Files
ThothII/harness/tests/test_adapter_factory.py
T

190 lines
6.1 KiB
Python

import pytest
from tht.adapters.dwh import PostgresDwhAdapter, ThothRestDwhAdapter
from tht.adapters.factory import build_dwh, build_vector_store
from tht.adapters.vector import PgVectorStore, QdrantVectorStore, ThothHttpVectorStore
from tht.config import Config, ConfigError
def _config(*, dwh_type="thoth_rest", vector_type="thoth_vector_http", reader=True, writer=True):
dwh = (
{
"type": "thoth_rest",
"database": {"database": "analytics", "schema": "mart"},
"endpoint": {"base_url": "https://dwh.test/", "api_key": "reader"},
}
if dwh_type == "thoth_rest"
else {
"type": "postgres_direct",
"connection": {
"host": "db",
"database": "analytics",
"schema": "mart",
"user": "reader",
"password": "secret",
},
}
)
vectors = (
{
"type": "thoth_vector_http",
**(
{"reader": {"base_url": "https://vectors.test/", "api_key": "reader"}}
if reader
else {}
),
**(
{"writer": {"base_url": "https://vectors.test/", "api_key": "writer"}}
if writer
else {}
),
}
if vector_type == "thoth_vector_http"
else {
"type": "pgvector_direct",
**(
{
"reader": {
"host": "vector-db",
"database": "postgres",
"schema": "vectors",
"user": "reader",
"password": "secret",
}
}
if reader
else {}
),
**(
{
"writer": {
"host": "vector-db",
"database": "postgres",
"schema": "vectors",
"user": "writer",
"password": "secret",
}
}
if writer
else {}
),
}
)
legacy_database = (
dwh["connection"]
if dwh_type == "postgres_direct"
else {
**dwh["database"],
"user": "rest",
"password": "",
"transport": "rest",
}
)
return Config.model_validate({"dwh": dwh, "vectors": vectors, "database": legacy_database})
@pytest.mark.parametrize(
("dwh_type", "adapter_type"),
[("postgres_direct", PostgresDwhAdapter), ("thoth_rest", ThothRestDwhAdapter)],
)
def test_factory_selects_dwh_adapter(dwh_type, adapter_type):
assert isinstance(build_dwh(_config(dwh_type=dwh_type)), adapter_type)
def test_factory_selects_http_vector_and_requires_writer():
config = _config(writer=False)
assert isinstance(build_vector_store(config), ThothHttpVectorStore)
with pytest.raises(ConfigError, match="writer"):
build_vector_store(config, require_write=True)
def test_factory_builds_writer_only_http_vector_when_write_is_required():
config = _config(reader=False, writer=True)
store = build_vector_store(config, require_write=True)
assert isinstance(store, ThothHttpVectorStore)
assert store.capabilities.search is False
assert store.capabilities.upsert is True
def test_factory_selects_direct_vector_store_and_requires_writer():
config = _config(vector_type="pgvector_direct", writer=False)
assert isinstance(build_vector_store(config), PgVectorStore)
with pytest.raises(ConfigError, match="writer"):
build_vector_store(config, require_write=True)
def test_factory_builds_writer_only_direct_vector_when_write_is_required():
store = build_vector_store(
_config(vector_type="pgvector_direct", reader=False), require_write=True
)
assert isinstance(store, PgVectorStore)
assert store.capabilities.search is False
assert store.capabilities.upsert is True
def test_factory_selects_qdrant_for_schema_v3_runtime():
config = Config.model_validate(
{
"dwh": {
"type": "postgres_direct",
"connection": {
"host": "db",
"database": "analytics",
"schema": "mart",
"user": "reader",
"password": "secret",
},
},
"database": {
"host": "db",
"database": "analytics",
"schema": "mart",
"user": "reader",
"password": "secret",
"transport": "direct",
},
"vectors": {
"type": "qdrant",
"base_url": "http://qdrant:6333",
"collection": "psd-clinical",
},
"embeddings": {
"provider": "ollama_internal",
"base_url": "http://embedding:11434",
"model": "qwen3-embedding:0.6b",
"dim": 1024,
},
}
)
config._workspace_id = "psd-clinical"
config._workspace_revision = "a" * 40
store = build_vector_store(config, require_write=True)
assert isinstance(store, QdrantVectorStore)
assert store.capabilities.search is True
assert store.capabilities.upsert is True
def test_factory_reuses_legacy_direct_connection_for_server_writes_only():
server = _config(vector_type="pgvector_direct", writer=False)
server.vectors.connection = server.vectors.reader
server.vectors.reader = None
store = build_vector_store(server, require_write=True)
assert store.capabilities.search is True
assert store.capabilities.upsert is True
server.profile = "workstation"
with pytest.raises(ConfigError, match="writer"):
build_vector_store(server, require_write=True)
def test_factory_propagates_non_default_statement_timeout():
config = _config(dwh_type="postgres_direct")
config.execution.statement_timeout_ms = 12_345
assert build_dwh(config)._statement_timeout_ms == 12_345