feat: resumefy - [Shubham, Nitin & Megha]

community/resumefy/README.MD

@@ -0,0 +1,129 @@
+# 🎯 Resumefy: all judgy and grumpy!
+
+## Overview
+
+Welcome to the **Resumefy** app! This project is a powerful, AI-driven application that uses **TuneStudio's** language model API to review resumes, provide feedback, and answer follow-up questions. By uploading your resume in PDF format, the app will:
+1. Extract and parse the resume text.
+2. Send the parsed content to TuneStudio's AI model for analysis and feedback.
+3. Display feedback that includes recommendations for improvement.
+4. Offer prefilled prompts to ask additional questions or allow custom follow-ups to improve your resume further.
+
+This solution is ideal for job seekers, professionals looking to fine-tune their resumes, or anyone interested in receiving tailored feedback from a state-of-the-art AI model.
+
+---
+
+## 🛠️ Key Features
+
+- **Easy PDF Upload**: Users can simply upload their resume in PDF format to start the review process.
+- **AI-Powered Feedback**: Leverages the TuneStudio API to analyze and provide constructive feedback on resume structure, content, and formatting.
+- **Follow-up Interaction**: Offers both prefilled suggestions and the ability to ask custom follow-up questions about the resume.
+- **Streamlit Interface**: A simple, intuitive, and responsive user interface for easy interaction with the AI.
+
+---
+
+## 📷 Screenshots
+
+### 1. Resume Upload Process
+![Uploading Resume](./1.png)
+> **Step 1**: Upload your resume as a PDF file to begin the review process.
+
+---
+
+### 2. Feedback from AI
+![Feedback from AI](./2.png)
+> **Step 2**: The AI reviews your resume and provides detailed feedback. This includes suggestions on how to improve various aspects of your resume, such as formatting, skills presentation, and overall content.
+
+---
+
+### 3. Follow-up Recommendations
+![Follow-up Recommendations](./3.png)
+> **Step 3**: You can engage with the AI further by asking follow-up questions. We offer prefilled suggestions, or you can submit your custom questions for a more personalized experience.
+
+---
+
+## 🚀 How It Works
+
+1. **Upload the Resume**: Users upload their resume in PDF format via the Streamlit app.
+2. **PDF Parsing**: The app extracts the text from the uploaded resume using `PyPDF2`.
+3. **TuneStudio API Integration**: The parsed resume text is sent to TuneStudio's model (`lamina/Test`), which reviews the content and returns feedback.
+4. **Receive Feedback**: The feedback is displayed in the app. This feedback includes suggestions for improving your resume’s format, structure, and content.
+5. **Follow-up Interaction**: After receiving the initial feedback, users can select from prefilled follow-up prompts or ask custom questions to get deeper insights into their resume.
+
+---
+
+## 💡 Use Cases
+
+- **Job Seekers**: Get actionable insights to fine-tune your resume and increase your chances of standing out in a competitive job market.
+- **Professionals**: Looking to advance in your career? Use the AI feedback to polish your resume and showcase your skills better.
+- **Career Coaches**: Automate part of your resume review process by integrating this tool to get immediate feedback for your clients.
+- **Recruiters**: Use this tool to quickly evaluate and provide feedback on potential candidates' resumes.
+
+---
+
+## 🔧 Installation & Setup
+
+1. Clone this repository:
+   ```bash
+   git clone https://github.com/freakpirate/cookbook.git
+   ```
+2. Install the required dependencies:
+   ```bash
+   pip install streamlit requests PyPDF2
+   ```
+3. Add your TuneStudio API key:
+   - Replace `YOUR_TUNESTUDIO_API_KEY` in the `resume_reviewer.py` file with your actual TuneStudio API key.
+
+4. Run the Streamlit app:
+   ```bash
+   streamlit run resume_reviewer.py
+   ```
+
+5. Open your browser and go to the local URL provided by Streamlit to start interacting with the app.
+
+---
+
+## 📝 Example Flow
+
+1. **Upload your resume**: Start by uploading your resume as a PDF file.
+2. **Get initial feedback**: Once your resume is uploaded, the AI reviews it and provides detailed feedback.
+3. **Ask further questions**: Use the prefilled prompts or ask your own follow-up questions to get additional insights.
+4. **Improve your resume**: Apply the suggestions and feedback to create a standout resume.
+
+---
+
+## 🌟 Why Use This App?
+
+- **Efficient Resume Review**: Automate the process of getting personalized resume feedback from an AI that has been fine-tuned for review tasks.
+- **Interactive**: Engage with the AI through follow-up questions for deeper analysis and more comprehensive feedback.
+- **Easy-to-use Interface**: Simple upload process and a user-friendly interface for a seamless experience.
+
+---
+
+## 💻 Tech Stack
+
+- **Streamlit**: Provides the frontend interface for the application.
+- **TuneStudio API**: Powers the AI-driven feedback using their language model.
+- **PyPDF2**: Used for parsing and extracting text from uploaded PDF resumes.
+- **Python**: Core language for the app logic.
+
+---
+
+## 🔑 License
+
+This project is open-source and available under the [MIT License](LICENSE).
+
+---
+
+## 🎉 Contributions
+
+Feel free to fork this repository and submit pull requests. Any ideas for improving the app or additional features are welcome!
+
+---
+
+## 🙋‍♂️ Contact
+
+If you have any questions or feedback, feel free to open an issue or reach out via [[email protected]](mailto:[email protected]).
+
+---
+
+Give it a try and improve your resume today! 💼✨

