import gradio as gr
import logging

# Set up logging
logger = logging.getLogger('gradio_app.processors.topic')


def process_topic_modeling(analysis_results, prompt, analyses):
    """
    Process Topic Modeling analysis and return UI updates

    Args:
        analysis_results (dict): Complete analysis results
        prompt (str): The prompt being analyzed
        analyses (dict): Analysis data for the prompt

    Returns:
        tuple: UI component updates
    """
    topic_results = analyses["topic_modeling"]

    # Check for errors in topic modeling
    if "error" in topic_results:
        return (
            analysis_results,
            False,  # Don't show raw JSON
            False,  # Don't show visualization area
            gr.update(visible=False),
            gr.update(visible=False),
            gr.update(visible=False),
            gr.update(visible=False),
            gr.update(visible=False),
            gr.update(visible=False),
            gr.update(visible=False),
            gr.update(visible=False),
            gr.update(visible=False),
            True,  # Show status message
            gr.update(visible=True, value=f"❌ **Topic modeling error:** {topic_results['error']}"),
            gr.update(visible=False)  # bias_visualizations
        )

    visualization_area_visible = True
    models = topic_results.get("models", [])
    method = topic_results.get("method", "lda").upper()
    n_topics = topic_results.get("n_topics", 3)

    if len(models) < 2:
        from analysis_runner import default_no_visualization
        return default_no_visualization(analysis_results)

    prompt_title_visible = True
    prompt_title_value = f"## Analysis of Prompt: \"{prompt[:100]}...\""

    models_compared_visible = True
    models_compared_value = f"### Topic Modeling Analysis ({method}, {n_topics} topics)"

    # Initialize component visibility and values
    model1_title_visible = False
    model1_title_value = ""
    model1_words_visible = False
    model1_words_value = ""
    model2_title_visible = False
    model2_title_value = ""
    model2_words_visible = False
    model2_words_value = ""
    similarity_title_visible = False
    similarity_metrics_visible = False
    similarity_metrics_value = ""

    # Extract and format topic information
    topics = topic_results.get("topics", [])

    if topics:
        # Format topic info for display
        topic_info = []
        for topic in topics[:5]:  # Show first 5 topics
            topic_id = topic.get("id", 0)
            words = topic.get("words", [])[:5]  # Top 5 words per topic

            if words:
                topic_info.append(f"**Topic {topic_id + 1}**: {', '.join(words)}")

        if topic_info:
            model1_title_visible = True
            model1_title_value = "#### Discovered Topics"
            model1_words_visible = True
            model1_words_value = "\n".join(topic_info)

    # Get topic distributions for models
    model_topics = topic_results.get("model_topics", {})

    if model_topics:
        model1_name = models[0]
        model2_name = models[1]

        # Format topic distribution info
        if model1_name in model_topics and model2_name in model_topics:
            model2_title_visible = True
            model2_title_value = "#### Topic Distribution"
            model2_words_visible = True

            # Simple distribution display
            dist1 = model_topics[model1_name]
            dist2 = model_topics[model2_name]

            model2_words_value = f"""
            **{model1_name}**: {', '.join([f"Topic {i + 1}: {v:.2f}" for i, v in enumerate(dist1[:5])])}

            **{model2_name}**: {', '.join([f"Topic {i + 1}: {v:.2f}" for i, v in enumerate(dist2[:5])])}
            """

    # Add similarity metrics if available
    comparisons = topic_results.get("comparisons", {})
    if comparisons:
        comparison_key = f"{model1_name} vs {model2_name}"

        if comparison_key in comparisons:
            metrics = comparisons[comparison_key]
            js_div = metrics.get("js_divergence", 0)

            # Add interpretation
            similarity_text = ""
            if js_div < 0.2:
                similarity_text = "very similar"
            elif js_div < 0.4:
                similarity_text = "somewhat similar"
            elif js_div < 0.6:
                similarity_text = "moderately different"
            else:
                similarity_text = "very different"

            similarity_title_visible = True
            similarity_metrics_visible = True
            similarity_metrics_value = f"""
            - **Topic Distribution Divergence**: {js_div:.4f}
            - The topic distributions between models are **{similarity_text}**
            - *Lower divergence values indicate more similar topic distributions*
            """

    return (
        analysis_results,  # analysis_results_state
        False,  # analysis_output visibility
        True,  # visualization_area_visible
        gr.update(visible=True),  # analysis_title
        gr.update(visible=prompt_title_visible, value=prompt_title_value),  # prompt_title
        gr.update(visible=models_compared_visible, value=models_compared_value),  # models_compared
        gr.update(visible=model1_title_visible, value=model1_title_value),  # model1_title
        gr.update(visible=model1_words_visible, value=model1_words_value),  # model1_words
        gr.update(visible=model2_title_visible, value=model2_title_value),  # model2_title
        gr.update(visible=model2_words_visible, value=model2_words_value),  # model2_words
        gr.update(visible=similarity_title_visible),  # similarity_metrics_title
        gr.update(visible=similarity_metrics_visible, value=similarity_metrics_value),  # similarity_metrics
        False,  # status_message_visible
        gr.update(visible=False),  # status_message
        gr.update(visible=False)  # bias_visualizations - Not visible for Topic Modeling
    )