Update app.py

app.py

@@ -2,15 +2,26 @@ import gradio as gr
 import sqlite3
 import json
 import numpy as np
-import numpy as np
+import os
 from sentence_transformers import SentenceTransformer
+from huggingface_hub import hf_hub_download
 
 EMBEDDING_MODEL = "nomic-ai/nomic-embed-text-v1.5"
 embedder = SentenceTransformer(EMBEDDING_MODEL, trust_remote_code=True)
 
-# Ensure the database file exists locally.
-db_path = "hpo_genes.db"
+HF_TOKEN = os.environ.get("HF_TOKEN")
+if not HF_TOKEN:
+    raise ValueError("Missing Hugging Face API token.")
+
+db_repo = "UoS-HGIG/hpo_genes"
+db_filename = "hpo_genes.db"
+db_path = db_filename
 
+# Download database if not present locally
+if not os.path.exists(db_path):
+    db_path = hf_hub_download(
+        repo_id=db_repo, filename=db_filename, repo_type="dataset", use_auth_token=HF_TOKEN
+    )
 
 def find_best_hpo(finding, region, threshold):
     query = f"{finding} {region}" if region else finding
@@ -18,21 +29,28 @@ def find_best_hpo(finding, region, threshold):
 
     conn = sqlite3.connect(db_path)
     cursor = conn.cursor()
-    cursor.execute("SELECT hpo_id, hpo_name, embedding FROM hpo_embeddings")
+
+    try:
+        cursor.execute("SELECT hpo_id, hpo_name, embedding FROM hpo_embeddings")
+    except sqlite3.OperationalError as e:
+        conn.close()
+        raise ValueError(f"Database error: {e}")
 
     best_match, best_score = None, -1
 
     for hpo_id, hpo_name, embedding_str in cursor.fetchall():
         hpo_embedding = np.array(json.loads(embedding_str))
-        score = np.dot(query_embedding, hpo_embedding)
+        similarity = np.dot(query_embedding, hpo_embedding) / (
+            np.linalg.norm(query_embedding) * np.linalg.norm(hpo_embedding)
+        )
 
-        if score > best_score:
-            best_score = score
+        if similarity > best_score:
+            best_score = similarity
             best_match = {"hpo_id": hpo_id, "hpo_name": hpo_name}
 
+    conn.close()
     return best_match if best_score >= threshold else None
 
-
 def get_genes(hpo_id):
     conn = sqlite3.connect(db_path)
     cursor = conn.cursor()
@@ -41,26 +59,27 @@ def get_genes(hpo_id):
     conn.close()
     return genes
 
-
 def hpo_mapper(finding, region, threshold):
     if not finding:
         return "Please provide a pathological finding.", "", ""
 
-    match = find_best_hpo(finding, region, threshold)
+    try:
+        match = find_best_hpo(finding, region, threshold)
+    except ValueError as e:
+        return str(e), "", ""
 
     if match:
         genes = get_genes(match["hpo_id"])
         return match["hpo_id"], match["hpo_name"], ", ".join(genes)
 
-    return "No match found", "", ""
-
+    return "No match found.", "", ""
 
 demo = gr.Interface(
     fn=hpo_mapper,
     inputs=[
         gr.Textbox(label="Pathological Finding"),
         gr.Textbox(label="Anatomical Region (optional)"),
-        gr.Slider(0.5, 1.0, step=0.01, value=0.74, label="Similarity Threshold")
+        gr.Slider(0.5, 1.0, step=0.01, value=0.74, label="Similarity Threshold"),
     ],
     outputs=[
         gr.Textbox(label="HPO ID"),
@@ -73,12 +92,15 @@ demo = gr.Interface(
         '(e.g. "terminal ileum") to map it to the closest Human Phenotype Ontology (HPO) '
         'term and retrieve associated genes.\n\n'
         '**Reference:**\n'
-        'Application of Generative Artificial Intelligence to Utilise Unstructured Clinical Data for Acceleration of Inflammatory Bowel Disease Research\n'
-        'Alex Z Kadhim, Zachary Green, Iman Nazari, Jonathan Baker, Michael George, Ashley Heinson, Matt Stammers, Christopher M Kipps, R Mark Beattie, James J Ashton, Sarah Ennis\n'
-        'medRxiv 2025.03.07.25323569; doi: [https://doi.org/10.1101/2025.03.07.25323569](https://doi.org/10.1101/2025.03.07.25323569)\n\n'
+        'Application of Generative Artificial Intelligence to Utilise Unstructured Clinical Data '
+        'for Acceleration of Inflammatory Bowel Disease Research\n'
+        'Alex Z Kadhim, Zachary Green, Iman Nazari, Jonathan Baker, Michael George, Ashley Heinson, '
+        'Matt Stammers, Christopher M Kipps, R Mark Beattie, James J Ashton, Sarah Ennis\n'
+        'medRxiv 2025.03.07.25323569; doi: '
+        '[https://doi.org/10.1101/2025.03.07.25323569](https://doi.org/10.1101/2025.03.07.25323569)\n\n'
         'HPO to gene mappings obtained from [Jax](https://hpo.jax.org/data/annotations)'
-    )
+    ),
 )
 
 if __name__ == "__main__":
-    demo.launch()
+    demo.launch()