Update app.py
Browse files
app.py
CHANGED
|
@@ -9,11 +9,20 @@ from together import Together
|
|
| 9 |
from email.mime.multipart import MIMEMultipart
|
| 10 |
from email.mime.text import MIMEText
|
| 11 |
from email.mime.audio import MIMEAudio
|
|
|
|
|
|
|
|
|
|
|
|
|
| 12 |
|
| 13 |
# Image-to-text function
|
| 14 |
def img2txt(url: str) -> str:
|
|
|
|
| 15 |
captioning_model = pipeline("image-to-text", model="Salesforce/blip-image-captioning-base")
|
|
|
|
|
|
|
| 16 |
text = captioning_model(url, max_new_tokens=20)[0]["generated_text"]
|
|
|
|
|
|
|
| 17 |
return text
|
| 18 |
|
| 19 |
# Text-to-story generation function
|
|
@@ -43,6 +52,7 @@ def txt2story(prompt: str, top_k: int, top_p: float, temperature: float) -> str:
|
|
| 43 |
|
| 44 |
# Text-to-speech function
|
| 45 |
def txt2speech(text: str) -> None:
|
|
|
|
| 46 |
tts = gTTS(text=text, lang='en')
|
| 47 |
tts.save("audio_story.mp3")
|
| 48 |
|
|
@@ -61,6 +71,47 @@ def get_user_preferences() -> Dict[str, str]:
|
|
| 61 |
}
|
| 62 |
return preferences
|
| 63 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 64 |
# Main Streamlit application
|
| 65 |
def main():
|
| 66 |
st.set_page_config(
|
|
@@ -92,6 +143,10 @@ def main():
|
|
| 92 |
|
| 93 |
with col2:
|
| 94 |
if uploaded_file is not None:
|
|
|
|
|
|
|
|
|
|
|
|
|
| 95 |
# Display uploaded image
|
| 96 |
st.markdown("## πΌοΈ Your Image")
|
| 97 |
bytes_data = uploaded_file.read()
|
|
@@ -99,43 +154,53 @@ def main():
|
|
| 99 |
file.write(bytes_data)
|
| 100 |
st.image(uploaded_file, use_column_width=True)
|
| 101 |
|
| 102 |
-
#
|
| 103 |
-
|
| 104 |
try:
|
| 105 |
-
|
| 106 |
-
|
| 107 |
-
|
| 108 |
-
st.session_state.caption = scenario # Store caption in session state
|
| 109 |
-
|
| 110 |
-
# Step 2: Generate Story
|
| 111 |
-
with st.spinner("π Generating Story..."):
|
| 112 |
-
prompt = f"""Based on the image description: '{scenario}',
|
| 113 |
-
create a {preferences['genre']} story set in {preferences['setting']}
|
| 114 |
-
in {preferences['continent']}. The story should have a {preferences['tone']}
|
| 115 |
-
tone and explore the theme of {preferences['theme']}. The main conflict
|
| 116 |
-
should be {preferences['conflict']}. The story should have a {preferences['twist']}
|
| 117 |
-
and end with a {preferences['ending']} ending."""
|
| 118 |
-
|
| 119 |
-
story = txt2story(prompt, top_k=5, top_p=0.8, temperature=1.5)
|
| 120 |
-
st.session_state.story = story # Store story in session state
|
| 121 |
-
|
| 122 |
-
# Step 3: Generate Audio Version
|
| 123 |
-
with st.spinner("π§ Generating Audio Version..."):
|
| 124 |
-
txt2speech(story)
|
| 125 |
-
st.session_state.audio_file_path = "audio_story.mp3" # Store audio path in session state
|
| 126 |
-
|
| 127 |
except Exception as e:
|
| 128 |
-
st.error(f"An error occurred: {str(e)}")
|
| 129 |
st.warning("Please try again or contact support if the problem persists.")
|
| 130 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 131 |
# Display results if story exists in session state
|
| 132 |
if st.session_state.story:
|
| 133 |
st.markdown("---")
|
| 134 |
|
| 135 |
-
# Image caption
|
| 136 |
-
with st.expander("π Image Caption", expanded=True):
|
| 137 |
-
st.write(st.session_state.caption)
|
| 138 |
-
|
| 139 |
# Story text
|
| 140 |
with st.expander("π Generated Story", expanded=True):
|
| 141 |
st.write(st.session_state.story)
|
|
@@ -144,5 +209,25 @@ def main():
|
|
| 144 |
with st.expander("π§ Audio Version", expanded=True):
|
| 145 |
st.audio(st.session_state.audio_file_path)
|
| 146 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 147 |
if __name__ == '__main__':
|
| 148 |
main()
|
|
|
|
| 9 |
from email.mime.multipart import MIMEMultipart
|
| 10 |
from email.mime.text import MIMEText
|
| 11 |
from email.mime.audio import MIMEAudio
|
| 12 |
+
#from dotenv import load_dotenv
|
| 13 |
+
|
| 14 |
+
# Load environment variables
|
| 15 |
+
#load_dotenv()
|
| 16 |
|
| 17 |
# Image-to-text function
|
| 18 |
def img2txt(url: str) -> str:
|
| 19 |
+
print("Initializing captioning model...")
