import os
import json
import vertexai
from vertexai.generative_models import GenerativeModel
import vertexai.preview.generative_models as generative_models
import gradio as gr

# Read the service account key JSON file path from environment variable
SERVICE_ACCOUNT_KEY_PATH = os.getenv("GOOGLE_APPLICATION_CREDENTIALS")

if not SERVICE_ACCOUNT_KEY_PATH:
    raise ValueError("The GOOGLE_APPLICATION_CREDENTIALS environment variable is not set.")

with open(SERVICE_ACCOUNT_KEY_PATH) as f:
    service_account_info = json.load(f)
os.environ["GOOGLE_APPLICATION_CREDENTIALS"] = SERVICE_ACCOUNT_KEY_PATH

def generate(text):
    try:
        vertexai.init(project="idyllic-now-424815-h2", location="us-central1")
        model = GenerativeModel(
            "gemini-1.5-flash-001",
            system_instruction=[
                'Objective', text, 'Instructions', """You are an AI model designed to provide concise information about big data analytics across various fields without mentioning the question. Respond with a focused, one-line answer that captures the essence of the key risk, benefit, or trend associated with the topic.

            
input: What do you consider the most significant risk of over-reliance on big data analytics in stock market risk management?
output: Increased market volatility.

input: What is a major benefit of big data analytics in healthcare?
output: Enhanced patient care through personalized treatment.

input: What is a key challenge of big data analytics in retail?
output: Maintaining data privacy and security.

input: What is a primary advantage of big data analytics in manufacturing?
output: Improved production efficiency and predictive maintenance.

input: What is a significant risk associated with big data analytics in education?
output: Potential widening of the achievement gap if data is not used equitably. """

            ]
        )
        generation_config = {
            'max_output_tokens': 3019,
            'temperature': 1,
            'top_p': 0.32,
        }
        safety_settings = {
            generative_models.HarmCategory.HARM_CATEGORY_HATE_SPEECH: generative_models.HarmBlockThreshold.BLOCK_NONE,
            generative_models.HarmCategory.HARM_CATEGORY_DANGEROUS_CONTENT: generative_models.HarmBlockThreshold.BLOCK_NONE,
            generative_models.HarmCategory.HARM_CATEGORY_SEXUALLY_EXPLICIT: generative_models.HarmBlockThreshold.BLOCK_NONE,
            generative_models.HarmCategory.HARM_CATEGORY_HARASSMENT: generative_models.HarmBlockThreshold.BLOCK_NONE,
        }
        responses = model.generate_content(
            [text],
            generation_config=generation_config,
            safety_settings=safety_settings,
            stream=True,
        )

        response_text = ""
        for response in responses:
            response_text += response.text

        return response_text if response_text else "No valid response generated or response was blocked."

    except Exception as e:
        return str(e)

# # Custom HTML and JavaScript for "Copy to Clipboard" functionality
# js = """

#     function copyToClipboard() {
#         var copyText = document.getElementById("output-textbox");
#         copyText.select();
#         document.execCommand("copy");
#     }
# """

iface = gr.Interface(
    fn=generate,

    inputs=gr.Textbox(lines=2, placeholder="Enter text here..."),
    outputs="text",
    title="Chuunibyou Text Generator",
    description="Transform text into an elaborate and formal style with a nobleman tone.",
    live=False
)

def launch_custom_interface():
    iface.launch()
    with gr.TabbedInterface(fn=generate, inputs=gr.Textbox(lines=2, placeholder="Enter text here..."), outputs=gr.HTML(label="Output")) as ti:
        ti.add(custom_html)

if __name__ == "__main__":
    launch_custom_interface()