Spaces:
Sleeping
Sleeping
File size: 7,798 Bytes
b511d75 8d1bd48 19338e6 83a7945 19338e6 9abec8f 19338e6 70ebe1c 19338e6 8d1bd48 83a7945 19338e6 83a7945 b511d75 19338e6 2dea97c 19338e6 bc1dcdf 19338e6 bc1dcdf 19338e6 bc1dcdf 19338e6 bc1dcdf 19338e6 bc1dcdf 9abec8f cb93ca5 bc1dcdf cb93ca5 bc1dcdf 70ebe1c bc1dcdf 19338e6 bc1dcdf 19338e6 bc1dcdf 19338e6 bc1dcdf 9abec8f cb93ca5 9abec8f 19338e6 9abec8f bc1dcdf 9abec8f bc1dcdf 9abec8f bc1dcdf 19338e6 bc1dcdf 70ebe1c bc1dcdf 19338e6 2dea97c 19338e6 bc1dcdf 19338e6 9abec8f bc1dcdf 70ebe1c bc1dcdf cb93ca5 19338e6 bc1dcdf 70ebe1c 9abec8f bc1dcdf 19338e6 cb93ca5 19338e6 bc1dcdf 70ebe1c bc1dcdf 70ebe1c bc1dcdf cb93ca5 bc1dcdf cb93ca5 bc1dcdf 19338e6 bc1dcdf 70ebe1c bc1dcdf 70ebe1c bc1dcdf 70ebe1c bc1dcdf 70ebe1c 831308c cb93ca5 70ebe1c bc1dcdf 70ebe1c cb93ca5 bc1dcdf 70ebe1c bc1dcdf 70ebe1c bc1dcdf 19338e6 70ebe1c 19338e6 bc1dcdf 70ebe1c 19338e6 70ebe1c |
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 |
import os
import gradio as gr
import json
import time
import traceback
import io
import base64
from PIL import Image, ImageEnhance, ImageFilter
# --- Environment Configuration ---
GEMINI_KEY = os.environ.get("GEMINI_KEY", "")
DEFAULT_PORT = int(os.environ.get("PORT", 7860))
API_TIMEOUT = 30 # seconds
# --- Style Template Optimization ---
BASE_TEMPLATE = """Describe this design as a concise Flux 1.1 Pro prompt focusing on:
- Key visual elements
- Technical specifications
- Style consistency
- Functional requirements"""
STYLE_INSTRUCTIONS = {
"General": BASE_TEMPLATE,
"Realistic": f"{BASE_TEMPLATE}\nPHOTOREALISTIC RULES: Use photography terms, texture details, accurate lighting",
"Kawaii": f"{BASE_TEMPLATE}\nKAWAII RULES: Rounded shapes, pastel colors, cute expressions",
"Vector": f"{BASE_TEMPLATE}\nVECTOR RULES: Clean lines, geometric shapes, B&W gradients",
"Silhouette": f"{BASE_TEMPLATE}\nSILHOUETTE RULES: High contrast, minimal details, strong outlines"
}
# --- Flux Configuration ---
FLUX_SPECS = {
"aspect_ratios": ["1:1", "16:9", "4:3", "9:16"],
"formats": ["SVG", "PNG", "PDF"],
"color_modes": ["B&W", "CMYK", "RGB"],
"dpi_options": [72, 150, 300, 600]
}
# --- Image Processing Pipeline ---
def preprocess_image(img):
try:
if isinstance(img, str): # Handle file paths
img = Image.open(img)
img = img.convert("RGB")
img = ImageEnhance.Contrast(img).enhance(1.2)
img = img.filter(ImageFilter.SHARPEN)
return img
except Exception as e:
raise ValueError(f"๐ด Image processing failed: {str(e)}")
# --- Core Generation Engine ---
def generate_prompt(image, api_key, style, creativity, neg_prompt, aspect, color_mode, dpi):
try:
# Step 1: Input Validation
if not image:
return {"error": "โ ๏ธ Please upload an image."}, None, None
api_key = api_key or GEMINI_KEY
if not api_key:
return {"error": "๐ API key required - set in env (GEMINI_KEY) or input field."}, None, None
# Step 2: Gemini Setup
try:
import google.generativeai as genai
genai.configure(api_key=api_key)
model = genai.GenerativeModel("gemini-1.5-pro")
except ImportError:
return {"error": "๐ซ Failed to import google.generativeai. Install with: pip install google-generativeai"}, None, None
except Exception as e:
if "authentication" in str(e).lower():
return {"error": "๐ Invalid API key or authentication error"}, None, None
else:
return {"error": f"โ๏ธ API initialization error: {str(e)}"}, None, None
# Step 3: Preprocess Image
try:
img = preprocess_image(image)
img_bytes = io.BytesIO()
img.save(img_bytes, format="PNG")
img_b64 = base64.b64encode(img_bytes.getvalue()).decode()
