Add application file

app.py

@@ -0,0 +1,61 @@
+import streamlit as st
+from time import sleep
+import os
+from templates import *
+   
+
+# Page title
+title = '🦜🔗 DemoGPT'
+
+st.set_page_config(page_title=title)
+st.title(title)
+st.markdown(
+    """
+    This's just to showcase the capabilities of DemoGPT.
+
+    For custom applications, please open in [![Open in GitHub](https://img.shields.io/badge/GitHub-100000?style=for-the-badge&logo=github&logoColor=white)](https://github.com/melih-unsal/DemoGPT)
+    """
+)
+# Text input    
+
+openai_api_key = st.text_input('Enter your OpenAI API Key', placeholder='sk-...',type="password")
+demo_title = st.text_input('Enter your demo title', placeholder='Type your demo title')
+
+st.write("Examples")
+
+cols1 = st.columns([1,1,1.2])
+cols2 = st.columns([1.6,1.5,1])
+
+pid = None
+
+pressed = False
+
+if 'current' not in st.session_state:
+    st.session_state['current'] = ''
+    st.session_state['done'] = None 
+elif st.session_state['done']:
+    st.session_state['done'].empty()
+
+for col,example in zip(cols1,examples1):
+    if col.button(example):
+        st.session_state['current'] = example
+        pressed = True
+
+for col,example in zip(cols2,examples2):
+    if col.button(example):
+        st.session_state['current'] = example
+        pressed = True
+
+st.markdown('----')
+if st.session_state['current']:
+    with st.container():
+        if not openai_api_key:
+            st.warning('Please enter your OpenAI API Key', icon="⚠️")
+        else:
+            if pressed:
+                wait()
+                st.session_state['done'] = st.success('Done!')
+            example2pages[st.session_state['current']](openai_api_key,demo_title)
+st.markdown('----')
+REPO_URL = "https://github.com/melih-unsal/DemoGPT"
+st.markdown(f"project [repo on github]({REPO_URL}) waiting for your :star:")

