Spaces:
Sleeping
Sleeping
Add match image objective
Browse files
app.py
CHANGED
@@ -2,12 +2,13 @@ import gradio as gr
|
|
2 |
from playwright.async_api import async_playwright, Page
|
3 |
from PIL import Image
|
4 |
from io import BytesIO
|
5 |
-
from anthropic import Anthropic
|
6 |
from dotenv import load_dotenv
|
7 |
import os
|
8 |
from typing import Literal
|
9 |
import time
|
10 |
from base64 import b64encode
|
|
|
11 |
|
12 |
load_dotenv()
|
13 |
# check for ANTHROPIC_API_KEY
|
@@ -53,23 +54,25 @@ Assume that the content is being inserted into a template like this:
|
|
53 |
{apply_tailwind("your html here")}
|
54 |
"""
|
55 |
|
56 |
-
improve_prompt = """
|
57 |
-
Given the current draft of the webpage you generated for me as HTML and the screenshot of it rendered, improve the HTML to look nicer.
|
58 |
-
"""
|
59 |
|
|
|
|
|
|
|
|
|
60 |
|
61 |
-
|
|
|
|
|
|
|
62 |
with anthropic.messages.stream(
|
63 |
model=model,
|
64 |
-
max_tokens=
|
65 |
-
system=
|
66 |
-
messages=
|
67 |
-
{"role": "user", "content": prompt},
|
68 |
-
],
|
69 |
) as stream:
|
70 |
-
for
|
71 |
-
|
72 |
-
|
73 |
|
74 |
|
75 |
def format_image(image: bytes, media_type: Literal["image/png", "image/jpeg"]):
|
@@ -84,13 +87,14 @@ def format_image(image: bytes, media_type: Literal["image/png", "image/jpeg"]):
|
|
84 |
}
|
85 |
|
86 |
|
87 |
-
def
|
88 |
"""
|
89 |
history is a list of tuples of (content, image) corresponding to iterations of generation and rendering
|
90 |
"""
|
91 |
-
|
92 |
-
|
93 |
-
|
|
|
94 |
{"role": "user", "content": prompt},
|
95 |
*[
|
96 |
item
|
@@ -118,27 +122,57 @@ def stream_with_visual_feedback(prompt, history: list[tuple[str, bytes]]):
|
|
118 |
],
|
119 |
]
|
120 |
|
121 |
-
|
122 |
-
|
123 |
-
|
124 |
-
|
125 |
-
|
126 |
-
|
127 |
-
|
128 |
-
|
129 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
130 |
|
131 |
|
132 |
async def render_html(page: Page, content: str):
|
133 |
-
start_time =
|
134 |
await page.set_content(content)
|
135 |
# weird, can we set scale to 2.0 directly instead of "device", ie whatever server this is running on?
|
136 |
image_bytes = await page.screenshot(type="png", scale="device", full_page=True)
|
137 |
-
|
138 |
-
|
139 |
-
|
140 |
-
def t():
|
141 |
-
return time.perf_counter()
|
142 |
|
143 |
|
144 |
def apply_template(content, template):
|
@@ -151,48 +185,78 @@ def to_pil(image_bytes: bytes):
|
|
151 |
return Image.open(BytesIO(image_bytes))
|
152 |
|
153 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
154 |
async def generate_with_visual_feedback(
|
155 |
prompt,
|
156 |
template,
|
157 |
resolution: str = "512",
|
158 |
num_iterations: int = 1,
|
159 |
):
|
160 |
-
|
161 |
-
|
162 |
-
async with async_playwright() as p:
|
163 |
-
browser = await p.chromium.launch()
|
164 |
-
page = await browser.new_page(
|
165 |
-
viewport={"width": resolution[0], "height": resolution[1]}
|
166 |
-
)
|
167 |
-
last_yield = t()
|
168 |
history = []
|
169 |
for i in range(num_iterations):
|
170 |
-
|
171 |
-
|
172 |
if i == 0
|
173 |
-
else
|
174 |
)
|
175 |
content = ""
|
176 |
-
for
|
177 |
-
|
178 |
-
|
179 |
-
|
180 |
-
|
181 |
-
|
182 |
-
|
183 |
-
|
184 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
185 |
# always render the final image of each iteration
|
186 |
image_bytes, render_time = await render_html(
|
187 |
page, apply_template(content, template)
|
188 |
)
|
189 |
history.append((content, image_bytes))
|
190 |
-
