Audio Feature

.streamlit/config.toml

@@ -0,0 +1,6 @@
+[theme]
+primaryColor="#FF4B4B"
+backgroundColor="#0E1117"
+secondaryBackgroundColor="#262730"
+textColor="#FAFAFA"
+font="sans serif"

app.py

@@ -1,32 +1,28 @@
+import os
+import pyperclip
 import streamlit as st
-from langchain.prompts import PromptTemplate
+from dotenv import load_dotenv
+import speech_recognition as sr
 import google.generativeai as genai
+from langchain.prompts import PromptTemplate
 from langchain_google_genai import ChatGoogleGenerativeAI
-import pyperclip
-import os
-from dotenv import load_dotenv
+
 load_dotenv()
-os.environ['GOOGLE_API_KEY']='AIzaSyAA59xkxMEUqffuaNMzrTxK6m5AJpY3tCw'
 
 def model():
     genai.configure(api_key=os.getenv("GOOGLE_API_KEY"))
     return ChatGoogleGenerativeAI(model="gemini-pro")
 
 def getLLamaresponse(input_text,no_words,blog_style):
-    
     ## Prompt Template
-
     template="""
-        Write a {tone} blog for {blog_style} job profile for a topic {input_text}
-        within {no_words} words and ends with a conclusion. {info}.
+        Write a {tone} blog in {lang} as {blog_style} for a topic {input_text} within {no_words} words with references and video links also if possible and ends with a conclusion. {info}
             """
-    
-    prompt=PromptTemplate(input_variables=['tone',"blog_style","input_text",'no_words', "info"],
+    prompt=PromptTemplate(input_variables=['tone','lang',"blog_style","input_text",'no_words', "info"],
                           template=template)
-    
+    formated_prompt = prompt.format(tone=tone,lang=lang,blog_style=blog_style,input_text=input_text,no_words=no_words,info=info)
     ## Generate the ressponse from the LLama 2 model
-    response=llm.invoke(prompt.format(tone=tone,blog_style=blog_style,input_text=input_text,no_words=no_words,info=info))
-    print(response)
+    response=llm.invoke(formated_prompt)
     return response.content
 
 def copy_text(answer, copy_button=False):
@@ -34,6 +30,31 @@ def copy_text(answer, copy_button=False):
     if copy_button:
         st.toast("Text copied to clipboard!", icon="📋")
 
+def record_audio():
+    r = sr.Recognizer()
+    with st.spinner("Recording..."):
+        with sr.Microphone() as source:
+            r.adjust_for_ambient_noise(source)  # Adjust for ambient noise
+            with st.spinner("Say Something..."):
+                audio = r.listen(source, timeout=5)  # Set a timeout for listening
+        with st.spinner("Processing..."):
+            # Attempt speech recognition
+            try:
+                text = r.recognize_google(audio)
+                st.session_state['input_text'] = text
+                return text
+            except sr.UnknownValueError:
+                st.write("Sorry, I could not understand what you said. Please try again or write in text box.")
+                return ""
+            except sr.RequestError as e:
+                st.write(f"Could not request results; {e}")
+                return ""
+
+def input_state(input_text):
+    if isinstance(input_text, str):
+        st.session_state['input_text'] = input_text  # Update st.session_state with new text input
+
+
 
 st.set_page_config(page_title="AI-Blogger",
                     page_icon='🤖',
@@ -42,39 +63,41 @@ st.set_page_config(page_title="AI-Blogger",
 st.header("AI Blogger 🤖", divider='rainbow')
 
 
-input_text=st.text_input("Enter the Blog Topic")
+col1,col2=st.columns([20,3])
+with col1:
+    st.write('')
+    text = st.empty()
+    input_text = text.text_input("Enter the Blog Topic")
+    
+
+with col2:
+    st.write(':grey[.]')
+    recorder = st.button("🎙️", help="Blog Topic by Voice")
+
+if recorder:
+    recorded_text = record_audio()
+    if recorded_text:
+        input_text = text.text_input("Enter the Blog Topic",value=recorded_text)
 
 with st.sidebar:
     st.title(' :blue[_AI Generated Blog_] 🤖')
-
     # Refactored from <https://github.com/a16z-infra/llama2-chatbot>
     st.subheader('Parameters')
-    
     no_words = st.sidebar.slider('Maximum Characters', min_value=10, max_value=5000, value=1000, step=100)
-    blog_style=st.selectbox('Writing the blog for',
-                            ('Researchers','Data Scientist','Common People'),index=0)
-    tone=st.selectbox('Desired Tone',
-                            ('Informative','Casual','Persuasive', 'Formal', 'Humorous'),index=0)
-
+    blog_style=st.selectbox('Writing the blog for', ('Researchers','Data Scientist','Common People'),index=0)
+    tone=st.selectbox('Desired Tone', ('Informative','Casual','Persuasive', 'Formal', 'Humorous'),index=0)
+    lang = st.text_input('Language', 'English')
     info = st.text_input('Specific Instruction')
     with st.spinner("Loading Model..."):
         llm= model()
-        
-    
-submit=st.button("Generate Blog")
 
-
-## Final response
+submit=st.button("Generate Blog", on_click =lambda: input_state(input_text))
 if submit:
     with st.spinner("Generating Blog..."):
-        answer = getLLamaresponse(input_text,no_words,blog_style)
+        answer = getLLamaresponse(st.session_state['input_text'],no_words,blog_style)
         st.success('Blog Generated!', icon="✅")
         st.write(answer)
+        st.markdown(''':orange[Run in your system to access Copy to Clipboard Feature]''')
 
         ## Uncomment if using in system HF doesn't Support Copy to clipboard
         # st.button("📋", on_click=copy_text(answer))
-        st.markdown('''
-
-
-                :orange[Run in your system to access Copy to Clipboard Feature]
-        ''')

requirements.txt

@@ -4,4 +4,6 @@ langchain
 streamlit
 pyperclip
 huggingface_hub
-python-dotenv
+python-dotenv
+SpeechRecognition
+PyAudio