Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -7,12 +7,12 @@ import torch
|
|
7 |
from gtts import gTTS
|
8 |
import io
|
9 |
import time
|
10 |
-
import asyncio
|
11 |
from streamlit.components.v1 import html
|
|
|
12 |
|
13 |
if not asyncio.get_event_loop().is_running():
|
14 |
asyncio.set_event_loop(asyncio.new_event_loop())
|
15 |
-
|
16 |
# Initialize session state
|
17 |
if 'processed_data' not in st.session_state:
|
18 |
st.session_state.processed_data = {
|
@@ -21,11 +21,34 @@ if 'processed_data' not in st.session_state:
|
|
21 |
'audio': None
|
22 |
}
|
23 |
|
24 |
-
if 'image_data' not in st.session_state:
|
25 |
-
st.session_state.image_data = None
|
26 |
|
27 |
-
|
28 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
29 |
|
30 |
# Page setup
|
31 |
st.set_page_config(page_title="Your Image to Audio Story", page_icon="🦜")
|
@@ -66,168 +89,88 @@ def text2audio(story_text):
|
|
66 |
audio_io.seek(0)
|
67 |
return {'audio': audio_io, 'sampling_rate': 16000}
|
68 |
|
69 |
-
#
|
70 |
-
|
71 |
-
|
72 |
-
|
73 |
-
|
74 |
-
|
75 |
-
|
76 |
-
# JavaScript timer implementation
|
77 |
-
def active_timer():
|
78 |
-
timer_html = """
|
79 |
-
<div style="font-size:16px;color:#666;margin-bottom:10px;">
|
80 |
-
<span id="timer-display">⏱️ Elapsed: 00:00</span>
|
81 |
-
</div>
|
82 |
-
|
83 |
-
<script>
|
84 |
-
// Reset any existing timer
|
85 |
-
if (window.timerInterval) {
|
86 |
-
clearInterval(window.timerInterval);
|
87 |
-
window.timerInterval = null;
|
88 |
-
}
|
89 |
|
90 |
-
|
91 |
-
|
92 |
-
window.timerInterval = setInterval(function() {
|
93 |
-
var now = new Date().getTime();
|
94 |
-
var elapsed = now - startTime;
|
95 |
-
|
96 |
-
var minutes = Math.floor(elapsed / (1000 * 60));
|
97 |
-
var seconds = Math.floor((elapsed % (1000 * 60)) / 1000);
|
98 |
-
|
99 |
-
var formattedTime =
|
100 |
-
(minutes < 10 ? "0" : "") + minutes + ":" +
|
101 |
-
(seconds < 10 ? "0" : "") + seconds;
|
102 |
-
|
103 |
-
var timerElement = document.getElementById("timer-display");
|
104 |
-
if (timerElement) {
|
105 |
-
timerElement.innerHTML = "⏱️ Elapsed: " + formattedTime;
|
106 |
-
}
|
107 |
-
}, 100);
|
108 |
-
</script>
|
109 |
-
"""
|
110 |
-
return timer_html
|
111 |
-
|
112 |
-
def frozen_timer(time_display):
|
113 |
-
timer_html = """
|
114 |
-
<div style="font-size:16px;color:#00cc00;font-weight:bold;margin-bottom:10px;">
|
115 |
-
<span id="timer-display">⏱️ Elapsed: """ + time_display + """ ✓</span>
|
116 |
-
</div>
|
117 |
|
118 |
-
<script>
|
119 |
-
// Clear any existing timer
|
120 |
-
if (window.timerInterval) {
|
121 |
-
clearInterval(window.timerInterval);
|
122 |
-
window.timerInterval = null;
|
123 |
-
}
|
124 |
-
</script>
|
125 |
-
"""
|
126 |
-
return timer_html
|
127 |
|
128 |
-
|
129 |
-
|
130 |
-
|
131 |
-
|
132 |
-
|
133 |
-
"""
|
134 |
-
return timer_html
|
135 |
|
136 |
-
|
137 |
-
|
138 |
-
timer_container.markdown(empty_timer(), unsafe_allow_html=True)
|
139 |
|
140 |
-
#
|
141 |
-
|
142 |
-
image_container.image(st.session_state.image_data, caption="Uploaded Image", use_container_width=True)
|
143 |
|
144 |
-
# UI components
|
145 |
-
uploaded_file = st.file_uploader("Select an Image After the Models are Loaded...")
