File size: 13,708 Bytes
c336648 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 |
import time
from typing import Union
import gradio
from jinja2 import Template
from modules import script_callbacks, scripts
from modules.processing import Processed, StableDiffusionProcessing
from modules.ui_components import InputAccordion
from lib.custom_text_overlay.align import Position
from lib.custom_text_overlay.drawText import drawText
from lib.custom_text_overlay.logger import logger
from lib.custom_text_overlay.options import getOption, onUiSettings
from src.custom_text_overlay.extension import extensionId, extensionTitle
templateBracketLeft = '{{'
templateBracketRight = '}}'
normalFontSize = 32 # for 1024x1024 pixels or anything with the same megapixel value
hookType = 'postprocess_image'
keysFromImg = ['cfg_scale', 'width', 'height', 'seed', 'subseed', 'prompt', 'negative_prompt', 'steps']
specialReplacements = {
'all_seeds': 'seed',
'all_subseeds': 'subseed',
'all_prompts': 'prompt',
'all_negative_prompts': 'negative_prompt',
}
class CustomTextOverlay(scripts.Script):
def title(self):
return extensionTitle
def show(self, is_img2img):
return scripts.AlwaysVisible
def ui(self, is_img2img):
minWidth = 200
templateEngine = getOption('template_engine')
useExamples = getOption('examples')
def getTemplateInput(positionLabel: str, defaultValue: str = '') -> tuple[gradio.Checkbox, gradio.Textbox]:
if not useExamples:
defaultValue = ''
checkbox = gradio.Checkbox(label=f'{positionLabel} text', value=True)
textbox = gradio.Textbox(label=f'{positionLabel} text template', value=defaultValue, lines=1, show_label=False)
return (checkbox, textbox)
with InputAccordion(False, label=extensionTitle, elem_id=self.elem_id(extensionId)) as enabled:
with gradio.Accordion('Text', open=True):
with gradio.Row():
with gradio.Column(min_width=minWidth):
timeTemplate = "{{ ('%.1f'|format(time)).rstrip('0').rstrip('.') }} sec\n{{ ('%.1f'|format(steps / time)).rstrip('0').rstrip('.') }} steps/s" if templateEngine == 'jinja2' else '{{time}}s'
(textEnabled1, textTemplate1) = getTemplateInput('Top left', timeTemplate)
with gradio.Column(min_width=minWidth):
(textEnabled2, textTemplate2) = getTemplateInput('Top')
with gradio.Column(min_width=minWidth):
(textEnabled3, textTemplate3) = getTemplateInput('Top right')
with gradio.Row():
with gradio.Column(min_width=minWidth):
(textEnabled4, textTemplate4) = getTemplateInput('Left')
with gradio.Column(min_width=minWidth):
(textEnabled5, textTemplate5) = getTemplateInput('Center')
with gradio.Column(min_width=minWidth):
(textEnabled6, textTemplate6) = getTemplateInput('Right')
with gradio.Row():
with gradio.Column(min_width=minWidth):
seedTemplate = 'Seed {{seed}}'
(textEnabled7, textTemplate7) = getTemplateInput('Bottom left', seedTemplate)
with gradio.Column(min_width=minWidth):
(textEnabled8, textTemplate8) = getTemplateInput('Bottom')
with gradio.Column(min_width=minWidth):
(textEnabled9, textTemplate9) = getTemplateInput('Bottom right')
with gradio.Accordion('Style'):
gradio.HTML('<p><h2>General</h2><hr style="margin-top: 0; margin-bottom: 1em"/></p>')
with gradio.Row():
with gradio.Column(min_width=minWidth):
textColor = gradio.ColorPicker(label='Text color', value='#ffffff')
textScale = gradio.Slider(minimum=20, maximum=300, step=10, label='Text scale', value=120)
with gradio.Column(min_width=minWidth):
paddingScale = gradio.Slider(minimum=0, maximum=200, step=5, label='Padding scale', value=25)
marginScale = gradio.Slider(minimum=0, maximum=200, step=5, label='Margin scale', value=0)
gradio.HTML('<p style="margin-top: 2em"><h2>Outline</h2><hr style="margin-top: 0; margin-bottom: 1em"/></p>')
with gradio.Row():
outlineScale = gradio.Slider(minimum=0, maximum=25, step=1, label='Outline scale', value=12)
outlineColor = gradio.ColorPicker(label='Outline color', value='#000000')
outlineOpacity = gradio.Slider(minimum=0, maximum=100, step=5, label='Outline opacity', value=100)
gradio.HTML('<p style="margin-top: 2em"><h2>Background Box</h2><hr style="margin-top: 0; margin-bottom: 1em"/></p>')
with gradio.Row():
backgroundColor = gradio.ColorPicker(label='Background color', value='#000000')
