app.py

import gradio as gr
import torch
from transformers import RobertaForSequenceClassification, RobertaTokenizer
import numpy as np
import tempfile

# Load model and tokenizer
model_name = "SamanthaStorm/abuse-pattern-detector-v2"
model = RobertaForSequenceClassification.from_pretrained(model_name)
tokenizer = RobertaTokenizer.from_pretrained(model_name)

# Define labels
LABELS = [
    "gaslighting", "mockery", "dismissiveness", "control",
    "guilt_tripping", "apology_baiting", "blame_shifting", "projection",
    "contradictory_statements", "manipulation", "deflection", "insults",
    "obscure_formal", "recovery_phase", "suicidal_threat", "physical_threat",
    "extreme_control"
]

# Custom thresholds per label (feel free to adjust)
THRESHOLDS = {
    "gaslighting": 0.15,
    "mockery": 0.15,
    "dismissiveness": 0.25,
    "control": 0.03,
    "guilt_tripping": 0.15,
    "apology_baiting": 0.15,
    "blame_shifting": 0.15,
    "projection": 0.20,
    "contradictory_statements": 0.15,
    "manipulation": 0.15,
    "deflection": 0.15,
    "insults": 0.25,
    "obscure_formal": 0.20,
    "recovery_phase": 0.15,
    "suicidal_threat": 0.08,
    "physical_threat": 0.045,
    "extreme_control": 0.30,
}

# Label categories
PATTERN_LABELS = LABELS[:14]
DANGER_LABELS = LABELS[14:]

def calculate_abuse_level(scores, thresholds):
    triggered_scores = [score for label, score in zip(LABELS, scores) if score > thresholds[label]]
    if not triggered_scores:
        return 0.0
    return round(np.mean(triggered_scores) * 100, 2)

def interpret_abuse_level(score):
    if score > 80:
        return "Extreme / High Risk"
    elif score > 60:
        return "Severe / Harmful Pattern Present"
    elif score > 40:
        return "Likely Abuse"
    elif score > 20:
        return "Mild Concern"
    else:
        return "Very Low / Likely Safe"

def analyze_messages(input_text):
    input_text = input_text.strip()
    if not input_text:
        return "Please enter a message for analysis.", None

    # Tokenize and predict
    inputs = tokenizer(input_text, return_tensors="pt", truncation=True, padding=True)
    with torch.no_grad():
        outputs = model(**inputs)
    scores = torch.sigmoid(outputs.logits.squeeze(0)).numpy()

    # Count triggered labels
    pattern_count = sum(score > THRESHOLDS[label] for label, score in zip(PATTERN_LABELS, scores[:14]))
    danger_flag_count = sum(score > THRESHOLDS[label] for label, score in zip(DANGER_LABELS, scores[14:]))

    # Abuse level calculation
    abuse_level = calculate_abuse_level(scores, THRESHOLDS)
    abuse_description = interpret_abuse_level(abuse_level)

    # Resource logic
    if danger_flag_count >= 2:
        resources = (
            "**Immediate Help:** Call 911 if in danger.\n\n"
            "**Crisis Support:** National DV Hotline – [thehotline.org/plan-for-safety](https://www.thehotline.org/plan-for-safety/)\n"
            "**Legal Support:** WomensLaw – [womenslaw.org](https://www.womenslaw.org/)\n"
            "**Specialized Services:** RAINN, StrongHearts, LGBTQ+, immigrant, neurodivergent resources"
        )
    elif danger_flag_count == 1:
        resources = (
            "**Emotional Abuse Info:** [thehotline.org/resources](https://www.thehotline.org/resources/what-is-emotional-abuse/)\n"
            "**Relationship Education:** [joinonelove.org](https://www.joinonelove.org/)\n"
            "**Support Chat:** [thehotline.org](https://www.thehotline.org/)\n"
            "**Community Groups:** LGBTQ+, immigrant, and neurodivergent spaces"
        )
    else:
        resources = (
            "**Healthy Relationships:** [loveisrespect.org](https://www.loveisrespect.org/)\n"
            "**Find a Therapist:** [psychologytoday.com](https://www.psychologytoday.com/us/therapists)\n"
            "**Relationship Tools:** [relate.org.uk](https://www.relate.org.uk/)\n"
            "**Online Peer Support:** (including identity-focused groups)"
        )

    result_md = (
        f"### 📋 Analysis Summary\n\n"
        f"**Abuse Pattern Count:** {pattern_count}\n"
        f"**Danger Cues Detected:** {danger_flag_count}\n"
        f"**Abuse Level:** {abuse_level}% ({abuse_description})\n\n"
        f"### 🛟 Suggested Support Resources\n{resources}"
    )

    # Save to .txt
    with tempfile.NamedTemporaryFile(delete=False, suffix=".txt", mode="w") as f:
        f.write(result_md)
        report_path = f.name

    return result_md, report_path

# Interface
with gr.Blocks() as demo:
    gr.Markdown("# 🔍 Abuse Pattern Detector")
    gr.Markdown("Paste one or more messages for analysis (multi-line supported).")

    text_input = gr.Textbox(label="Text Message(s)", lines=10, placeholder="Paste messages here...")
    result_output = gr.Markdown()
    file_output = gr.File(label="📥 Download Analysis (.txt)")

    text_input.submit(analyze_messages, inputs=text_input, outputs=[result_output, file_output])
    analyze_btn = gr.Button("Analyze")
    analyze_btn.click(analyze_messages, inputs=text_input, outputs=[result_output, file_output])

if __name__ == "__main__":
    demo.launch()