|
146 |
-
|
147 |
-
# Process new uploaded file
|
148 |
-
if uploaded_file is not None:
|
149 |
-
# Save the image data to session state
|
150 |
-
bytes_data = uploaded_file.getvalue()
|
151 |
-
st.session_state.image_data = bytes_data
|
152 |
-
|
153 |
-
# Display the image
|
154 |
-
image_container.image(bytes_data, caption="Uploaded Image", use_container_width=True)
|
155 |
-
|
156 |
-
if st.session_state.get('current_file') != uploaded_file.name:
|
157 |
-
st.session_state.current_file = uploaded_file.name
|
158 |
-
|
159 |
-
# Start timer
|
160 |
-
st.session_state.timer_active = True
|
161 |
-
timer_container.markdown(active_timer(), unsafe_allow_html=True)
|
162 |
-
|
163 |
-
# Progress indicators
|
164 |
-
status_text = status_container.empty()
|
165 |
-
progress_bar = progress_container.progress(0)
|
166 |
-
|
167 |
-
try:
|
168 |
-
# Save uploaded file
|
169 |
-
with open(uploaded_file.name, "wb") as file:
|
170 |
-
file.write(bytes_data)
|
171 |
-
|
172 |
# Stage 1: Image to Text
|
173 |
status_text.markdown("**🖼️ Generating caption...**")
|
|
|
174 |
st.session_state.processed_data['scenario'] = img2text(uploaded_file.name)
|
175 |
progress_bar.progress(33)
|
176 |
-
|
177 |
# Stage 2: Text to Story
|
178 |
status_text.markdown("**📖 Generating story...**")
|
|
|
179 |
st.session_state.processed_data['story'] = text2story(
|
180 |
st.session_state.processed_data['scenario']
|
181 |
)
|
182 |
progress_bar.progress(66)
|
183 |
-
|
184 |
# Stage 3: Story to Audio
|
185 |
status_text.markdown("**🔊 Synthesizing audio...**")
|
|
|
186 |
st.session_state.processed_data['audio'] = text2audio(
|
187 |
st.session_state.processed_data['story']
|
188 |
)
|
189 |
progress_bar.progress(100)
|
190 |
-
|
191 |
-
# Final status
|
192 |
status_text.success("**✅ Generation complete!**")
|
193 |
-
|
194 |
-
# Show results
|
195 |
-
with results_container:
|
196 |
-
st.write("**Caption:**", st.session_state.processed_data['scenario'])
|
197 |
-
st.write("**Story:**", st.session_state.processed_data['story'])
|
198 |
-
|
199 |
-
except Exception as e:
|
200 |
-
status_text.error(f"**❌ Error:** {str(e)}")
|
201 |
-
progress_bar.empty()
|
202 |
-
raise e
|
203 |
-
|
204 |
-
# Display results if available
|
205 |
-
if st.session_state.processed_data.get('scenario'):
|
206 |
-
with results_container:
|
207 |
-
st.write("**Caption:**", st.session_state.processed_data['scenario'])
|
208 |
|
209 |
-
|
210 |
-
|
|
|
211 |
st.write("**Story:**", st.session_state.processed_data['story'])
|
212 |
|
213 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
214 |
if st.button("Play Audio of the Story Generated"):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
215 |
if st.session_state.processed_data.get('audio'):
|
216 |
-
# Make sure the image is still displayed
|
217 |
-
if st.session_state.image_data is not None:
|
218 |
-
image_container.image(st.session_state.image_data, caption="Uploaded Image", use_container_width=True)
|
219 |
-
|
220 |
-
# Freeze the timer (with a default value since we can't easily get the real time)
|
221 |
-
timer_container.markdown(frozen_timer("00:30"), unsafe_allow_html=True)
|
222 |
-
|
223 |
-
# Set timer as inactive since we've frozen it
|
224 |
-
st.session_state.timer_active = False
|
225 |
-
|
226 |
-
# Play the audio
|
227 |
audio_data = st.session_state.processed_data['audio']
|
228 |
-
st.audio(
|
229 |
-
audio_data['audio'].getvalue(),
|
230 |
-
format="audio/mp3"
|
231 |
-
)
|
232 |
else:
|
233 |
-
st.warning("Please generate a story first!")
|
|
|
7 |
from gtts import gTTS
|
8 |
import io
|
9 |
import time
|
|
|
10 |
from streamlit.components.v1 import html
|
11 |
+
import asyncio
|
12 |
|
13 |
if not asyncio.get_event_loop().is_running():
|
14 |
asyncio.set_event_loop(asyncio.new_event_loop())
|
15 |
+
|
16 |
# Initialize session state
|
17 |
if 'processed_data' not in st.session_state:
|
18 |
st.session_state.processed_data = {
|
|
|
21 |
'audio': None
|
22 |
}
|
23 |
|
|
|
|
|
24 |
|
25 |
+
# JavaScript timer component
|
26 |
+
def timer():
|
27 |
+
return """
|
28 |
+
<div id="timer" style="font-size:16px;color:#666;margin-bottom:10px;">⏱️ Elapsed: 00:00</div>
|
29 |
+
<script>
|
30 |
+
function updateTimer() {
|
31 |
+
var start = Date.now();
|
32 |
+
var timerElement = document.getElementById('timer');
|
33 |
+
|
34 |
+
var interval = setInterval(function() {
|
35 |
+
var elapsed = Date.now() - start;
|
36 |
+
var minutes = Math.floor(elapsed / 60000);
|
37 |
+
var seconds = Math.floor((elapsed % 60000) / 1000);
|
38 |
+
timerElement.innerHTML = '⏱️ Elapsed: ' +
|
39 |
+
(minutes < 10 ? '0' : '') + minutes + ':' +
|
40 |
+
(seconds < 10 ? '0' : '') + seconds;
|
41 |
+
}, 1000);
|
42 |
+
|
43 |
+
// Make the cleanup function available globally
|
44 |
+
window.stopTimer = function() {
|
45 |
+
clearInterval(interval);
|
46 |
+
};
|
47 |
+
}
|
48 |
+
updateTimer();
|
49 |
+
</script>
|
50 |
+
"""
|
51 |
+
|
52 |
|
53 |
# Page setup
|
54 |
st.set_page_config(page_title="Your Image to Audio Story", page_icon="🦜")
|
|
|
89 |
audio_io.seek(0)
|
90 |
return {'audio': audio_io, 'sampling_rate': 16000}
|
91 |
|
92 |
+
# UI components
|
93 |
+
uploaded_file = st.file_uploader("Select an Image After the Models are Loaded...")
