"tiling": "tiling",
"do_not_save_samples": "do_not_save_samples",
"do_not_save_grid": "do_not_save_grid"}
def process_string_tag(tag):
return tag
def process_int_tag(tag):
return int(tag)
def process_float_tag(tag):
return float(tag)
def process_boolean_tag(tag):
return True if (tag == "true") else False
prompt_tags = {
"sd_model": None,
"outpath_samples": process_string_tag,
"outpath_grids": process_string_tag,
"prompt_for_display": process_string_tag,
"prompt": process_string_tag,
"negative_prompt": process_string_tag,
"styles": process_string_tag,
"seed": process_int_tag,
"subseed_strength": process_float_tag,
"subseed": process_int_tag,
"seed_resize_from_h": process_int_tag,
"seed_resize_from_w": process_int_tag,
"sampler_index": process_int_tag,
"batch_size": process_int_tag,
"n_iter": process_int_tag,
"steps": process_int_tag,
"cfg_scale": process_float_tag,
"width": process_int_tag,
"height": process_int_tag,
"restore_faces": process_boolean_tag,
"tiling": process_boolean_tag,
"do_not_save_samples": process_boolean_tag,
"do_not_save_grid": process_boolean_tag
}
avatar_prompts = list()
avatar_names = list()
avatar_negatives = list()
avatar_name = ""
def cmdargs(line):
args = shlex.split(line)
pos = 0
res = {}
while pos < len(args):
arg = args[pos]
assert arg.startswith("--"), f'must start with "--": {arg}'
tag = arg[2:]
func = prompt_tags.get(tag, None)
assert func, f'unknown commandline option: {arg}'
assert pos+1 < len(args), f'missing argument for command line option {arg}'
val = args[pos+1]
res[tag] = func(val)
pos += 2
return res
def add_param(key, value, cur_str):
cur_str += '--{key} {value} '.format(key=key, value=value)
return cur_str
def parse_size(i_width, i_height, str_size, cur_str):
i_width = str_size.split('x')[0]
i_height = str_size.split('x')[1]
def parse_virariant_size(str_size, cur_str):
width = str_size.split('x')[0]
height = str_size.split('x')[1]
cur_str = add_param('seed_resize_from_w', width, cur_str)
cur_str = add_param('seed_resize_from_h', height, cur_str)
return cur_str
def parse_param(param_str):
m_batch_size = BATCH_SIZE
m_n_iter = N_ITER
m_steps = STEPS
m_cfg_scale = CFG_SCALE
m_width = WIDTH
m_height = HEIGHT
m_sampler_index = SAMPLER_INDEX
# m_tiling = TILING
m_restore_faces = RESTORE_FACE
m_do_not_save_grid = DO_NOT_SAVE_GRID
# m_sd_model = sd_model
cur_line = ""
for item in param_str.split(', '):
if item == '':
continue
group = item.split(': ')
key = group[0]
value = group[1]
if key == 'Steps':
m_steps = value
elif key == "CFG scale":
m_cfg_scale = value
elif value == 'Sampler':
m_sampler_index = map_sampler_to_idx[value]
elif key == 'Size':
parse_size(m_width, m_height, value, cur_line)
elif key == 'Seed resize from':
cur_line = parse_virariant_size(value, cur_line)
elif key == 'Seed':
cur_line = add_param("seed", value, cur_line)
elif key == 'Variation seed strength':
cur_line = add_param("subseed_strength", value, cur_line)
elif key == 'Variation seed':
cur_line = add_param("subseed", value, cur_line)
# elif key == 'Model hash':
# cur_line = add_param("sd_model", m_sd_model, cur_line)
cur_line = add_param("batch_size", m_batch_size, cur_line)
cur_line = add_param("n_iter", m_n_iter, cur_line)
cur_line = add_param("steps", m_steps, cur_line)
cur_line = add_param("cfg_scale", m_cfg_scale, cur_line)
cur_line = add_param("sampler_index", m_sampler_index, cur_line)
cur_line = add_param("width", m_width, cur_line)
cur_line = add_param("height", m_height, cur_line)
cur_line = add_param("restore_faces", m_restore_faces, cur_line)
# cur_line = add_param("tiling", m_tiling, cur_line)
cur_line = add_param("do_not_save_grid", m_do_not_save_grid, cur_line)
return cur_line
def parse_yaml_dict(rawDict, tag, avatar_prompt, avatar_name, default_negative):
# depth-first-search
if 'value' in rawDict.keys() or 'negative' in rawDict.keys():
if SKIP_EXISTS:
if os.path.exists(os.path.join(OUTPATH_SAMPLES, tag, avatar_name+'.png')) or os.path.exists(os.path.join(OUTPATH_SAMPLES,tag,'Not-available.png')):
print("Skip "+str(tag))
return ""
