import gradio as gr
import os
from openai import OpenAI

# Set your OpenAI API key
client = OpenAI(api_key=os.environ['OPENAI_API_KEY'])

jd_summary_global = ""  # Global variable to store the job description summary

def process_jd(text):
    global jd_summary_global  # Declare the global variable
    if not text.strip():  # Check if the text is empty or contains only whitespace
        jd_summary_global = "No JD"  # Update the global variable
        return "No JD"
    
    try:
        # Structuring a prompt to ask GPT-3.5 to summarize the job description
        prompt = f"Summarize the following job description into its job nature, responsibilities, and requirements:\n\n{text}"
        
        # Uploading text to OpenAI 
        response = client.chat.completions.create(model="gpt-3.5-turbo", messages=[{"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": prompt}])
        jd_summary = response.choices[0].message.content.strip()
        jd_summary_global = jd_summary  # Update the global variable
        return jd_summary
    except Exception as e:
        return str(e)

def cv_rating(cv_data):
    global jd_summary_global  # Declare the global variable
    global cv_rating_global
    
    if len(jd_summary_global) <= 1 or jd_summary_global == "No JD":  
        return "No JD in the previous tab."
    if len(cv_data) <= 1:
        return "No CV data"
    try:
        # Construct a prompt to ask GPT-3.5 to rate the CV based on the job description summary
        prompt = f"""
        Job Description Summary: {jd_summary_global}
        CV Data: {cv_data}
        
        Rate the compatibility of the CV with the job description and provide strengths, weaknesses, and recommendations to strengthen the CV.
        """
        # Uploading text to OpenAI 
        response = client.chat.completions.create(model="gpt-3.5-turbo", messages=[{"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": prompt}])
        cv_rating_global= response.choices[0].message.content.strip()
        return cv_rating_global
    except Exception as e:
        return str(e)

def create_cover_letter(additional_info):
    global jd_summary_global  # Declare the global variable
    global cv_rating_global
    if len(jd_summary_global) <= 1 or jd_summary_global == "No JD":  
        return "No JD in the previous tab."
    if len(cv_rating_global) <= 1:
        return "No CV data"
    try:
        # Constructing a prompt for GPT-3.5 to create a tailored cover letter
        prompt = f"""
        Job Description: {jd_summary_global}
        CV Data: {cv_rating_global}
        Additional Information: {additional_info}

        Create a tailored cover letter based on the job description and CV data provided:
        """
        # Uploading text to OpenAI 
        response = client.chat.completions.create(model="gpt-3.5-turbo", messages=[{"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": prompt}])
        return response.choices[0].message.content.strip()
    except Exception as e:
        return str(e)

def interview_qa(additional_info):
    global cv_rating_global
    if len(cv_rating_global) <= 1:
        return "No CV data"
    try:
        # Constructing a prompt for GPT-3.5 to create interview questions and answers
        prompt = f"""
        CV Data: {cv_rating_global}
        Additional Information: {additional_info}

        Generate at least 10 interview questions and provide potential answers based on the CV data and additional information provided:
        """
        # Uploading text to OpenAI 
        response = client.chat.completions.create(model="gpt-3.5-turbo", messages=[{"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": prompt}])
        return response.choices[0].message.content.strip()
    except Exception as e:
        return str(e)
        
def suggest_cv_content(additional_info):
    global jd_summary_global  # Accessing the global variable for job description summary
    global cv_rating_global   # Accessing the global variable for CV data
    
    if len(jd_summary_global) <= 1 or jd_summary_global == "No JD":  
        return "No JD in the previous tab."
    if len(cv_rating_global) <= 1:
        return "No CV data"
    
    try:
        # Constructing a prompt for GPT-3.5 to suggest tailored CV content
        prompt = f"""
        Given the following job description, generate a new CV to better match the job description. Also, ensure the suggestions are formatted in a way that is compatible with most ATS solutions.

        Job Description: {jd_summary_global}
        CV Data: {cv_rating_global}
        Additional Information: {additional_info}
        """
        # Uploading text to OpenAI 
        response = client.chat.completions.create(model="gpt-3.5-turbo",
            messages=[{"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": prompt}]
        )
        return response.choices[0].message.content.strip()
    except Exception as e:
        return str(e)

jd_sum = gr.Interface(
    fn=process_jd,  # function to process the text
    inputs=gr.Textbox(lines=30, label="Job Description"),
    outputs=gr.Textbox(lines=30, label="JD Summary", show_copy_button=True),
    live=False,
    title="Job Description Summarizer",
    description="An app to summarize job descriptions into job nature, responsibilities, and requirements. \
        For more info, check out: https://github.com/jmesplana/BespokeCV",
    api_name="jd_sum"
)

cv_rate_interface = gr.Interface(
    fn=cv_rating,
    inputs=gr.Textbox(lines=30, label="CV Data", placeholder="Paste the CV data here"),
    outputs=gr.Textbox(lines=30, label="ATS Rating System", show_copy_button=True),
    live=False,
    title="CV Rating",
    description="An app to rate CV compatibility with job description, providing strengths, weaknesses, and recommendations.",
    api_name="cv_rate_interface"
)

cover_letter_interface = gr.Interface(
    fn=create_cover_letter,
    inputs=[gr.Textbox(lines=10, label="Additional Information", placeholder="Add any additional information or preferences for your cover letter here")],
    outputs=gr.Textbox(lines=30, label="Output", show_copy_button=True),
    live=False,
    title="Cover Letter Creator",
    description="An app to create a tailored cover letter based on job description and CV data. You may input additional information in the additional information box to add highlight specific experiences/projects and/or skills.",
    api_name="cover_letter_interface"
)

interview_qa_interface = gr.Interface(
    fn=interview_qa,
    inputs=[gr.Textbox(lines=10, label="Additional Information", placeholder="Add any specific questions or additional information here")],
    outputs=gr.Textbox(lines=30, label="Output", show_copy_button=True),
    live=False,
    title="Interview Q&A",
    description="An app to generate interview questions and answers based on CV data and additional information.",
    api_name="interview_qa"
)

cv_suggestion_interface = gr.Interface(
    fn=suggest_cv_content,
    inputs=[gr.Textbox(lines=10, label="Additional Information", placeholder="Add any specific requests or additional information here")],
    outputs=gr.Textbox(lines=30, label="Output", show_copy_button=True),
    live=False,
    title="CV Content Suggestion",
    description="An app to suggest CV content tailored to the job description, optimized for ATS compatibility.",
    api_name="cv_suggestion"
)


bespokecv = gr.TabbedInterface([jd_sum, cv_rate_interface,cover_letter_interface,interview_qa_interface,cv_suggestion_interface],
                               tab_names=['Job Description Summarizer','CV ATS Rating','Cover Letter Generator','Interview Q&A','Suggested CV'])

if __name__ == "__main__":
    bespokecv.launch(share=True)