except Exception as e:
return {"error": f"๐ผ๏ธ Image preparation failed: {str(e)}"}, None, None
# Step 4: Build Instruction Prompt
try:
instruction = f"{STYLE_INSTRUCTIONS[style]}\nAVOID: {neg_prompt}\n"
instruction += f"ASPECT: {aspect}, COLORS: {color_mode}, DPI: {dpi}\n"
except KeyError:
return {"error": "๐ ๏ธ Invalid style selected. Please choose from available options."}, None, None
# Step 5: Call Gemini API
try:
response = model.generate_content(
contents=[instruction, {"mime_type": "image/png", "data": img_b64}],
generation_config={"temperature": creativity}
)
raw_prompt = response.text
except Exception as e:
return {"error": f"๐ค Prompt generation failed: {str(e)}"}, None, None
# Step 6: Quality & Token Stats
validation = {"score": 8, "issues": [], "suggestions": []}
input_tokens = len(img_b64) // 4
output_tokens = len(raw_prompt.split())
return {
"prompt": raw_prompt,
"validation": validation,
"stats": {"input": input_tokens, "output": output_tokens}
}, validation, {"tokens": f"Input: {input_tokens}, Output: {output_tokens}"}
except Exception as e:
traceback.print_exc()
return {"error": f"๐ฅ Unexpected error: {str(e)}"}, None, None
# --- Modern Copy Function ---
def copy_text(text):
return text, "โ Copied to clipboard!", gr.Button.update(variant="secondary")
# --- Main Interface ---
def build_interface():
global STYLE_INSTRUCTIONS # Ensure access to global dict
with gr.Blocks(title="Flux Pro Generator", theme="soft") as app:
gr.Markdown("# ๐จ Flux Pro Prompt Generator")
gr.Markdown("Generate optimized design prompts from images using Google's Gemini.")
with gr.Row():
with gr.Column(scale=1):
api_key = gr.Textbox(
label="๐ Gemini API Key",
value=GEMINI_KEY,
type="password",
info="Set GEMINI_KEY environment variable for production."
)
img_input = gr.Image(
label="๐ผ๏ธ Upload Design",
type="pil",
sources=["upload"],
interactive=True
)
style = gr.Dropdown(
list(STYLE_INSTRUCTIONS.keys()),
value="General",
label="๐จ Target Style"
)
with gr.Accordion("โ๏ธ Advanced Settings", open=False):
creativity = gr.Slider(0.0, 1.0, 0.7, label="๐ง Creativity Level")
neg_prompt = gr.Textbox(label="๐ซ Negative Prompts", placeholder="What to avoid")
aspect = gr.Dropdown(FLUX_SPECS["aspect_ratios"], value="1:1", label="Aspect Ratio")
color_mode = gr.Dropdown(FLUX_SPECS["color_modes"], value="RGB", label="Color Mode")
dpi = gr.Dropdown([str(d) for d in FLUX_SPECS["dpi_options"]], value="300", label="Output DPI")
gen_btn = gr.Button("โจ Generate Prompt", variant="primary")
with gr.Column(scale=2):
status_msg = gr.Textbox(label="๐ข Status", interactive=False)
prompt_output = gr.Textbox(
label="๐ Optimized Prompt",
lines=8,
interactive=True,
show_copy_button=True
)
with gr.Row():
copy_btn = gr.Button("๐ Copy to Clipboard", variant="secondary")
quality_report = gr.JSON(label="๐ Quality Report")
token_stats = gr.JSON(label="๐งฎ Token Usage")
# Event Handling
gen_btn.click(
fn=generate_prompt,
inputs=[
img_input, api_key, style, creativity,
neg_prompt, aspect, color_mode, dpi
],
outputs=[prompt_output, quality_report, token_stats],
api_name="generate"
).then(
fn=lambda: "โ
Prompt generated successfully!",
outputs=status_msg
)
copy_btn.click(
fn=copy_text,
inputs=prompt_output,
outputs=[prompt_output, status_msg, copy_btn],
js="""
(text) => {
if(text) navigator.clipboard.writeText(text);
return [text, 'โ Copied!', { variant: 'secondary' }];
}
"""
)
return app
# --- Production Launch ---
if __name__ == "__main__":
app = build_interface()
app.launch() |