cur = ""
parsed_param = False
m_positive = avatar_prompt
m_negative = default_negative
for item in rawDict.items():
key = item[0]
value = item[1]
if key == 'param':
params = parse_param(rawDict['param'])
parsed_param = True
elif key == 'value':
m_positive = value + m_positive
elif key == 'negative':
m_negative = value +','+ m_negative
cur += "--{key} \"{value}\" ".format(key='prompt', value= m_positive)
cur += "--{key} \"{value}\" ".format(key='negative_prompt', value= m_negative)
if parsed_param == False:
params = parse_param("")
cur += params
cur += '--outpath_samples \"'+os.path.join(OUTPATH_SAMPLES, str(tag), '')+'\\\"'
cur += ' --outpath_grids \"'+os.path.join(OUTPATH_GRIDS, str(tag), '')+'\\\"'
# cur = add_param('outpath_samples=\"'+os.path.join(OUTPATH_SAMPLES, str(tag), '')+'\"', cur)
# cur = add_param('outpath_grids=\"'+os.path.join(OUTPATH_GRIDS, str(tag), '')+'\"', cur)
return cur
else:
for item in rawDict.items():
key = item[0]
ret = parse_yaml_dict(rawDict[key], tag if key=='' else key, avatar_prompt, avatar_name, default_negative)
if len(ret) != 0:
if tag not in EXCLUDED_TAGS:
OUTPUTS_DICT.append({'name': key,
'prompt': item[1]['value'] if 'value' in item[1].keys() else '',
'negative_prompt': item[1]['negative'] if 'negative' in item[1].keys() else ''})
if tag in OUTPUTS.keys():
OUTPUTS[tag].append(ret)
else:
OUTPUTS[tag] = [ret]
return ""
def rename_preview(avatar_name):
if avatar_name == '':
print("Please select avatar name first.")
return
root = OUTPATH_SAMPLES
for folder in os.listdir(root):
files = os.listdir(os.path.join(root, folder))
if 'Not-available.png' in files:
print('Skip '+ folder + ' not available.')
continue
if avatar_name + '.png' in files:
continue
for each_avatar in avatar_names:
if each_avatar + '.png' in files:
files.remove(each_avatar + '.png')
if len(files) == 1:
os.rename(os.path.join(root, folder, files[0]), os.path.join(root, folder, avatar_name + '.png'))
else:
print('There are 0 or more than 1 files in ' + folder)
def load_prompt_file(file):
if (file is None):
lines = []
else:
lines = [x.strip() for x in file.decode('utf8', errors='ignore').split("\n")]
return None, "\n".join(lines), gr.update(lines=7)
def copy_from_prompt_app():
return []
def open_folder(f):
if not os.path.exists(f):
print(f'Folder "{f}" does not exist. After you create an image, the folder will be created.')
return
elif not os.path.isdir(f):
print(f"""
WARNING
An open_folder request was made with an argument that is not a folder.
This could be an error or a malicious attempt to run code on your computer.
Requested path was: {f}
""", file=sys.stderr)
return
if not shared.cmd_opts.hide_ui_dir_config:
path = os.path.normpath(f)
if platform.system() == "Windows":
os.startfile(path)
elif platform.system() == "Darwin":
sp.Popen(["open", path])
else:
sp.Popen(["xdg-open", path])
class PromptStyle(typing.NamedTuple):
name: str
prompt: str
negative_prompt: str
def save_styles() -> None:
if len(OUTPUTS.keys()) == 0:
return
path = os.path.join(root_path, 'styles.csv')
# Write to temporary file first, so we don't nuke the file if something goes wrong
fd, temp_path = tempfile.mkstemp(".csv")
with os.fdopen(fd, "w", encoding="utf-8-sig", newline='') as file:
# _fields is actually part of the public API: typing.NamedTuple is a replacement for collections.NamedTuple,
# and collections.NamedTuple has explicit documentation for accessing _fields. Same goes for _asdict()
writer = csv.DictWriter(file, fieldnames=PromptStyle._fields)
writer.writeheader()
for row in OUTPUTS_DICT:
writer.writerow({'name': row['name'], 'prompt': row['prompt'], 'negative_prompt': row['negative_prompt']})
# writer.writerows(style._asdict() for k, style in self.styles.items())
# Always keep a backup file around
if os.path.exists(path):
shutil.move(path, path + ".bak")
shutil.move(temp_path, path)
def load_prompt(file, default_negative, dropdown, skip_exist):
global SKIP_EXISTS
SKIP_EXISTS = skip_exist
if dropdown == '' or file is None:
return
rawDict = yaml.load(file, Loader = yaml.BaseLoader)
