import gradio as gr
import pandas as pd
from fpdf import FPDF
import sib_api_v3_sdk
import base64
import json
import os

class AdaptiveAI:
    def __init__(self, subjects, scores_list):
        self.df = pd.DataFrame({
            "Subject": subjects,
            "Score": [int(s) for s in scores_list]
        })
        self.days = ["Monday", "Tuesday", "Wednesday", "Thursday", "Friday", "Saturday", "Sunday"]

    def generate_plan(self, feedback_data=None):
        def get_base_settings(score):
            if score < 40: return 150, "Focus on fundamental concepts."
            if score < 65: return 120, "Practice more topical questions."
            if score < 80: return 90, "Review mistakes & past year papers."
            return 60, "Quick revision & advance topics."

        self.df['Time_Mins'], self.df['Advice'] = zip(*self.df['Score'].apply(get_base_settings))
        self.df = self.df.sort_values(by='Score').reset_index(drop=True)
        self.df['Day'] = self.days[:len(self.df)]

        if feedback_data:
            for fb in feedback_data:
                sub = fb['Subject']
                idx = self.df[self.df['Subject'] == sub].index
                if not idx.empty:
                    current_t = self.df.loc[idx, 'Time_Mins'].values[0]
                    if fb['Status'] == 'stuck' or fb['Level'] <= 2:
                        self.df.loc[idx, 'Time_Mins'] = current_t + 30
                    elif fb['Status'] == 'finished' and fb['Level'] >= 4:
                        self.df.loc[idx, 'Time_Mins'] = max(45, current_t - 20)
        return self.df

def create_pdf(df):
    pdf = FPDF()
    pdf.add_page()
    pdf.set_font("Arial", 'B', 16)
    pdf.cell(0, 15, "AI ADAPTIVE STUDY PLAN", ln=True, align='C')
    pdf.set_font("Arial", size=10)
    for _, row in df.iterrows():
        line = f"{row['Day']} | {row['Subject']} | {row['Time_Mins']} mins | {row['Advice']}"
        pdf.cell(0, 10, line, ln=True, border=1)
    pdf_path = "study_plan.pdf"
    pdf.output(pdf_path)
    return pdf_path

def send_email(email_to, pdf_path):
    api_key = "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"
    configuration = sib_api_v3_sdk.Configuration()
    configuration.api_key['api-key'] = api_key
    api_instance = sib_api_v3_sdk.TransactionalEmailsApi(sib_api_v3_sdk.ApiClient(configuration))
    
    with open(pdf_path, "rb") as f:
        pdf_b64 = base64.b64encode(f.read()).decode('utf-8')

    smtp_email = sib_api_v3_sdk.SendSmtpEmail(
        to=[{"email": email_to}],
        sender={"name": "AI Tutor", "email": "syuantengzhiya@gmail.com"},
        subject="Your New Study Plan",
        html_content="<p>Attached is your personalized plan.</p>",
        attachment=[{"content": pdf_b64, "name": "Plan.pdf"}]
    )
    api_instance.send_transac_email(smtp_email)

def api_interface(student_email, subjects_json, scores_json, feedback_json=None):
    try:
        subjects = json.loads(subjects_json)   # ["Math", "Science", ...]
        scores   = json.loads(scores_json)     # [75, 60, ...]
        feedback = json.loads(feedback_json) if feedback_json and feedback_json.strip() else None

        engine = AdaptiveAI(subjects, scores)
        plan_df = engine.generate_plan(feedback)
        pdf_file = create_pdf(plan_df)
        send_email(student_email, pdf_file)

        result = plan_df[['Day','Subject','Time_Mins','Advice']].to_dict('records')
        return json.dumps(result)
    except Exception as e:
        return f"Error: {str(e)}"

demo = gr.Interface(
    fn=api_interface,
    inputs=[
        gr.Textbox(label="Email"),
        gr.Textbox(label="Subjects JSON"),   # ["Math","Science",...]
        gr.Textbox(label="Scores JSON"),     # [75, 60, ...]
        gr.Textbox(label="Feedback JSON")
    ],
    outputs="text"
)
demo.launch()