GuyPaddock · August 8, 2024 03:31 · Aug 8, 2024
diff --git a/index_data.py b/index_data.py
@@ -0,0 +1,45 @@
+import json
+from elasticsearch import Elasticsearch
+
+# Load JSON data
+with open("documents.json", "r") as file:
+    documents_raw = json.load(file)
+
+# Flatten the documents structure
+documents = []
+for course in documents_raw:
+    course_name = course['course']
+    for doc in course['documents']:
+        doc['course'] = course_name
+        documents.append(doc)
+
+# Initialize Elasticsearch client
+es_client = Elasticsearch('http://localhost:9200')
+
+# Define index settings and mappings
+def create_index(index_name="course-questions"):
+    index_settings = {
+        "settings": {
+            "number_of_shards": 1,
+            "number_of_replicas": 0
+        },
+        "mappings": {
+            "properties": {
+                "text": {"type": "text"},
+                "section": {"type": "text"},
+                "question": {"type": "text"},
+                "course": {"type": "keyword"}
+            }
+        }
+    }
+    if not es_client.indices.exists(index=index_name):
+        es_client.indices.create(index=index_name, body=index_settings)
+
+# Create the index
+create_index()
+
+# Index documents
+for doc in documents:
+    es_client.index(index="course-questions", body=doc)
+
+print("Data indexed successfully")