From 7bcde6857e838a0e649291f953b68d110321da2b Mon Sep 17 00:00:00 2001
From: Jen Hamon <jhamon@pinecone.io>
Date: Fri, 12 Jan 2024 18:14:30 -0500
Subject: [PATCH] Initial load generation script

---
 scripts/generate_usage.py | 90 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 90 insertions(+)
 create mode 100755 scripts/generate_usage.py

diff --git a/scripts/generate_usage.py b/scripts/generate_usage.py
new file mode 100755
index 00000000..1a2fc4a3
--- /dev/null
+++ b/scripts/generate_usage.py
@@ -0,0 +1,90 @@
+import os
+import random
+import string
+from pinecone.grpc import PineconeGRPC
+
+def read_env_var(name):
+    value = os.environ.get(name)
+    if value is None:
+        raise Exception('Environment variable {} is not set'.format(name))
+    return value
+
+def random_string(length):
+    return ''.join(random.choice(string.ascii_lowercase) for i in range(length))
+
+def random_embedding_values(dimension=2):
+    return [random.random() for _ in range(dimension)]
+
+def write_gh_output(name, value):
+    with open(os.environ['GITHUB_OUTPUT'], 'a') as fh:
+        print(f'{name}={value}', file=fh)
+
+DIMENSION = 1536 # common for openai embeddings
+
+def create_index_if_not_exists(pc, index_name):
+    if index_name not in pc.list_indexes().names():
+        print(f'Index {index_name} does not exist, creating it')
+        pc.create_index(
+            name=index_name,
+            metric='cosine',
+            dimension=DIMENSION,
+            spec={
+                'serverless': {
+                    'cloud': read_env_var('CLOUD'),
+                    'region': read_env_var('REGION'),
+                }
+            }
+        )
+
+upserted_ids = set()
+
+def main():
+    pc = PineconeGRPC(api_key=read_env_var('PINECONE_API_KEY'))
+    index_name = read_env_var('INDEX_NAME')
+    iterations = int(read_env_var('ITERATIONS'))
+
+    create_index_if_not_exists(pc, index_name)
+
+    index = pc.Index(name=index_name)
+    for i in range(iterations):
+        try:
+            # Upsert some vectors
+            items_to_upsert = random.randint(1, 100)
+            vector_list = [
+                {
+                    'id': random_string(10), 
+                    'values': random_embedding_values(DIMENSION),
+                    'metadata': {
+                        'genre': random.choice(['action', 'comedy', 'drama']),
+                        'runtime': random.randint(60, 120)
+                    }
+                } for x in range(items_to_upsert)
+            ]
+            index.upsert(vectors=vector_list)
+            print('Upserted {} vectors'.format(items_to_upsert))
+
+            for v in vector_list:
+                upserted_ids.add(v['id'])
+
+            # Fetch some vectors
+            ids_to_fetch = random.sample(upserted_ids, k=random.randint(1, 20))
+            print('Fetching {} vectors'.format(len(ids_to_fetch)))
+            fetched_vectors = index.fetch(ids=ids_to_fetch)
+
+            # Query some vectors
+            print('Querying 10 times')
+            for i in range(10):
+                # Query by vector values
+                query_vector = random_embedding_values(DIMENSION)
+                query_results = index.query(vector=query_vector, top_k=10)
+            
+            # Delete some vectors
+            print('Deleting some vectors')
+            id_to_delete = random.sample(upserted_ids, k=random.randint(1, 10))
+            index.delete(ids=id_to_delete)
+        except Exception as e:
+            print('Exception: {}'.format(e))
+
+if __name__ == '__main__':
+    main()
+