requirements.txt

@@ -0,0 +1,4 @@
+streamlit
+langchain
+openai
+tiktoken

templates.py

@@ -0,0 +1,253 @@
+def wait():
+    import streamlit as st
+    import time
+
+    progress_texts = ["Generating Code...:pencil:","Creating App...:running:","Rendering the demo page...:tv:"]
+    num_of_texts = len(progress_texts)
+    progress_texts_iter =  iter(progress_texts)
+    my_bar = st.progress(0, "Initializing...")
+    with st.spinner('Processing...'):
+        start = end = 0
+        for i in range(num_of_texts):
+            text = next(progress_texts_iter)
+            start = end
+            end = start + 100 // num_of_texts
+            for percent_complete in range(start, end):
+                time.sleep(0.03*(num_of_texts-i))
+                my_bar.progress(percent_complete + 1, text=text)
+    my_bar.empty()
+   
+def language_translator(openai_api_key,demo_title="My Lang App"):
+    import streamlit as st
+    from langchain import LLMChain
+    from langchain.chat_models import ChatOpenAI
+    from langchain.prompts.chat import (
+        ChatPromptTemplate,
+        SystemMessagePromptTemplate,
+        HumanMessagePromptTemplate,
+    )
+
+    def language_translator(input_language, output_language, text):
+        chat = ChatOpenAI(openai_api_key=openai_api_key, temperature=0)
+
+        template = "You are a helpful assistant that translates {input_language} to {output_language}. Please provide the text to translate."
+        system_message_prompt = SystemMessagePromptTemplate.from_template(template)
+        human_template = "{text}"
+        human_message_prompt = HumanMessagePromptTemplate.from_template(human_template)
+        chat_prompt = ChatPromptTemplate.from_messages([system_message_prompt, human_message_prompt])
+
+        chain = LLMChain(llm=chat, prompt=chat_prompt)
+        result = chain.run(input_language=input_language, output_language=output_language, text=text)
+        return result
+
+    st.header(demo_title)
+    
+    input_language = st.text_input("Input Language")
+    output_language = st.text_input("Output Language")
+    text = st.text_area("Text")
+
+    if st.button("Translate"):
+        result = language_translator(input_language, output_language, text)
+        st.write(result)
+        st.balloons()
+
+def blog_post_generator(openai_api_key,demo_title="My Blogger"):
+    import streamlit as st
+    from langchain import LLMChain
+    from langchain.chat_models import ChatOpenAI
+    from langchain.prompts.chat import (
+        ChatPromptTemplate,
+        SystemMessagePromptTemplate,
+        HumanMessagePromptTemplate,
+    )
+
+    def generate_blog_post(title):
+        print("Generating blog post")
+        chat = ChatOpenAI(openai_api_key=openai_api_key, temperature=0)
+
+        template = "You are a helpful assistant that generates a blog post from the title: {title}. Please provide some content."
+        system_message_prompt = SystemMessagePromptTemplate.from_template(template)
+        human_template = "{text}"
+        human_message_prompt = HumanMessagePromptTemplate.from_template(human_template)
+        chat_prompt = ChatPromptTemplate.from_messages([system_message_prompt, human_message_prompt])
+
+        chain = LLMChain(llm=chat, prompt=chat_prompt)
+        result = chain.run(title=title, text="")
+        return result
+
+    st.header(demo_title)
+
+    title = st.text_input("Enter the title of your blog post")
+    if st.button("Generate Blog Post"):
+        print("Generate")
+        with st.spinner("Generating the blog post..."):
+            result = generate_blog_post(title)
+            st.write(result)
+            st.balloons()
+
+def grammer_corrector(openai_api_key,demo_title="My Grammerly"):
+    import streamlit as st
+    from langchain import LLMChain
+    from langchain.chat_models import ChatOpenAI
+    from langchain.prompts.chat import (
+        ChatPromptTemplate,
+        SystemMessagePromptTemplate,
+        HumanMessagePromptTemplate,
+    )
+
+    def correct_grammar(text):
+        chat = ChatOpenAI(openai_api_key=openai_api_key, temperature=0)
+
+        template = "You are a helpful assistant that corrects grammar. Please provide the text you want to correct."
+        system_message_prompt = SystemMessagePromptTemplate.from_template(template)
+        human_template = "{text}"
+        human_message_prompt = HumanMessagePromptTemplate.from_template(human_template)
+        chat_prompt = ChatPromptTemplate.from_messages([system_message_prompt, human_message_prompt])
+
+        chain = LLMChain(llm=chat, prompt=chat_prompt)
+        result = chain.run(text=text)
+        return result
+
+    st.header(demo_title)
+
+    text = st.text_input("Enter the text you want to correct")
+    if st.button("Correct Grammar"):
+        result = correct_grammar(text)
+        st.write(result)
+        st.balloons()
+
+def lyrics_generator(openai_api_key,demo_title="Lyrics Maker"):
+    import streamlit as st
+    from langchain import LLMChain
+    from langchain.chat_models import ChatOpenAI