|
94 |
+
|
95 |
+
if uploaded_file is not None:
|
96 |
+
# Initialize progress containers
|
97 |
+
status_text = st.empty()
|
98 |
+
progress_bar = st.progress(0)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
99 |
|
100 |
+
# Start JavaScript timer
|
101 |
+
html(timer(), height=50)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
102 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
103 |
|
104 |
+
try:
|
105 |
+
# Save uploaded file
|
106 |
+
bytes_data = uploaded_file.getvalue()
|
107 |
+
with open(uploaded_file.name, "wb") as file:
|
108 |
+
file.write(bytes_data)
|
|
|
|
|
109 |
|
110 |
+
if st.session_state.get('current_file') != uploaded_file.name:
|
111 |
+
st.session_state.current_file = uploaded_file.name
|
|
|
112 |
|
113 |
+
# Display image
|
114 |
+
st.image(uploaded_file, caption="Uploaded Image", use_container_width=True)
|
|
|
115 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
116 |
# Stage 1: Image to Text
|
117 |
status_text.markdown("**🖼️ Generating caption...**")
|
118 |
+
progress_bar.progress(0)
|
119 |
st.session_state.processed_data['scenario'] = img2text(uploaded_file.name)
|
120 |
progress_bar.progress(33)
|
121 |
+
|
122 |
# Stage 2: Text to Story
|
123 |
status_text.markdown("**📖 Generating story...**")
|
124 |
+
progress_bar.progress(33)
|
125 |
st.session_state.processed_data['story'] = text2story(
|
126 |
st.session_state.processed_data['scenario']
|
127 |
)
|
128 |
progress_bar.progress(66)
|
129 |
+
|
130 |
# Stage 3: Story to Audio
|
131 |
status_text.markdown("**🔊 Synthesizing audio...**")
|
132 |
+
progress_bar.progress(66)
|
133 |
st.session_state.processed_data['audio'] = text2audio(
|
134 |
st.session_state.processed_data['story']
|
135 |
)
|
136 |
progress_bar.progress(100)
|
137 |
+
|
138 |
+
# Final status and stop timer
|
139 |
status_text.success("**✅ Generation complete!**")
|
140 |
+
html("<script>document.getElementById('timer').style.color = '#00cc00';</script>")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
141 |
|
142 |
+
# Show results
|
143 |
+
# st.subheader("Results")
|
144 |
+
st.write("**Caption:**", st.session_state.processed_data['scenario'])
|
145 |
st.write("**Story:**", st.session_state.processed_data['story'])
|
146 |
|
147 |
+
except Exception as e:
|
148 |
+
html("<script>document.getElementById('timer').remove();</script>")
|
149 |
+
status_text.error(f"**❌ Error:** {str(e)}")
|
150 |
+
progress_bar.empty()
|
151 |
+
raise e
|
152 |
+
|
153 |
+
|
154 |
+
finally:
|
155 |
+
pass # Timer cleanup handled by JavaScript
|
156 |
+
|
157 |
+
|
158 |
+
# Audio playback
|
159 |
if st.button("Play Audio of the Story Generated"):
|
160 |
+
# Stop the timer
|
161 |
+
st.markdown(
|
162 |
+
"""
|
163 |
+
<script>
|
164 |
+
if (window.stopTimer) {
|
165 |
+
window.stopTimer();
|
166 |
+
}
|
167 |
+
</script>
|
168 |
+
""",
|
169 |
+
unsafe_allow_html=True
|
170 |
+
)
|
171 |
+
|
172 |
if st.session_state.processed_data.get('audio'):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
173 |
audio_data = st.session_state.processed_data['audio']
|
174 |
+
st.audio(audio_data['audio'].getvalue(), format="audio/mp3")
|
|
|
|
|
|
|
175 |
else:
|
176 |
+
st.warning("Please generate a story first!")
|