|
| 20 |
captioning_model = pipeline("image-to-text", model="Salesforce/blip-image-captioning-base")
|
| 21 |
+
|
| 22 |
+
print("Generating text from the image...")
|
| 23 |
text = captioning_model(url, max_new_tokens=20)[0]["generated_text"]
|
| 24 |
+
|
| 25 |
+
print(text)
|
| 26 |
return text
|
| 27 |
|
| 28 |
# Text-to-story generation function
|
|
|
|
| 52 |
|
| 53 |
# Text-to-speech function
|
| 54 |
def txt2speech(text: str) -> None:
|
| 55 |
+
print("Converting text to speech using gTTS...")
|
| 56 |
tts = gTTS(text=text, lang='en')
|
| 57 |
tts.save("audio_story.mp3")
|
| 58 |
|
|
|
|
| 71 |
}
|
| 72 |
return preferences
|
| 73 |
|
| 74 |
+
# Function to send the story via email using smtplib
|
| 75 |
+
def send_story_email(recipient_email: str, story_text: str, audio_file_path: str) -> bool:
|
| 76 |
+
try:
|
| 77 |
+
# Email configuration
|
| 78 |
+
smtp_server = os.environ.get("SMTP_SERVER") # e.g., "smtp.gmail.com"
|
| 79 |
+
smtp_port = int(os.environ.get("SMTP_PORT", 587)) # e.g., 587 for TLS
|
| 80 |
+
sender_email = os.environ.get("SENDER_EMAIL") # Your email
|
| 81 |
+
sender_password = os.environ.get("SENDER_PASSWORD") # Your email password
|
| 82 |
+
|
| 83 |
+
# Create the email
|
| 84 |
+
msg = MIMEMultipart()
|
| 85 |
+
msg['From'] = sender_email
|
| 86 |
+
msg['To'] = recipient_email
|
| 87 |
+
msg['Subject'] = "Your Generated Story"
|
| 88 |
+
|
| 89 |
+
# Attach the story text
|
| 90 |
+
msg.attach(MIMEText(f"Here's your generated story:\n\n{story_text}\n\nEnjoy!", 'plain'))
|
| 91 |
+
|
| 92 |
+
# Attach the audio file
|
| 93 |
+
with open(audio_file_path, 'rb') as audio_file:
|
| 94 |
+
audio_part = MIMEAudio(audio_file.read(), _subtype='mp3')
|
| 95 |
+
audio_part.add_header('Content-Disposition', 'attachment', filename=os.path.basename(audio_file_path))
|
| 96 |
+
msg.attach(audio_part)
|
| 97 |
+
|
| 98 |
+
# Send the email
|
| 99 |
+
with smtplib.SMTP(smtp_server, smtp_port) as server:
|
| 100 |
+
server.starttls() # Upgrade to a secure connection
|
| 101 |
+
server.login(sender_email, sender_password)
|
| 102 |
+
server.send_message(msg)
|
| 103 |
+
|
| 104 |
+
return True
|
| 105 |
+
|
| 106 |
+
except Exception as e:
|
| 107 |
+
print(f"Error sending email: {str(e)}")
|
| 108 |
+
return False
|
| 109 |
+
|
| 110 |
+
# Basic email validation function
|
| 111 |
+
def validate_email(email: str) -> bool:
|
| 112 |
+
pattern = r'^[\w\.-]+@[\w\.-]+\.\w+$'
|
| 113 |
+
return re.match(pattern, email) is not None
|
| 114 |
+
|
| 115 |
# Main Streamlit application
|
| 116 |
def main():
|
| 117 |
st.set_page_config(
|
|
|
|
| 143 |
|
| 144 |
with col2:
|
| 145 |
if uploaded_file is not None:
|
| 146 |
+
# Reset session state variables when a new image is uploaded
|
| 147 |
+
st.session_state.story = ""
|
| 148 |
+
st.session_state.audio_file_path = ""
|
| 149 |
+
|
| 150 |
# Display uploaded image
|
| 151 |
st.markdown("## πΌοΈ Your Image")
|
| 152 |
bytes_data = uploaded_file.read()
|
|
|
|
| 154 |
file.write(bytes_data)
|
| 155 |
st.image(uploaded_file, use_column_width=True)
|
| 156 |
|
| 157 |
+
# Generate image caption as soon as image is uploaded
|
| 158 |
+
with st.spinner("π Generating image caption..."):
|
| 159 |
try:
|
| 160 |
+
scenario = img2txt("uploaded_image.jpg")
|
| 161 |
+
st.session_state.caption = scenario # Store caption in session state
|
| 162 |
+
st.success("Image caption generated.")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 163 |
except Exception as e:
|
| 164 |
+
st.error(f"An error occurred while generating image caption: {str(e)}")
|
| 165 |
st.warning("Please try again or contact support if the problem persists.")