+    from langchain.prompts.chat import (
+        ChatPromptTemplate,
+        SystemMessagePromptTemplate,
+        HumanMessagePromptTemplate,
+    )
+
+    def generate_song(title):
+        chat = ChatOpenAI(openai_api_key=openai_api_key, temperature=0)
+
+        template = "You are a helpful assistant that generates a song from the title: {title}. Please provide some lyrics."
+        system_message_prompt = SystemMessagePromptTemplate.from_template(template)
+        human_template = "{text}"
+        human_message_prompt = HumanMessagePromptTemplate.from_template(human_template)
+        chat_prompt = ChatPromptTemplate.from_messages([system_message_prompt, human_message_prompt])
+
+        chain = LLMChain(llm=chat, prompt=chat_prompt)
+        result = chain.run(title=title, text="")
+        return result
+
+    st.header(demo_title)
+
+    title = st.text_input("Enter the song title:")
+    if st.button("Generate Song"):
+        with st.spinner("Generating song..."):
+            result = generate_song(title)
+            st.write(result)
+            st.balloons()
+
+def twit_generator(openai_api_key,demo_title="My AutoTwitter"):
+    import streamlit as st
+    from langchain import LLMChain
+    from langchain.chat_models import ChatOpenAI
+    from langchain.prompts.chat import (
+        ChatPromptTemplate,
+        SystemMessagePromptTemplate,
+        HumanMessagePromptTemplate,
+    )
+
+    def twitter(hashtag):
+        chat = ChatOpenAI(openai_api_key=openai_api_key, temperature=0.1)
+
+        template = "You are a helpful assistant that generate twit from {hashtag}. Please provide the hashtag to generate a twit."
+        system_message_prompt = SystemMessagePromptTemplate.from_template(template)
+        human_template = "Only generate the corresponding twit for this hashtag {hashtag}"
+        human_message_prompt = HumanMessagePromptTemplate.from_template(human_template)
+        chat_prompt = ChatPromptTemplate.from_messages([system_message_prompt, human_message_prompt])
+
+        chain = LLMChain(llm=chat, prompt=chat_prompt)
+        result = chain.run(hashtag=hashtag)
+        return result
+
+    st.header(demo_title)
+    
+    hashtag = st.text_input("Hashtag",placeholder="#")
+
+    if st.button("Generate"):
+        result = twitter(hashtag)
+        st.write(result)
+        st.balloons()
+
+def email_generator(openai_api_key,demo_title="My AutoTwitter"):
+    import streamlit as st
+    from langchain import LLMChain
+    from langchain.chat_models import ChatOpenAI
+    from langchain.prompts.chat import (
+        ChatPromptTemplate,
+        SystemMessagePromptTemplate,
+        HumanMessagePromptTemplate,
+    )
+
+    def email(sender_name,receiver_name,purpose,keywords,tone):
+        chat = ChatOpenAI(openai_api_key=openai_api_key, temperature=0.1)
+
+        template = "You are a helpful assistant that generate email to a person according to the given purpose, keywords and tone."
+        system_message_prompt = SystemMessagePromptTemplate.from_template(template)
+        human_template = """Generate email for a person according to the given purpose, keywords and tone.
+        Sender Name:{sender_name}
+        Receiver Name:{receiver_name}
+        Purpose:{purpose}
+        Keywords:{keywords}
+        Tone:{tone}
+        Directly start to type an email
+        """
+        human_message_prompt = HumanMessagePromptTemplate.from_template(human_template)
+        chat_prompt = ChatPromptTemplate.from_messages([system_message_prompt, human_message_prompt])
+
+        chain = LLMChain(llm=chat, prompt=chat_prompt)
+        result = chain.run(sender_name=sender_name, receiver_name=receiver_name, purpose=purpose, keywords=keywords, tone=tone)
+        return result
+
+    st.header(demo_title)
+    
+    sender_name = st.text_input("Name of the sender")
+    receiver_name = st.text_input("Receiver of the sender")
+    purpose = st.text_input("Purpose of email")
+    keywords = st.text_input("Primary keywords",placeholder="comma separated list of keywords")
+    tone = st.text_input("Tone of the email")
+
+    if st.button("Generate"):
+        with st.spinner("Generating email..."):
+            result = email(sender_name,receiver_name,purpose,keywords,tone)
+            st.write(result)
+            st.balloons()
+
+examples1 = [
+    "Language Translator 📝",
+    "Grammer Corrector 🛠",
+    "Blog post generator from title 📔"
+    ]
+
+examples2=[
+    "Lyrics generator from song title 🎤",
+    "Twit generation from hashtag 🐦",
+    'Email generator :email:'
+    ] 
+
+examples = examples1 + examples2
+
+pages1 = [language_translator,grammer_corrector,blog_post_generator]
+pages2=[lyrics_generator,twit_generator,email_generator]
+
+pages = pages1 + pages2
+
+example2pages={
+    example:page
+    for example,page in zip(examples,pages)
+}
+
+
+__all__ = ['language_translator','grammer_corrector','blog_post_generator','lyrics_generator','twit_generator',
+           'example2pages', 'examples', 'examples1', 'examples2', 'wait']