Update app.py

app.py

@@ -6,12 +6,12 @@ import json
 from sentence_transformers import SentenceTransformer
 
 def load_data():
-    with h5py.File('patent_embeddings.h5', 'r') as f:
+    with h5py.File('path_to_your_embeddings_file.h5', 'r') as f:
         embeddings = f['embeddings'][:]
-        patent_numbers = [pn.decode('utf-8') for pn in f['patent_numbers'][:]]
+        patent_numbers = f['patent_numbers'][:]
     
     metadata = {}
-    with open('patent_metadata.jsonl', 'r') as f:
+    with open('path_to_your_metadata_file.jsonl', 'r') as f:
         for line in f:
             data = json.loads(line)
             metadata[data['patent_number']] = data
@@ -36,14 +36,12 @@ model = SentenceTransformer('all-mpnet-base-v2')
 
 def exact_text_match(query, metadata):
     matches = []
+    query_lower = query.lower()
     for patent_number, data in metadata.items():
-        abstract = data.get('abstract', '')
-        claims = data.get('claims', '')
-        if isinstance(abstract, float):
-            abstract = ''
-        if isinstance(claims, float):
-            claims = ''
-        if query.lower() in abstract.lower() or query.lower() in claims.lower():
+        text = data.get('text', '')
+        if isinstance(text, float):
+            text = ''
+        if query_lower in text.lower():
             matches.append((patent_number, 1.0))  # Score of 1.0 for exact match
     return matches
 
@@ -63,14 +61,14 @@ def search(query, top_k=5):
     
     results = []
     for i, idx in enumerate(indices[0]):
-        patent_number = patent_numbers[idx]
+        patent_number = patent_numbers[idx].decode('utf-8')
         if patent_number not in metadata:
             print(f"Warning: Patent number {patent_number} not found in metadata")
             continue
         patent_data = metadata[patent_number]
         result = f"Patent Number: {patent_number}\n"
-        abstract = patent_data.get('abstract', 'No abstract available')
-        result += f"Abstract: {abstract[:200]}...\n"
+        text = patent_data.get('text', 'No text available')
+        result += f"Text: {text[:200]}...\n"
         result += f"Similarity Score: {distances[0][i]:.4f}\n\n"
         results.append(result)
     
@@ -78,11 +76,15 @@ def search(query, top_k=5):
     if not results or distances[0][0] < 0.5:
         print("Falling back to exact text match")
         exact_matches = exact_text_match(query, metadata)
+        if exact_matches:
+            print(f"Exact matches found: {exact_matches}")
+        else:
+            print("No exact matches found")
         for patent_number, score in exact_matches[:top_k]:
             patent_data = metadata[patent_number]
             result = f"Patent Number: {patent_number}\n"
-            abstract = patent_data.get('abstract', 'No abstract available')
-            result += f"Abstract: {abstract[:200]}...\n"
+            text = patent_data.get('text', 'No text available')
+            result += f"Text: {text[:200]}...\n"
             result += f"Exact Match Score: {score:.4f}\n\n"
             results.append(result)
     
@@ -98,4 +100,4 @@ iface = gr.Interface(
 )
 
 if __name__ == "__main__":
-    iface.launch(share=True)
+    iface.launch()