yield to_pil(image_bytes), content, render_time
|
191 |
-
# cleanup
|
192 |
-
await browser.close()
|
193 |
|
194 |
|
195 |
-
|
196 |
generate_with_visual_feedback,
|
197 |
inputs=[
|
198 |
gr.Textbox(
|
@@ -212,6 +276,26 @@ demo = gr.Interface(
|
|
212 |
],
|
213 |
)
|
214 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
215 |
|
216 |
if __name__ == "__main__":
|
217 |
prepare_playwright_if_needed()
|
|
|
2 |
from playwright.async_api import async_playwright, Page
|
3 |
from PIL import Image
|
4 |
from io import BytesIO
|
5 |
+
from anthropic import Anthropic
|
6 |
from dotenv import load_dotenv
|
7 |
import os
|
8 |
from typing import Literal
|
9 |
import time
|
10 |
from base64 import b64encode
|
11 |
+
from contextlib import asynccontextmanager
|
12 |
|
13 |
load_dotenv()
|
14 |
# check for ANTHROPIC_API_KEY
|
|
|
54 |
{apply_tailwind("your html here")}
|
55 |
"""
|
56 |
|
|
|
|
|
|
|
57 |
|
58 |
+
def messages_text_to_web(prompt):
|
59 |
+
return [
|
60 |
+
{"role": "user", "content": prompt},
|
61 |
+
]
|
62 |
|
63 |
+
|
64 |
+
# returns the full text of the response each time
|
65 |
+
def stream_claude(messages, system=system_prompt, max_tokens=2000):
|
66 |
+
text = ""
|
67 |
with anthropic.messages.stream(
|
68 |
model=model,
|
69 |
+
max_tokens=max_tokens,
|
70 |
+
system=system,
|
71 |
+
messages=messages,
|
|
|
|
|
72 |
) as stream:
|
73 |
+
for chunk in stream.text_stream:
|
74 |
+
text += chunk
|
75 |
+
yield text
|
76 |
|
77 |
|
78 |
def format_image(image: bytes, media_type: Literal["image/png", "image/jpeg"]):
|
|
|
87 |
}
|
88 |
|
89 |
|
90 |
+
def visual_feedback_messages(prompt, history: list[tuple[str, bytes]]):
|
91 |
"""
|
92 |
history is a list of tuples of (content, image) corresponding to iterations of generation and rendering
|
93 |
"""
|
94 |
+
improve_prompt = """
|
95 |
+
Given the current draft of the webpage you generated for me as HTML and the screenshot of it rendered, improve the HTML to look nicer.
|
96 |
+
"""
|
97 |
+
return [
|
98 |
{"role": "user", "content": prompt},
|
99 |
*[
|
100 |
item
|
|
|
122 |
],
|
123 |
]
|
124 |
|
125 |
+
|
126 |
+
def match_image_messages(image_bytes: bytes, history: list[tuple[bytes, bytes]]):
|
127 |
+
improve_prompt = """
|
128 |
+
Given the current draft of the webpage you generated for me as HTML and the original screenshot, improve the HTML to match closer to the original screenshot.
|
129 |
+
"""
|
130 |
+
|
131 |
+
return [
|
132 |
+
{
|
133 |
+
"role": "user",
|
134 |
+
"content": [
|
135 |
+
{
|
136 |
+
"type": "text",
|
137 |
+
"text": "Please generate a webpage that matches the image below as closely as possible:",
|
138 |
+
},
|
139 |
+
format_image(image_bytes, "image/png"),
|
140 |
+
],
|
141 |
+
},
|
142 |
+
*[
|
143 |
+
item
|
144 |
+
for content, image_bytes in history
|
145 |
+
for item in [
|
146 |
+
{
|
147 |
+
"role": "assistant",
|
148 |
+
"content": content,
|
149 |
+
},
|
150 |
+
{
|
151 |
+
"role": "user",
|
152 |
+
"content": [
|
153 |
+
{
|
154 |
+
"type": "text",
|
155 |
+
"text": "Here is a screenshot of the above HTML code rendered in a browser:",
|
156 |
+
},
|
157 |
+
format_image(image_bytes, "image/png"),
|
158 |
+
{
|
159 |
+
"type": "text",
|
160 |
+
"text": improve_prompt,
|
161 |
+
},
|
162 |
+
],
|
163 |
+
},
|
164 |
+
]
|
165 |
+
],
|
166 |
+
]
|
167 |
|
168 |
|
169 |
async def render_html(page: Page, content: str):
|
170 |
+
start_time = time.perf_counter()
|
171 |
await page.set_content(content)
