Files
qdrant/openapi/tests/openapi_integration/test_init_from_collection.py
Tim Visée 1e4ce145a2 Parameterize integration tests (#2375)
* Clean up imports

* Clean up whitespace

* Use drop collection function from helpers

* Use parameterized integration tests to test on_disk variations

* Promote on_disk_vectors parameter into dedicated fixture

* Merge basic retrieve test with parameterization

* Flatten functions

* Split validation test functions

* Don't parameterize on_disk states when testing validation

* Parameterize init_from_collection test sequence

* Parameterize on_disk_payload in tests with fixture

* Parameterize new batch update tests

* Fix integration test collection setup distance after incorrect merge
2023-08-09 12:32:27 +02:00

207 lines
6.1 KiB
Python

import pytest
from .helpers.collection_setup import drop_collection
from .helpers.fixtures import on_disk_vectors, on_disk_payload
from .helpers.helpers import request_with_validation
collection_name = 'test_collection'
source_collection_name = 'source_collection'
@pytest.fixture(autouse=True, scope="module")
def setup():
yield
drop_collection(collection_name=collection_name)
drop_collection(collection_name=source_collection_name)
# Delete and create collection
def advanced_collection_setup(collection_name, size, distance, on_disk_vectors, on_disk_payload):
response = request_with_validation(
api='/collections/{collection_name}',
method="DELETE",
path_params={'collection_name': collection_name},
)
assert response.ok
response = request_with_validation(
api='/collections/{collection_name}',
method="PUT",
path_params={'collection_name': collection_name},
body={
"vectors": {
"size": size,
"distance": distance,
"on_disk": on_disk_vectors,
},
"on_disk_payload": on_disk_payload
}
)
assert response.ok
response = request_with_validation(
api='/collections/{collection_name}',
method="GET",
path_params={'collection_name': collection_name},
)
assert response.ok
# Delete and create collection
def advanced_collection_multi_setup(collection_name, vectors_config):
response = request_with_validation(
api='/collections/{collection_name}',
method="DELETE",
path_params={'collection_name': collection_name},
)
assert response.ok
response = request_with_validation(
api='/collections/{collection_name}',
method="PUT",
path_params={'collection_name': collection_name},
body={
"vectors": vectors_config
}
)
assert response.ok
response = request_with_validation(
api='/collections/{collection_name}',
method="GET",
path_params={'collection_name': collection_name},
)
assert response.ok
def create_from_collection(collection_name, source_collection_name, size, distance, on_disk_vectors, on_disk_payload):
response = request_with_validation(
api='/collections/{collection_name}',
method="DELETE",
path_params={'collection_name': collection_name},
)
assert response.ok
response = request_with_validation(
api='/collections/{collection_name}',
method="PUT",
path_params={'collection_name': collection_name},
body={
"vectors": {
"size": size,
"distance": distance,
"on_disk": on_disk_vectors,
},
"on_disk_payload": on_disk_payload,
"init_from": {
"collection": source_collection_name
}
}
)
return response
def create_multi_from_collection(collection_name, source_collection_name, vectors_config):
response = request_with_validation(
api='/collections/{collection_name}',
method="DELETE",
path_params={'collection_name': collection_name},
)
assert response.ok
response = request_with_validation(
api='/collections/{collection_name}',
method="PUT",
path_params={'collection_name': collection_name},
body={
"vectors": vectors_config,
"init_from": {
"collection": source_collection_name
}
}
)
return response
@pytest.mark.parametrize("ok,source,size,distance", [
(True, source_collection_name, 4, 'Dot'), # ok
(False, "i-do-not-exist", 4, 'Dot'), # fail: non existing source collection
(False, source_collection_name, 8, 'Dot'), # fail: bad size
(False, source_collection_name, 4, 'Cosine'), # fail: bad distance
])
def test_init_from_collection(
ok,
source,
size,
distance,
on_disk_vectors,
on_disk_payload,
):
advanced_collection_setup(source_collection_name, 4, 'Dot', on_disk_vectors, on_disk_payload)
response = create_from_collection(collection_name, source, size, distance, on_disk_vectors, on_disk_payload)
assert response.ok == ok
def test_init_from_collection_multivec(on_disk_vectors):
config = {
"image": {
"size": 4,
"distance": "Dot",
"on_disk": on_disk_vectors,
},
"audio": {
"size": 8,
"distance": "Dot",
"on_disk": on_disk_vectors,
}
}
# test successful init from collection
advanced_collection_multi_setup(source_collection_name, config)
response = create_multi_from_collection(collection_name, source_collection_name, config)
assert response.ok
larger_config = {
"image": {
"size": 4,
"distance": "Dot",
"on_disk": on_disk_vectors,
},
"audio": {
"size": 8,
"distance": "Dot",
"on_disk": on_disk_vectors,
},
"video": { # new vector
"size": 16,
"distance": "Dot",
"on_disk": on_disk_vectors,
}
}
# test successful init from collection (source is subset from target)
advanced_collection_multi_setup(source_collection_name, config)
response = create_multi_from_collection(collection_name, source_collection_name, larger_config)
assert response.ok
larger_config = {
"image": {
"size": 4,
"distance": "Dot",
"on_disk": on_disk_vectors,
},
"audio": {
"size": 9, # bad size
"distance": "Dot",
"on_disk": on_disk_vectors,
},
"video": {
"size": 16,
"distance": "Dot",
"on_disk": on_disk_vectors,
}
}
# test failed init from collection (target changes size of existing vector)
advanced_collection_multi_setup(source_collection_name, config)
response = create_multi_from_collection(collection_name, source_collection_name, larger_config)
assert not response.ok