Skip to content

Server Pattern Tests

View source code

There are 6 tests covering three access patterns: standalone embedded, server-managed embedded, and HTTP API (remote access). All tests validate expected usage flows and performance characteristics.

The 3 Access Patterns

  1. Standalone: db = arcadedb.create_database("./mydb") — simplest, no server overhead
  2. Server-managed: server.start() → db = server.create_database() — HTTP API available simultaneously
  3. HTTP API: requests.post(url, json=...) — remote/cross-process access

Test Cases

Starts server first, creates database through server, and validates both embedded access and HTTP availability. Asserts connection to studio URL and data retrieval via Java API.

2) thread safety

Creates server-managed database with 20 items, then spawns 5 threads concurrently querying ranges. Asserts all threads access successfully without errors.

3) context manager

Uses with arcadedb.create_server(...) as server: for automatic start/stop, creates database, inserts note, verifies count.

5) embedded performance comparison

Creates both standalone and server-managed databases with 500 complex multi-field records and runs 100 varied queries (filters, aggregations, date ranges, LIKE patterns). Measures time and ops/sec. Asserts server-managed performance is similar (within 1.5x, typically 0.99x) to standalone. Validates that embedded access through server is direct JVM call, not HTTP.

6) HTTP API access pattern

Starts server, creates database via Java API (HTTP API lacks create), inserts 5 products via HTTP, queries via HTTP (returns JSON), performs 100 mixed CRUD ops (inserts, filtered queries, updates, aggregations, LIKE searches) via both HTTP and Java API. Records time and ops/sec. Calculates ratio (HTTP typically 5–7x slower) and per-op overhead (~0.7–1ms). Uses requests.Session() with auth for realistic connection pooling.

Quick Patterns

Just need a database: 

db = arcadedb.create_database("./mydb")

Need HTTP API available: 

server = arcadedb.create_server(root_path="./databases")
server.start()
db = server.create_database("mydb")
# Java API is direct JVM call (0.99x perf of standalone)
# HTTP also available at http://localhost:2480

Pre-populating before server: 

db = arcadedb.create_database("./temp_db")
# ... populate ...
db.close()  # ⚠️ Release file lock!
# Move to server directory, then start server

Remote/HTTP access: 

import requests
session = requests.Session()
session.auth = ("root", "password")
r = session.post("http://localhost:2480/api/v1/query/mydb",
                  json={"language": "sql", "command": "SELECT ..."})
# 5–7x slower than Java API, ~0.7ms overhead per op