backgroundOpacity = gradio.Slider(minimum=0, maximum=100, step=5, label='Background opacity', value=0)
return [enabled, textScale, textColor, backgroundColor, backgroundOpacity, paddingScale, marginScale, outlineScale, outlineColor, outlineOpacity, textEnabled1, textEnabled2, textEnabled3, textEnabled4, textEnabled5, textEnabled6, textEnabled7, textEnabled8, textEnabled9, textTemplate1, textTemplate2, textTemplate3, textTemplate4, textTemplate5, textTemplate6, textTemplate7, textTemplate8, textTemplate9]
def process(self, processing: StableDiffusionProcessing, enabled: bool, textScale: int, textColor: str, backgroundColor: str, backgroundOpacity: int, paddingScale: int, marginScale: int, outlineScale: int, outlineColor: str, outlineOpacity: int, textEnabled1: bool, textEnabled2: bool, textEnabled3: bool, textEnabled4: bool, textEnabled5: bool, textEnabled6: bool, textEnabled7: bool, textEnabled8: bool, textEnabled9: bool, textTemplate1: str, textTemplate2: str, textTemplate3: str, textTemplate4: str, textTemplate5: str, textTemplate6: str, textTemplate7: str, textTemplate8: str, textTemplate9: str):
if not enabled:
return
self.startTime = time.perf_counter()
def collectReplacements(self, staticReplacements: dict = {}, replacementSources: dict = {}, imageIndex: int = 0, timeSeconds: float = 0):
tempateEngine = getOption('template_engine')
logger.info(tempateEngine)
replacements = self.makeReplacementTable(staticReplacements, keysFromImg, replacementSources)
if timeSeconds is not None:
if tempateEngine == 'jinja2':
replacements['time'] = timeSeconds
else:
replacements['time.00'] = f'{timeSeconds:.2f}'
replacements['time.0'] = f'{timeSeconds:.1f}'
replacements['time'] = int(timeSeconds)
if imageIndex is not None:
replacements['image_index'] = imageIndex
if tempateEngine == 'jinja2':
for sourceName, source in replacementSources.items():
replacements[sourceName] = source
for arrayKey, singleKey in specialReplacements.items():
for replacementSource in replacementSources.values():
if hasattr(replacementSource, arrayKey):
value = getattr(replacementSource, arrayKey)
if value is not None:
if imageIndex < len(value):
replacements[singleKey] = str(value[imageIndex])
break
return replacements
def makeReplacementTable(self, baseReplacements: dict, needles: list, haystacks: dict):
replacements = baseReplacements.copy()
for needle in needles:
for hackstackName, hackstack in haystacks.items():
value = getattr(hackstack, needle, None)
if value is not None:
logger.debug(f'{templateBracketLeft}{needle}{templateBracketRight} = {hackstackName}.{needle}')
replacements[needle] = value
break
else:
logger.debug(f'{templateBracketLeft}{needle}{templateBracketRight} = ?')
replacements[needle] = '?'
return replacements
def applyReplacements(self, templateString: str, replacements: dict):
templateEngine = getOption('template_engine')
templateHandler = self.applyReplacementsJinja if templateEngine == 'jinja2' else self.applyReplacementsBasic
inputTemplate = templateString.strip()
output = templateHandler(inputTemplate, replacements)
if inputTemplate == output:
return inputTemplate
logger.debug(f'Resolving template “{inputTemplate}” to “{output}” with engine {templateEngine} and keys {", ".join(replacements.keys())}')
return output
def applyReplacementsJinja(self, templateString: str, replacements: dict):
jinjaTemplate = Template(templateString)
return jinjaTemplate.render(replacements)
def applyReplacementsBasic(self, templateString: str, replacements: dict):
for key, value in replacements.items():
if isinstance(value, int) or isinstance(value, float):
value = str(value)
if isinstance(value, list):
value = ', '.join(value)
if not isinstance(value, str):
continue
templateString = templateString.replace(f'{templateBracketLeft}{key}{templateBracketRight}', value)
return templateString
def postprocess_image(self, processing: StableDiffusionProcessing, processed, enabled: bool, textScale: int, textColor: str, backgroundColor: str, backgroundOpacity: int, paddingScale: int, marginScale: int, outlineScale: int, outlineColor: str, outlineOpacity: int, textEnabled1: bool, textEnabled2: bool, textEnabled3: bool, textEnabled4: bool, textEnabled5: bool, textEnabled6: bool, textEnabled7: bool, textEnabled8: bool, textEnabled9: bool, textTemplate1: str, textTemplate2: str, textTemplate3: str, textTemplate4: str, textTemplate5: str, textTemplate6: str, textTemplate7: str, textTemplate8: str, textTemplate9: str):