community/resumefy/app.py

@@ -0,0 +1,148 @@
+import streamlit as st
+import requests
+import PyPDF2
+import json
+
+# Set TuneStudio API credentials and URL
+TUNE_API_URL = "https://proxy.tune.app/chat/completions"
+TUNE_API_KEY = "YOUR_API_KEY"  # Replace with your actual TuneStudio API key
+MODEL = "anthropic/claude-3.5-sonnet"  # The model you want to use
+
+headers = {
+    "Authorization": f"Bearer {TUNE_API_KEY}",
+    "Content-Type": "application/json",
+}
+
+
+def parse_pdf(file):
+    """Parses uploaded PDF file and returns text."""
+    reader = PyPDF2.PdfReader(file)
+    text = ""
+    for page_num in range(len(reader.pages)):
+        page = reader.pages[page_num]
+        text += page.extract_text()
+    return text
+
+
+def get_tune_response(messages, temperature=0.8, max_tokens=900, stream=False):
+    """Sends a prompt and messages to the TuneStudio API and gets the response."""
+    data = {
+        "temperature": temperature,
+        "messages": messages,
+        "model": MODEL,
+        "stream": stream,
+        "frequency_penalty": 0,
+        "max_tokens": max_tokens,
+    }
+    response = requests.post(TUNE_API_URL, headers=headers, json=data)
+
+    if stream:
+        # Handle streaming responses (if TuneStudio supports it)
+        for line in response.iter_lines():
+            if line:
+                l = line[6:]
+                if l != b"[DONE]":
+                    return json.loads(l)
+    else:
+        return response.json()
+
+
+def main():
+    st.title("Resumefy: all judgy and grumpy!")
+
+    # Step 1: Upload the resume PDF
+    uploaded_file = st.file_uploader("Upload your resume as a PDF", type="pdf")
+
+    if uploaded_file:
+        # Step 2: Parse the PDF and extract text
+        with st.spinner("Reading through your resume..."):
+            resume_text = parse_pdf(uploaded_file)
+            st.success("PDF processed successfully!")
+
+        # Display extracted text for transparency
+        st.subheader("Mirror of your life:")
+        st.text_area("Resume Text", value=resume_text, height=300)
+
+        # Step 3: Generate a prompt to send to the LLM
+        initial_prompt = (
+            f"I have the following resume. Can you please review it and provide feedback? "
+            f"Here is the resume:\n{resume_text}\n"
+            "Please include suggestions for improvement and general recommendations."
+        )
+
+        # Prepare messages for the chat completion API
+        messages = [
+            {
+                "role": "system",
+                "content": "You are a helpful assistant that reviews resumes.",
+            },
+            {"role": "user", "content": initial_prompt},
+        ]
+
+        # Step 4: Send the prompt to TuneStudio and get the feedback
+        if st.button("Judge Me!"):
+            with st.spinner("Sending resume to Resumefy for feedback..."):
+                feedback = get_tune_response(messages)
+            st.subheader("Feedback:")
+            st.write(feedback["choices"][0]["message"]["content"])
+
+            # Add feedback to the conversation history
+            messages.append(
+                {
+                    "role": "assistant",
+                    "content": feedback["choices"][0]["message"]["content"],
+                }
+            )
+
+            # Prefilled prompts for follow-up questions
+            st.subheader("Ask Further Questions:")
+            st.write("Here are some suggestions you might ask the model:")
+            suggested_prompts = [
+                "Can you suggest ways to highlight my skills better?",
+                "What are the most important points I should focus on?",
+                "How can I make my resume stand out?",
+                "Can you help me tailor this resume for a specific job role?",
+            ]
+
+            # Display the prefilled prompts as buttons
+            for prompt in suggested_prompts:
+                if st.button(prompt):
+                    with st.spinner(f"Asking: '{prompt}'..."):
+                        # Send the follow-up question to TuneStudio
+                        messages.append({"role": "user", "content": prompt})
+                        follow_up_response = get_tune_response(messages)
+                        st.subheader(f"Response to: '{prompt}'")
+                        st.write(follow_up_response["choices"][0]["message"]["content"])
+                        messages.append(
+                            {
+                                "role": "assistant",
+                                "content": follow_up_response["choices"][0]["message"][
+                                    "content"
+                                ],
+                            }
+                        )
+
+            # Step 5: Allow the user to ask custom follow-up questions
+            st.subheader("Ask a Custom Question:")
+            user_question = st.text_input("Your question:")
+
+            if st.button("Submit Question"):
+                if user_question.strip() != "":
+                    with st.spinner(f"Asking: '{user_question}'..."):
+                        # Send the custom question to TuneStudio
+                        messages.append({"role": "user", "content": user_question})
+                        custom_response = get_tune_response(messages)
+                        st.subheader(f"Response to: '{user_question}'")
+                        st.write(custom_response["choices"][0]["message"]["content"])
+                        messages.append(
+                            {
+                                "role": "assistant",
+                                "content": custom_response["choices"][0]["message"][
+                                    "content"
+                                ],
+                            }
+                        )
+
+
+if __name__ == "__main__":
+    main()