code/interface/old/annotators.py

import gradio as gr
import json
import random
import os
from datetime import datetime

# --- PATH CONFIGURATION ---
DATA_PATH = "/home/mshahidul/readctrl/data/data_annotator_data/crowdsourcing_input_en_v2.json"
SAVE_ROOT = "/home/mshahidul/readctrl/data/annotators_validate_data"
QUESTIONS_FILE = "/home/mshahidul/readctrl/code/interface/sp50_questions.json"

# --- SESSION CONFIGURATION ---
NUM_QUESTIONS = 30        
NUM_DUPLICATES = 4       
NUM_LITERACY_QUERIES = 10 
DUPLICATE_INTERVAL = 8   

# --- ANNOTATION GUIDE TEXT ---
GUIDE_HTML = """
<div style="background-color: #f9f9f9; padding: 15px; border-left: 6px solid #2196F3; border-radius: 4px;">
    <h3>Rating Guide: Medical Text Difficulty</h3>
    <p>Please rate the difficulty of the documents based on the following scale:</p>
    <table style="width:100%; border-collapse: collapse; text-align: left;">
        <tr style="background-color: #e3f2fd;">
            <th style="padding: 8px; border: 1px solid #ddd;">Score</th>
            <th style="padding: 8px; border: 1px solid #ddd;">Description</th>
        </tr>
        <tr>
            <td style="padding: 8px; border: 1px solid #ddd;"><b>1 - 2</b></td>
            <td style="padding: 8px; border: 1px solid #ddd;"><b>Very Easy:</b> Clear language, no medical jargon. Like a 5th-grade textbook.</td>
        </tr>
        <tr>
            <td style="padding: 8px; border: 1px solid #ddd;"><b>3 - 4</b></td>
            <td style="padding: 8px; border: 1px solid #ddd;"><b>Easy:</b> Common medical terms (e.g., "fever", "heart") used in simple sentences.</td>
        </tr>
        <tr>
            <td style="padding: 8px; border: 1px solid #ddd;"><b>5 - 6</b></td>
            <td style="padding: 8px; border: 1px solid #ddd;"><b>Moderate:</b> Some technical terms. Requires focused reading but understandable.</td>
        </tr>
        <tr>
            <td style="padding: 8px; border: 1px solid #ddd;"><b>7 - 8</b></td>
            <td style="padding: 8px; border: 1px solid #ddd;"><b>Hard:</b> Heavy use of medical jargon. Read like a clinical report.</td>
        </tr>
        <tr>
            <td style="padding: 8px; border: 1px solid #ddd;"><b>9 - 10</b></td>
            <td style="padding: 8px; border: 1px solid #ddd;"><b>Very Hard:</b> Specialist-level text. Extremely dense and difficult to follow.</td>
        </tr>
    </table>
</div>
"""

def load_questions():
    with open(QUESTIONS_FILE, "r") as f:
        all_q = json.load(f)
    return random.sample(all_q, min(NUM_LITERACY_QUERIES, len(all_q)))

class AnnotationSession:
    def __init__(self, dataset, questions):
        base_samples = random.sample(dataset, NUM_QUESTIONS)
        self.queue = list(base_samples)
        for i in range(NUM_DUPLICATES):
            self.queue.insert(DUPLICATE_INTERVAL + i, base_samples[i])
        
        self.current_index = 0
        self.results = []
        self.questions = questions
        self.session_folder = None 

with open(DATA_PATH, "r") as f:
    full_dataset = json.load(f)

session = AnnotationSession(full_dataset, load_questions())

def start_and_save_literacy(*answers):
    timestamp = datetime.now().strftime("%Y-%m-%d_%H-%M-%S")
    session_folder = os.path.join(SAVE_ROOT, timestamp)
    os.makedirs(session_folder, exist_ok=True)
    session.session_folder = session_folder 

    literacy_data = []
    for i, ans in enumerate(answers):
        q_info = session.questions[i]
        literacy_data.append({
            "question_id": q_info['id'],
            "question_text": q_info['question'],
            "user_answer": ans,
            "is_correct": ans == q_info['correct']
        })
    
    with open(os.path.join(session_folder, "literacy_results.json"), "w") as f:
        json.dump(literacy_data, f, indent=4)
    
    first_pair = session.queue[0]
    return (
        gr.update(visible=False), 
        gr.update(visible=True),  
        first_pair['original_doc'], 
        first_pair['wiki_anchor'],
        f"Item 1 of {len(session.queue)}"
    )

def submit_rating(doc_slider, wiki_slider):
    # 1. Capture the current result
    current_pair = session.queue[session.current_index]
    session.results.append({
        "original_index": current_pair.get('index', 'unknown'),
        "queue_position": session.current_index,
        "doc_rating": doc_slider,
        "wiki_rating": wiki_slider,
        "timestamp": datetime.now().strftime("%Y-%m-%d %H:%M:%S")
    })
    
    # 2. Incremental Save: Write to file immediately on every click
    annotation_file = os.path.join(session.session_folder, "annotation_results.json")
    with open(annotation_file, "w") as f:
        json.dump(session.results, f, indent=4)
    
    # 3. Show Pop-up Notification (Gradio Info Toast)
    gr.Info(f"Progress Saved: Item {session.current_index + 1} recorded.")

    # Increment index
    session.current_index += 1
    
    # 4. Check if session is finished
    if session.current_index < len(session.queue):
        next_pair = session.queue[session.current_index]
        return (
            next_pair['original_doc'], 
            next_pair['wiki_anchor'], 
            f"Item {session.current_index + 1} of {len(session.queue)}",
            5, 5  # Reset sliders to neutral middle value
        )
    else:
        # Final update for the UI when done
        return (
            "✅ ALL TASKS COMPLETED", 
            "The data has been saved to your session folder. You may close this tab.", 
            "Status: Finished", 
            0, 0
        )

# --- UI ---
with gr.Blocks(theme=gr.themes.Soft()) as demo:
    gr.Markdown("# Medical Text Readability Annotation")
    
    # Instructions available at all times via Accordion
    with gr.Accordion("See Annotation Instructions & Scale Guide", open=True):
        gr.HTML(GUIDE_HTML)

    with gr.Column(visible=True) as intro_box:
        gr.Markdown(f"### Pre-Task: Health Literacy Check ({NUM_LITERACY_QUERIES} Questions)")
        literacy_inputs = []
        for q in session.questions:
            radio = gr.Radio(choices=q['options'], label=q['question'])
            literacy_inputs.append(radio)
        btn_start = gr.Button("Start Annotation", variant="primary")

    with gr.Column(visible=False) as task_box:
        progress = gr.Label(label="Progress")
        with gr.Row():
            with gr.Column():
                doc_display = gr.Textbox(interactive=False, lines=12, label="Document D (Medical Text)")
                doc_slider = gr.Slider(1, 10, step=1, label="Difficulty (1: Simple → 10: Technical)", value=0)
            with gr.Column():
                wiki_display = gr.Textbox(interactive=False, lines=12, label="Document W (Wikipedia Text)")
                wiki_slider = gr.Slider(1, 10, step=1, label="Difficulty (1: Simple → 10: Technical)", value=0)
        btn_submit = gr.Button("Submit & Next", variant="primary")

    btn_start.click(
        start_and_save_literacy, 
        inputs=literacy_inputs, 
        outputs=[intro_box, task_box, doc_display, wiki_display, progress]
    )
    
    btn_submit.click(
        submit_rating,
        inputs=[doc_slider, wiki_slider],
        outputs=[doc_display, wiki_display, progress, doc_slider, wiki_slider]
    )

demo.launch(share=True)