if not enabled or hookType != 'postprocess_image':
return
enabledTextTemplates: list[int] = []
for textTemplateIndex in range(1, 10):
if locals()[f'textEnabled{textTemplateIndex}'] and locals()[f'textTemplate{textTemplateIndex}'].strip() != '':
enabledTextTemplates.append(textTemplateIndex)
if len(enabledTextTemplates) == 0:
return
endTime = time.perf_counter()
generationTimeSeconds = endTime - self.startTime
self.startTime = endTime
furtherReplacementsSources = {
'img': processed.image,
'processed': processed,
'processing': processing,
}
replacements = self.collectReplacements(timeSeconds=generationTimeSeconds, replacementSources=furtherReplacementsSources, imageIndex=processing.iteration)
for textTemplateIndex in enabledTextTemplates:
textTemplate = locals()[f'textTemplate{textTemplateIndex}'].strip()
text = self.applyReplacements(textTemplate, replacements)
position = Position(textTemplateIndex)
pixels = processed.image.width * processed.image.height
fontSize = int((textScale / 100) * (normalFontSize * (pixels / 1048576)))
margin = int((marginScale / 100) * fontSize)
padding = int((paddingScale / 100) * fontSize)
outline = int((outlineScale / 100) * fontSize)
processed.image = drawText(processed.image, text=text, fontSize=fontSize, textColor=textColor, position=position, backgroundColor=backgroundColor, backgroundOpacity=backgroundOpacity, margin=margin, padding=padding, outline=outline, outlineColor=outlineColor, outlineOpacity=outlineOpacity)
def postprocess(self, processing: StableDiffusionProcessing, processed: Processed, enabled: bool, textScale: int, textColor: str, backgroundColor: str, backgroundOpacity: int, paddingScale: int, marginScale: int, outlineScale: int, outlineColor: str, outlineOpacity: int, textEnabled1: bool, textEnabled2: bool, textEnabled3: bool, textEnabled4: bool, textEnabled5: bool, textEnabled6: bool, textEnabled7: bool, textEnabled8: bool, textEnabled9: bool, textTemplate1: str, textTemplate2: str, textTemplate3: str, textTemplate4: str, textTemplate5: str, textTemplate6: str, textTemplate7: str, textTemplate8: str, textTemplate9: str):
if not enabled or hookType != 'postprocess':
return
enabledTextTemplates: list[int] = []
for textTemplateIndex in range(1, 10):
if locals()[f'textEnabled{textTemplateIndex}'] and locals()[f'textTemplate{textTemplateIndex}'].strip() != '':
enabledTextTemplates.append(textTemplateIndex)
if len(enabledTextTemplates) == 0:
return
endTime = time.perf_counter()
generationTimeSeconds = endTime - self.startTime
images = processed.images[processed.index_of_first_image:]
for imageIndex in range(len(images)):
processedImageIndex = imageIndex + processed.index_of_first_image
img = images[imageIndex]
replacements = {
'processed_image_index': processedImageIndex,
'real_image_index': imageIndex,
}
furtherReplacementsSources = {
'img': img,
'processed': processed,
'processing': processing,
}
replacements = self.collectReplacements(staticReplacements=replacements, replacementSources=furtherReplacementsSources, imageIndex=imageIndex, timeSeconds=generationTimeSeconds)
for textTemplateIndex in enabledTextTemplates:
textTemplate = locals()[f'textTemplate{textTemplateIndex}'].strip()
text = self.applyReplacements(textTemplate, replacements)
position = Position(textTemplateIndex)
pixels = img.width * img.height
fontSize = int((textScale / 100) * (normalFontSize * (pixels / 1048576)))
margin = int((marginScale / 100) * fontSize)
padding = int((paddingScale / 100) * fontSize)
outline = int((outlineScale / 100) * fontSize)
img = drawText(img, text=text, fontSize=fontSize, textColor=textColor, position=position, backgroundColor=backgroundColor, backgroundOpacity=backgroundOpacity, margin=margin, padding=padding, outline=outline, outlineColor=outlineColor, outlineOpacity=outlineOpacity)
processed.images[processedImageIndex] = img
script_callbacks.on_ui_settings(onUiSettings)
|