default_negative = default_negative + ',' + avatar_negatives[avatar_names.index(dropdown)]
parse_yaml_dict(rawDict, "", avatar_prompts[avatar_names.index(dropdown)], dropdown, default_negative)
prompt_txt = ""
keys = list(filter(lambda x: x not in EXCLUDED_TAGS, OUTPUTS.keys()))
for style in keys:
for each_line in OUTPUTS[style]:
prompt_txt += each_line + '\n'
return [prompt_txt, gr.Row.update(visible=True)]
def load_avartar(avatar_dict, customize_tags_positive):
avatars = yaml.load(avatar_dict, yaml.BaseLoader)
for name, prompt in avatars.items():
avatar_names.append(name)
if 'value' in prompt.keys():
avatar_prompts.append(customize_tags_positive + ', ' + prompt['value'])
else:
avatar_prompts.append(customize_tags_positive + ', ' + '')
if 'negative' in prompt.keys():
avatar_negatives.append(prompt['negative'])
else:
avatar_negatives.append('')
return [gr.Dropdown.update(choices=avatar_names, value=avatar_names[0]), gr.Column.update(visible=True), gr.Group.update(visible=True)]
def scan_outputs(avatar_name):
if avatar_name is None or len(avatar_name) == 0:
print("Please select avatar name first.")
return
root = OUTPATH_SAMPLES
global qc_dict
qc_dict = {}
if not os.path.exists(root):
os.mkdir(root)
for folder in os.listdir(root):
if os.path.isdir(os.path.join(root, folder)) == False:
continue
files = os.listdir(os.path.join(root, folder))
if 'Not-available.png' in files:
print('Skip '+ folder + ' not available.')
continue
if avatar_name + '.png' in files:
continue
for each_avatar in avatar_names:
if each_avatar + '.png' in files:
files.remove(each_avatar + '.png')
if len(files) == 0:
continue
print("Insert file:")
for file in files:
print(os.path.join(root, folder, file))
qc_dict[folder] = [os.path.join(root, folder, file) for file in files]
if len(qc_dict.keys()) == 0:
return gr.Dropdown.update(choices=[])
return gr.Dropdown.update(choices=list(qc_dict.keys()), value=list(qc_dict.keys())[0])
def update_gallery(dropdown, avatar):
root = OUTPATH_SAMPLES
global trg_img, current_folder
current_folder = os.path.join(root, dropdown)
trg_img = os.path.join(root, dropdown, avatar + '.png')
print("Detected folders:")
print(qc_dict)
print("Selected folder:")
print(dropdown)
print("Detected files:")
print(qc_dict[dropdown])
return qc_dict[dropdown]
def clean_select_picture(filename):
if current_folder == '':
print("Please select qc tag.")
return
for file in os.listdir(current_folder):
is_avatar = False
for each_avatar in avatar_names:
if each_avatar + '.png' == file:
is_avatar = True
break
if '-' in file and file.split('-')[1] in filename:
print("rename", os.path.join(current_folder, file), trg_img)
os.rename(os.path.join(current_folder, file), trg_img)
elif is_avatar == False:
print(file, "delete", os.path.join(current_folder, file))
os.remove(os.path.join(current_folder, file))
def image_url(filedata):
if type(filedata) == dict and filedata["is_file"]:
filename = filedata["name"]
tempdir = os.path.normpath(tempfile.gettempdir())
normfn = os.path.normpath(filename)
assert normfn.startswith(tempdir), 'trying to open image file not in temporary directory'
image = Image.open(filename)
clean_select_picture(os.path.basename(filename))
return Image.open(filename)
if type(filedata) == list:
if len(filedata) == 0:
return None
filedata = filedata[0]
if filedata.startswith("data:image/png;base64,"):
filedata = filedata[len("data:image/png;base64,"):]
filedata = base64.decodebytes(filedata.encode('utf-8'))
image = Image.open(io.BytesIO(filedata))
return image
def dropdown_change():
global OUTPUTS, OUTPUTS_DICT
default_negative = []
OUTPUTS = {}
OUTPUTS_DICT = []
return [ gr.File.update(value=None), gr.Textbox.update(value=None)]
class Script(scripts.Script):
def title(self):
return "Prompt gallery"
def ui(self, is_img2img):
with gr.Group():
with gr.Column():
label_avatar = gr.Label("Upload avatars config")
avatar_dict = gr.File(label="Upload avatar prompt inputs", type='bytes')
# copy_from_app_button = gr.Button("Copy From Prompt Preview")
with gr.Group():
