GraphBatch Tests

View source code

These tests cover the engine-backed GraphBatch helper used for bulk graph ingest.

Covered Behavior

1) create vertices and edges

Creates Person vertices through GraphBatch, buffers Knows edges, and verifies the final outgoing traversal from Alice.

2) create_vertices returns RIDs

Verifies that create_vertices(...) returns RID strings for every requested row, including sparse property rows such as None.

3) invalid WAL flush mode

Confirms that invalid wal_flush values fail early with ValueError instead of being silently accepted.

4) parallel flush smoke

Exercises parallel_flush=True and verifies final vertex and edge counts plus graph connectivity.

5) retry and memory knobs

Exercises commit_retries, commit_retry_delay_ms, chunk_cache_capacity and max_deferred_incoming_edges on a five-vertex chain. The bounds are set deliberately tiny (a 2-entry chunk cache, a 1-edge deferred cap) so the bounded paths are the ones taken. Both are pure accelerators, so the test asserts the correct answer (5 vertices, 4 edges, and vertex 4 as the only incoming neighbour of vertex 5); it does not run a second, unbounded load to compare against.

6) invalid knob values are rejected

Confirms the four knobs above reach the Java builder, by asserting its own range validation fires. This is the test that discriminates a wired parameter from an accepted-and-ignored one, because none of the four changes an observable result. It asserts on ArcadeDBError and the engine's "must be" wording rather than grepping for the parameter name: an unwired keyword raises TypeError: ... unexpected keyword argument 'commit_retries', whose message contains the parameter name, so a name-substring check passes on exactly the broken code it is meant to catch.

7) a one-way edge in a two-way type is refused

test_graph_batch_refuses_one_way_edges_in_a_two_way_type: with bidirectional=False, new_edge on an edge type declared two-way (the CREATE EDGE TYPE default) raises ArcadeDBError naming the type and the word "bidirectional", and the type holds 0 edges; on a type declared UNIDIRECTIONAL the same batch writes its edge (1). Fails on 26.9.1, which accepted the edge (ArcadeData/arcadedb#8625).

8) what sees a one-way edge

test_one_way_edges_are_seen_by_patterns_not_by_in: 50 questions each tagged twice through a UNIDIRECTIONAL type (100 edges). Cypher counts 100 written from the source, from the target, and undirected; SQL MATCH from the target returns 100 rows; in(), inE(), both(), and bothE() from the target return 0, out() from the source 100, and the target vertex's get_in_edges() is empty. Fails on 26.9.1, where the incoming Cypher pattern returned 0.

9) a failed create_vertices leaves no transaction open

test_graph_batch_create_vertices_failure_rolls_back (JSON-bulk and property-matrix paths): two rows with the same key on a unique index make create_vertices raise ArcadeDBError, db.is_transaction_active() is False afterwards, the batch still creates the next vertex, and an INSERT outside any transaction is refused. Fails before humemai/arcadedb-embedded-python#121, where the transaction stayed active and the insert was accepted and lost at close. test_graph_batch_create_vertices_keeps_the_callers_transaction checks the other half: a transaction the caller opened before the call is still active after the failure. test_graph_batch_create_vertices_keyboard_interrupt_rolls_back raises KeyboardInterrupt after the engine's transaction has begun and expects the rollback to run.

Why It Matters

GraphBatch is the repository's preferred bulk graph-ingest path from Python, so these tests protect the performance-oriented API surface and its configuration validation behavior.