app.py

import gradio as gr
import torch
from transformers import pipeline

# -------------------
# 1. Model definitions
# -------------------
MODELS = {
    "econbert": "climatebert/econbert",
    "controversy-classification": "climatebert/ClimateControversyBERT_classification",
    "controversy-bert": "climatebert/ClimateControversyBert",
    "netzero-reduction": "climatebert/netzero-reduction",
    "transition-physical": "climatebert/transition-physical",
    "renewable": "climatebert/renewable",
    "climate-detector": "climatebert/distilroberta-base-climate-detector",
    "climate-commitment": "climatebert/distilroberta-base-climate-commitment",
    "climate-tcfd": "climatebert/distilroberta-base-climate-tcfd",
    "climate-s": "climatebert/distilroberta-base-climate-s",
    "climate-specificity": "climatebert/distilroberta-base-climate-specificity",
    "climate-sentiment": "climatebert/distilroberta-base-climate-sentiment",
    "environmental-claims": "climatebert/environmental-claims",
    "climate-f": "climatebert/distilroberta-base-climate-f",
    "climate-d-s": "climatebert/distilroberta-base-climate-d-s",
    "climate-d": "climatebert/distilroberta-base-climate-d",
}

# -------------------
# 2. Human-readable label maps
# -------------------
LABEL_MAPS = {
    "climate-commitment": {
        "LABEL_0": "Not about climate commitments",
        "LABEL_1": "About climate commitments",
    },
    "climate-detector": {
        "LABEL_0": "Not climate-related",
        "LABEL_1": "Climate-related",
    },
    "climate-sentiment": {
        "LABEL_0": "Negative",
        "LABEL_1": "Neutral",
        "LABEL_2": "Positive",
    },
    "climate-specificity": {
        "LABEL_0": "Low specificity",
        "LABEL_1": "Medium specificity",
        "LABEL_2": "High specificity",
    },
    "netzero-reduction": {
        "LABEL_0": "No net-zero / reduction commitment",
        "LABEL_1": "Net-zero / reduction commitment",
    },
    "transition-physical": {
        "LABEL_0": "Transition risk",
        "LABEL_1": "Physical risk",
    },
    "renewable": {
        "LABEL_0": "Not about renewables",
        "LABEL_1": "About renewables",
    },
}

# -------------------
# 3. Pipeline cache
# -------------------
pipelines = {}

def load_model(model_key):
    """Load and cache a model pipeline."""
    if model_key not in pipelines:
        repo_id = MODELS[model_key]
        device = 0 if torch.cuda.is_available() else -1
        print(f"🔹 Loading model: {model_key} ({repo_id})")
        pipelines[model_key] = pipeline(
            "text-classification",
            model=repo_id,
            device=device,
            torch_dtype=torch.float16 if device == 0 else None,
            truncation=True,
            max_length=512
        )
    return pipelines[model_key]

# -------------------
# 4. Inference across all models
# -------------------
def predict_all_models(text):
    """Run inference across all ClimateBERT models and return structured output."""
    if not text.strip():
        return "⚠️ Please enter some text."

    results_summary = []
    for model_key, repo in MODELS.items():
        try:
            model = load_model(model_key)
            outputs = model(text)
            label_map = LABEL_MAPS.get(model_key, {})

            formatted = "\n".join([
                f"• {label_map.get(r['label'], r['label'])}: {r['score']:.2f}"
                for r in outputs
            ])

            results_summary.append(f"### {model_key}\n{formatted}")
        except Exception as e:
            results_summary.append(f"### {model_key}\n❌ Error: {str(e)}")

    return "\n\n".join(results_summary)

# -------------------
# 5. Gradio UI
# -------------------
with gr.Blocks(title="🌍 ClimateBERT All-Models Analyzer") as demo:
    gr.Markdown("""
    # 🌍 ClimateBERT Multi-Model Analysis
    This app runs **all ClimateBERT models** on your input text (`mergedMarkdown` style).  
    It detects sentiment, specificity, renewables, commitments, and more — all at once.
    """)

    text_input = gr.Textbox(
        label="Input Text (mergedMarkdown)",
        placeholder="Paste the sustainability report, ESG statement, or corporate disclosure here...",
        lines=5
    )

    output = gr.Markdown(label="Model Outputs")

    run_btn = gr.Button("🔍 Run All Models")
    run_btn.click(predict_all_models, inputs=text_input, outputs=output)

    gr.Markdown("""
    ---
    **Note:** Each model captures a different aspect of climate-related discourse (e.g., sentiment, specificity, commitments, etc.).
    """)

if __name__ == "__main__":
    demo.launch(server_name="0.0.0.0", server_port=7860)