import os
import openai
import datetime
import utilities.parsed_policies as policies
import utilities.engagements as engagements
import utilities.pickfromlist as pick
import gradio as gr
import json
from jinja2 import Template
import csv

# openai.api_key = os.getenv("OPENAI_API_KEY")
openai.api_key = "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"

#Write log file function
def append_to_qalog(your_role, school_selection, output_format, input_text, gpt_response,response_time,question_cost):
    # Define the filename
    filename = "qalog.csv"
    
    # Check if file exists to decide if headers are needed
    file_exists = os.path.isfile(filename)
    
    # Open the file in append mode
    with open(filename, 'a', newline='') as csvfile:
        writer = csv.DictWriter(csvfile, fieldnames=["your_role", "school_selection", "output_format", "input_text", "gpt_response","response_time","question_cost"])
        
        # If the file didn't exist, write the headers first
        if not file_exists:
            writer.writeheader()
        
        # Append the data
        writer.writerow({
            "your_role": your_role,
            "school_selection": school_selection,
            "output_format": output_format,
            "input_text": input_text,
            "gpt_response": gpt_response,
            "response_time": response_time,
            "question_cost": question_cost
        })

#Chatbot Function
def chatbot(your_role,school_selection,output_format,input_text):

    start_time = datetime.datetime.now()

    # Read in the prompt header for engagement analysis
    with open('prompts/nili_analysis_header.txt', 'r') as file:
        header = file.read()

    # school_selection holds an array of one or more schools, so create a comma delimited str to process.
    #print(school_selection)
    str=",".join(school_selection)

    # Read the Hydrated policies

    # Split the input string by commas to get the list of university names
    university_names = str.split(',')

    policies = ''

    for university in university_names:
        # Construct the file path
        file_path = f"nili_hydrated_policies/{university}.txt"

        try:
            # Open the file and read its contents
            with open(file_path, 'r', encoding='utf-8') as file:
                content = file.read()

                # If 'policies' already has content, add separator before appending
                if policies:
                    policies += "\n----------\n"
                
                policies += content
        
        except FileNotFoundError:

            print(f"File for {university} not found!")

    with open('combined_policies.txt', 'w', encoding='utf-8') as out_file:
        out_file.write(policies)

    #print(policies)
    
    # Insert into prompt
    with open("prompts/nili_engagement_user_v2.j2", "r") as file:
    #with open("prompts/nili_engagement_user_v3.j2", "r") as file:    
        template_content = file.read()

    # Create a Jinja2 template from the content
    template = Template(template_content)

    # Render the template with the policy JSON
    analysis_input = template.render(policies=policies, question=input_text,format=output_format)

    with open('analysis_input.txt', 'w', encoding='utf-8') as out_file:
        out_file.write(analysis_input)

    response = openai.ChatCompletion.create(
        model="gpt-4",
      # model="gpt-3.5-turbo",
        temperature=0,
        messages=[
            {
                "role": "system",
                "content": header
            },
            {
                "role": "user",
                "content": analysis_input
            }
          ]
        )

    gpt_response = response.choices[0].message["content"]

    tokens_used = response.usage

    question_cost = (tokens_used.get('total_tokens', 0) / 1000) * .03

    #print(question_cost)

    with open('response.txt', 'w', encoding='utf-8') as out_file:
        out_file.write(gpt_response)

    """print(ui_examples)

    # Read the csv file
    with open('examples/examples_v3.csv', 'r') as csvfile:
        reader = csv.reader(csvfile)
        next(reader)  # Skip the header row
        for row in reader:
            ui_examples.append([convert_to_none(item) for item in row])

    print(ui_examples)"""

    # Define the style and content
    label_text = "NILI Response"
    color = "#6562F4"
    background_color = "white"
    border_radius = "10px"  # You can adjust this value as needed

    response_label = f'<h3 style="color: {color}; background-color: {background_color}; border-radius: {border_radius}; padding: 10px;display: inline-block;">{label_text}</h3>'

    end_time = datetime.datetime.now()

    response_time = end_time - start_time

    append_to_qalog(your_role, school_selection, output_format, input_text, gpt_response,response_time,question_cost)

    return response_label,gpt_response

def load_example(example_id):
    global ui_examples
    return ui_examples[example_id][0]



#Gradio UI
CIMStheme = gr.themes.Soft().set(button_primary_background_fill='#6562F4')

# Initialize an empty list to store the examples
ui_examples = []

nili_response = "Test"

# Function to convert 'None' string to None object
def convert_to_none(value):
    return None if value == 'None' else value

# Read the csv file
with open('examples/examples_v2.csv', 'r') as csvfile:
    reader = csv.reader(csvfile)
    next(reader)  # Skip the header row
    for row in reader:
        ui_examples.append([convert_to_none(item) for item in row])

#response_label='<!DOCTYPE html><html lang="en"><head><meta charset="UTF-8"><meta name="viewport" content="width=device-width, initial-scale=1.0"><title>Formatted Text</title><style>.formatted-text {color: black;font-weight: bold;text-align: left;}</style></head><body><div class="formatted-text">NILI Response</div></body></html>'
# Define the style and content
label_text = "NILI Response"
color = "#6562F4"
background_color = "white"
border_radius = "10px"  # You can adjust this value as needed

response_label = f'<h3 style="color: {color}; background-color: {background_color}; border-radius: {border_radius}; padding: 10px;display: inline-block;">{label_text}</h3>'

with gr.Blocks(CIMStheme) as iface:
    with gr.Row():
        with gr.Column(scale=2):
            gr.Image(label="Logo",value="CIMS Logo Purple.png",width=10,show_download_button=False,interactive=False,show_label=False,elem_id="logo",container=False)
        with gr.Column(scale=2):
            #gr.Textbox(value="# NILI - Powered by CIMS.AI",show_label=False,interactive=False,text_align="center",elem_id="CIMSTitle")
            gr.Markdown(value="# NILI - Powered by CIMS.AI")
        with gr.Column(scale=2):
            gr.Markdown("")
    with gr.Row():
        with gr.Column():
            gr.Interface(fn=chatbot,
                     inputs=[
                         gr.components.Dropdown(["Student Athlete","Parent","Athletic Director"],multiselect=False,info="Select a role.",label="User Role"),
                         gr.components.Dropdown(["Stanford","CU","LSU","CMU","Tulane","Cal Poly","Long Beach State University","Auburn"],multiselect=True,info="Select one or more schools. This will help set the context of your question.",label="School Context"),
                         gr.components.Dropdown(["Summary","Detailed Analysis","Table"],multiselect=False,info="Select the desired output format.",label="Output Format"),
                         gr.components.Textbox(lines=5, placeholder="Enter your question here", label="NIL Question")],
                     outputs=[
                         gr.components.Markdown(response_label),
                         gr.components.HTML(label="NILI Response")
                         ],
                     description="Ask any question about Name, Image, Likeness (NIL)",
                     allow_flagging="manual",
                     examples=ui_examples,
                     flagging_options=["The response is incorrect","The response is inappropriate","The response doesn't make sense"]
                      )
    with gr.Row():
        with gr.Column():
            gr.HTML('<center><i>CIMS.AI Confidential 2023</i></center>')

iface.launch(share=True)