with gr.Column(visible=False) as avatar_col:
label_presets = gr.Label("Presets")
dropdown = gr.Dropdown(label="Choose avatar", choices=[""], value="", type="value", elem_id="dropdown")
dropdown.save_to_config = True
with gr.Row():
checkbox_iterate = gr.Checkbox(label="Iterate seed every line", value=False)
skip_exist = gr.Checkbox(value=True, label="skip exist")
default_negative = gr.Textbox(label="default_negative", lines=1)
default_positive = gr.Textbox(label="default_positive", lines=1)
prompt_dict = gr.File(label="Upload prompt dictionary", type='bytes')
with gr.Row(visible = False) as save_prompts:
open_button = gr.Button("Open outputs directory")
export_button = gr.Button("Export to WebUI style")
prompt_display = gr.Textbox(label="List of prompt inputs", lines=1)
prompt_dict.change(fn=load_prompt, inputs=[prompt_dict, default_negative, dropdown, skip_exist], outputs=[prompt_display, save_prompts])
open_button.click(fn=lambda: open_folder(OUTPATH_SAMPLES), inputs=[], outputs=[])
export_button.click(fn=save_styles, inputs=[], outputs=[])
with gr.Group(visible=False) as qc_widgets:
label_preview = gr.Label("QC preview")
with gr.Row():
qc_refresh = gr.Button("QC scan")
preview_dropdown = gr.Dropdown(label="Select prompts", choices=[""], value="", type="value", elem_id="dropdown")
preview_gallery = gr.Gallery(label='Output', show_label=False, elem_id=f"preview_gallery").style(grid=4)
qc_refresh.click(fn=scan_outputs, inputs=[dropdown], outputs=preview_dropdown)
with gr.Row():
qc_show = gr.Button(f"Show pics")
qc_select = gr.Button(f"Select")
rename_button = gr.Button("Auto rename")
selected_img = gr.Image(label="Selected",show_label=False, source="upload", interactive=True, type="pil").style(height=480)
qc_show.click(fn=update_gallery, inputs=[preview_dropdown, dropdown], outputs=preview_gallery)
qc_select.click(
fn=lambda x: image_url(x),
_js="extract_image_from_gallery",
inputs=[preview_gallery],
outputs=[selected_img],
)
# qc_select.click(fn=select_picture, inputs=[dropdown, preview_dropdown, preview_gallery], outputs=[])
dropdown.change(fn=dropdown_change, inputs=[], outputs=[prompt_dict, prompt_display])
rename_button.click(fn=rename_preview, inputs=[dropdown], outputs=[])
# qc_select.click(fn=scan_outputs, inputs=[], outputs=[preview_dropdown])
avatar_dict.change(fn=load_avartar, inputs=[avatar_dict, default_positive], outputs=[dropdown, avatar_col, qc_widgets])
return [checkbox_iterate, avatar_dict, prompt_dict, default_negative, default_positive, dropdown, prompt_display, rename_button, label_avatar, open_button, export_button, skip_exist, label_presets, label_preview, preview_dropdown, preview_gallery, qc_select, qc_refresh, qc_show, selected_img]
def run(self, p, checkbox_iterate, avatar_dict, prompt_dict, default_negative, default_positive, dropdown, prompt_display, rename_button, label_avatar, open_button, export_button, skip_exist, label_presets, label_preview, preview_dropdown, preview_gallery, qc_select, qc_refresh, qc_show, selected_img):
lines = [x.strip() for x in prompt_display.splitlines()]
lines = [x for x in lines if len(x) > 0]
p.do_not_save_grid = True
job_count = 0
jobs = []
for line in lines:
if "--" in line:
try:
args = cmdargs(line)
except Exception:
print(f"Error parsing line [line] as commandline:", file=sys.stderr)
print(traceback.format_exc(), file=sys.stderr)
args = {"prompt": line}
else:
args = {"prompt": line}
n_iter = args.get("n_iter", 1)
if n_iter != 1:
job_count += n_iter
else:
job_count += 1
jobs.append(args)
print(f"Will process {len(lines)} lines in {job_count} jobs.")
if (checkbox_iterate and p.seed == -1):
p.seed = int(random.randrange(4294967294))
state.job_count = job_count
images = []
for n, args in enumerate(jobs):
state.job = f"{state.job_no + 1} out of {state.job_count}"
copy_p = copy.copy(p)
for k, v in args.items():
setattr(copy_p, k, v)
proc = process_images(copy_p)
images += proc.images
if (checkbox_iterate):
p.seed = p.seed + (p.batch_size * p.n_iter)
OUTPUTS = {}
rawDict = {}
return Processed(p, images, p.seed, "")