|
172 |
# weird, can we set scale to 2.0 directly instead of "device", ie whatever server this is running on?
|
173 |
image_bytes = await page.screenshot(type="png", scale="device", full_page=True)
|
174 |
+
dt = time.perf_counter() - start_time
|
175 |
+
return image_bytes, dt
|
|
|
|
|
|
|
176 |
|
177 |
|
178 |
def apply_template(content, template):
|
|
|
185 |
return Image.open(BytesIO(image_bytes))
|
186 |
|
187 |
|
188 |
+
@asynccontextmanager
|
189 |
+
async def browser(width, height):
|
190 |
+
async with async_playwright() as p:
|
191 |
+
browser = await p.chromium.launch()
|
192 |
+
page = await browser.new_page(viewport={"width": width, "height": height})
|
193 |
+
try:
|
194 |
+
yield page
|
195 |
+
finally:
|
196 |
+
await browser.close()
|
197 |
+
|
198 |
+
|
199 |
+
async def throttle(generator, every=0.25):
|
200 |
+
last_emit_time = 0
|
201 |
+
for item in generator:
|
202 |
+
current_time = time.perf_counter()
|
203 |
+
if current_time - last_emit_time >= every:
|
204 |
+
yield item
|
205 |
+
last_emit_time = current_time
|
206 |
+
# always emit the last item
|
207 |
+
yield item
|
208 |
+
|
209 |
+
|
210 |
async def generate_with_visual_feedback(
|
211 |
prompt,
|
212 |
template,
|
213 |
resolution: str = "512",
|
214 |
num_iterations: int = 1,
|
215 |
):
|
216 |
+
width = {"512": 512, "1024": 1024}[resolution]
|
217 |
+
async with browser(width, width) as page:
|
|
|
|
|
|
|
|
|
|
|
|
|
218 |
history = []
|
219 |
for i in range(num_iterations):
|
220 |
+
messages = (
|
221 |
+
messages_text_to_web(prompt)
|
222 |
if i == 0
|
223 |
+
else visual_feedback_messages(prompt, history)
|
224 |
)
|
225 |
content = ""
|
226 |
+
async for content in throttle(stream_claude(messages), every=0.25):
|
227 |
+
image_bytes, render_time = await render_html(
|
228 |
+
page, apply_template(content, template)
|
229 |
+
)
|
230 |
+
yield to_pil(image_bytes), content, render_time
|
231 |
+
history.append((content, image_bytes))
|
232 |
+
|
233 |
+
|
234 |
+
def to_image_bytes(image: Image.Image) -> bytes:
|
235 |
+
buffer = BytesIO()
|
236 |
+
image.save(buffer, format="PNG")
|
237 |
+
return buffer.getvalue()
|
238 |
+
|
239 |
+
|
240 |
+
async def match_image_with_visual_feedback(image, template, resolution, num_iterations):
|
241 |
+
width = {"512": 512, "1024": 1024}[resolution]
|
242 |
+
async with browser(width, width) as page:
|
243 |
+
history = []
|
244 |
+
for i in range(num_iterations):
|
245 |
+
image.thumbnail((width, width), Image.Resampling.LANCZOS)
|
246 |
+
messages = match_image_messages(to_image_bytes(image), history)
|
247 |
+
async for content in throttle(stream_claude(messages), 0.25):
|
248 |
+
image_bytes, render_time = await render_html(
|
249 |
+
page, apply_template(content, template)
|
250 |
+
)
|
251 |
+
yield to_pil(image_bytes), content, render_time
|
252 |
# always render the final image of each iteration
|
253 |
image_bytes, render_time = await render_html(
|
254 |
page, apply_template(content, template)
|
255 |
)
|
256 |
history.append((content, image_bytes))
|
|
|
|
|
|
|
257 |
|
258 |
|
259 |
+
demo_generate = gr.Interface(
|
260 |
generate_with_visual_feedback,
|
261 |
inputs=[
|
262 |
gr.Textbox(
|
|
|
276 |
],
|
277 |
)
|
278 |
|
279 |
+
demo_match_image = gr.Interface(
|
280 |
+
match_image_with_visual_feedback,
|
281 |
+
inputs=[
|
282 |
+
gr.Image(type="pil", label="Original Image", image_mode="RGB", format="png"),
|
283 |
+
gr.Dropdown(choices=["tailwind"], label="Template", value="tailwind"),
|
284 |
+
gr.Dropdown(choices=["512", "1024"], label="Page Width", value="512"),
|
285 |
+
gr.Slider(1, 10, 3, step=1, label="Iterations"),
|
286 |
+
],
|
287 |
+
outputs=[
|
288 |
+
gr.Image(type="pil", label="Rendered HTML", image_mode="RGB", format="png"),
|
289 |
+
gr.Textbox(lines=5, label="Code"),
|
290 |
+
gr.Number(label="Render Time", precision=2),
|
291 |
+
],
|
292 |
+
)
|
293 |
+
|
294 |
+
demo = gr.TabbedInterface(
|
295 |
+
[demo_match_image, demo_generate],
|
296 |
+
["Match Image", "Generate"],
|
297 |
+
)
|
298 |
+
|
299 |
|
300 |
if __name__ == "__main__":
|
301 |
prepare_playwright_if_needed()
|