Create app.py

app.py

@@ -0,0 +1,95 @@
+import gradio as gr
+from transformers import pipeline, AutoTokenizer, AutoModelForCausalLM
+from rdkit import Chem
+from rdkit.Chem import AllChem
+import py3Dmol
+import random
+import torch
+
+# Load models
+bio_gpt = pipeline("text-generation", model="microsoft/BioGPT-Large")
+chemberta_tokenizer = AutoTokenizer.from_pretrained("seyonec/ChemBERTa-zinc-base-v1")
+chemberta_model = AutoModelForCausalLM.from_pretrained("seyonec/ChemBERTa-zinc-base-v1")
+compliance_qa = pipeline("question-answering", model="nlpaueb/legal-bert-base-uncased")
+
+# Generate random SMILES string
+def generate_smiles():
+    sample_smiles = ["CCO", "CCN", "C1=CC=CC=C1", "C(C(=O)O)N", "CC(C)CC"]
+    return random.choice(sample_smiles)
+
+# Convert SMILES to 3D molecule HTML using py3Dmol
+def mol_to_3d_html(smiles):
+    mol = Chem.MolFromSmiles(smiles)
+    mol = Chem.AddHs(mol)
+    AllChem.EmbedMolecule(mol, AllChem.ETKDG())
+    AllChem.UFFOptimizeMolecule(mol)
+    block = Chem.MolToMolBlock(mol)
+
+    view = py3Dmol.view(width=400, height=400)
+    view.addModel(block, "mol")
+    view.setStyle({"stick": {}})
+    view.zoomTo()
+    return view.render().data  # Gradio-compatible HTML output
+
+# Main pipeline
+def run_pipeline(disease, symptoms):
+    print(f"Received input: disease={disease}, symptoms={symptoms}")
+    
+    # Literature generation
+    try:
+        prompt = f"Recent treatment and drug research for {disease} with symptoms: {symptoms}"
+        insights = bio_gpt(prompt, max_length=200, do_sample=True)[0]['generated_text']
+    except Exception as e:
+        print("BioGPT Error:", e)
+        insights = f"❌ Error generating literature: {e}"
+
+    # 3D Molecule from random SMILES
+    try:
+        smiles = generate_smiles()
+        html_3d = mol_to_3d_html(smiles)
+    except Exception as e:
+        print("3D Molecule Error:", e)
+        smiles = "N/A"
+        html_3d = f"<p>❌ Error rendering molecule: {str(e)}</p>"
+
+    # ChemBERTa Score
+    try:
+        inputs = chemberta_tokenizer(smiles, return_tensors="pt")
+        with torch.no_grad():
+            outputs = chemberta_model(**inputs)
+        score = round(outputs.logits.mean().item(), 3)
+    except Exception as e:
+        print("ChemBERTa Error:", e)
+        score = f"❌ Error: {e}"
+
+    # LegalBERT Compliance
+    try:
+        compliance = compliance_qa(
+            question="What does FDA require for drug testing?",
+            context="FDA requires extensive testing for new drug candidates including Phase I, II, and III clinical trials."
+        )["answer"]
+    except Exception as e:
+        print("LegalBERT Error:", e)
+        compliance = f"❌ Error: {e}"
+
+    return insights, smiles, html_3d, score, compliance
+
+# Gradio Interface
+iface = gr.Interface(
+    fn=run_pipeline,
+    inputs=[
+        gr.Textbox(label="🦠 Disease", placeholder="e.g., Lung Cancer"),
+        gr.Textbox(label="🩺 Symptoms", placeholder="e.g., fatigue, breathlessness"),
+    ],
+    outputs=[
+        gr.Textbox(label="📜 Literature Insights"),
+        gr.Textbox(label="🧪 Generated SMILES"),
+        gr.HTML(label="🧬 3D Molecule"),
+        gr.Textbox(label="📊 Molecular Property Score (ChemBERTa)"),
+        gr.Textbox(label="⚖️ Legal Compliance (FDA)"),
+    ],
+    title="💊 Drug Discovery using LLMs",
+    description="Enter disease & symptoms to explore literature, generate molecule structure, and simulate compliance.",
+)
+
+iface.launch()