|
| 166 |
|
| 167 |
+
# Display image caption
|
| 168 |
+
if st.session_state.caption:
|
| 169 |
+
st.markdown("## π Image Caption")
|
| 170 |
+
st.write(st.session_state.caption)
|
| 171 |
+
|
| 172 |
+
# Generate Story button
|
| 173 |
+
if st.session_state.caption:
|
| 174 |
+
if st.button("π¨ Generate Story"):
|
| 175 |
+
with st.spinner("π Generating story..."):
|
| 176 |
+
try:
|
| 177 |
+
# Create story prompt using st.session_state.caption
|
| 178 |
+
prompt = f"""Based on the image description: '{st.session_state.caption}',
|
| 179 |
+
create a {preferences['genre']} story set in {preferences['setting']}
|
| 180 |
+
in {preferences['continent']}. The story should have a {preferences['tone']}
|
| 181 |
+
tone and explore the theme of {preferences['theme']}. The main conflict
|
| 182 |
+
should be {preferences['conflict']}. The story should have a {preferences['twist']}
|
| 183 |
+
and end with a {preferences['ending']} ending."""
|
| 184 |
+
|
| 185 |
+
# Generate story
|
| 186 |
+
story = txt2story(prompt, top_k=5, top_p=0.8, temperature=1.5)
|
| 187 |
+
st.session_state.story = story # Store story in session state
|
| 188 |
+
st.success("Story generated.")
|
| 189 |
+
|
| 190 |
+
# Convert to audio
|
| 191 |
+
with st.spinner("π Generating audio story..."):
|
| 192 |
+
txt2speech(story)
|
| 193 |
+
st.session_state.audio_file_path = "audio_story.mp3" # Store audio path in session state
|
| 194 |
+
st.success("Audio story generated.")
|
| 195 |
+
|
| 196 |
+
except Exception as e:
|
| 197 |
+
st.error(f"An error occurred: {str(e)}")
|
| 198 |
+
st.warning("Please try again or contact support if the problem persists.")
|
| 199 |
+
|
| 200 |
# Display results if story exists in session state
|
| 201 |
if st.session_state.story:
|
| 202 |
st.markdown("---")
|
| 203 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 204 |
# Story text
|
| 205 |
with st.expander("π Generated Story", expanded=True):
|
| 206 |
st.write(st.session_state.story)
|
|
|
|
| 209 |
with st.expander("π§ Audio Version", expanded=True):
|
| 210 |
st.audio(st.session_state.audio_file_path)
|
| 211 |
|
| 212 |
+
# Email section
|
| 213 |
+
st.markdown("---")
|
| 214 |
+
st.markdown("## π§ Get Story via Email")
|
| 215 |
+
email = st.text_input(
|
| 216 |
+
"Enter your email address:",
|
| 217 |
+
help="We'll send you the story text and audio file"
|
| 218 |
+
)
|
| 219 |
+
|
| 220 |
+
if st.button("π€ Send to Email"):
|
| 221 |
+
if not email:
|
| 222 |
+
st.warning("Please enter an email address.")
|
| 223 |
+
elif not validate_email(email):
|
| 224 |
+
st.error("Please enter a valid email address.")
|
| 225 |
+
else:
|
| 226 |
+
with st.spinner("π¨ Sending email..."):
|
| 227 |
+
if send_story_email(email, st.session_state.story, st.session_state.audio_file_path):
|
| 228 |
+
st.success(f"Email sent to: {email}")
|
| 229 |
+
else:
|
| 230 |
+
st.error("β Failed to send email. Please try again.")
|
| 231 |
+
|
| 232 |
if __name__ == '__main__':
|
| 233 |
main()
|