mirror of
https://github.com/qdrant/qdrant.git
synced 2026-07-26 12:41:04 -05:00
9d67c6325b
Collection name under test is equal to the test module name, without `.py` suffix. * Helps by debugging/tracing failed tests and find relevant logs lines in qdrant log files * Opens up a possibility to run tests in parallel, given that there are no data sharing between test modules Change details: * defined module scoped `collection_name` fixture in `conftest.py` * removed `collection_name` module variable * each test signature modified to declare the dependency to `collection_name` fixture * `@pytest.mark.parametrize` migrated to `@pytest-cases.parametrize` in cases when `collection_name` was used as the value
337 lines
10 KiB
Python
337 lines
10 KiB
Python
import datetime
|
|
import math
|
|
import random
|
|
|
|
import pytest
|
|
from more_itertools import batched
|
|
|
|
from .helpers.collection_setup import basic_collection_setup, drop_collection
|
|
from .helpers.helpers import request_with_validation
|
|
|
|
total_points = 300
|
|
|
|
|
|
def upsert_points(collection_name, amount=100):
|
|
def maybe_repeated():
|
|
"""Descending sequence of possibly repeated floats"""
|
|
repeated_float = float(amount)
|
|
while True:
|
|
repeated_float = repeated_float if random.random() > 0.5 else repeated_float - 1.0
|
|
yield repeated_float
|
|
|
|
maybe_repeated_generator = maybe_repeated()
|
|
|
|
def date():
|
|
date = datetime.datetime.now()
|
|
while True:
|
|
if random.random() > 0.5:
|
|
date += datetime.timedelta(days=1)
|
|
yield date.isoformat() + "Z" # RFC-3339 format
|
|
|
|
date_generator = date()
|
|
|
|
points = [
|
|
{
|
|
"id": i,
|
|
"vector": [0.1 * i] * 4,
|
|
"payload": {
|
|
"city": "London" if i % 3 == 0 else "Moscow",
|
|
"is_middle_split": i > amount * 0.25 and i < amount * 0.75,
|
|
"price": float(amount - i),
|
|
"payload_id": i,
|
|
"multi_id": [i, amount - i + 1],
|
|
"maybe_repeated_float": next(maybe_repeated_generator),
|
|
"date_rfc3339": next(date_generator),
|
|
"date_simple": next(date_generator).split("T")[0],
|
|
},
|
|
}
|
|
for i in range(amount)
|
|
]
|
|
|
|
for batch in batched(points, 50):
|
|
response = request_with_validation(
|
|
api="/collections/{collection_name}/points",
|
|
method="PUT",
|
|
path_params={"collection_name": collection_name},
|
|
query_params={"wait": "true"},
|
|
body={"points": list(batch)},
|
|
)
|
|
assert response.ok
|
|
|
|
|
|
def create_payload_index(collection_name, field_name, field_schema):
|
|
response = request_with_validation(
|
|
api="/collections/{collection_name}/index",
|
|
method="PUT",
|
|
path_params={"collection_name": collection_name},
|
|
query_params={"wait": "true"},
|
|
body={"field_name": field_name, "field_schema": field_schema},
|
|
)
|
|
assert response.ok, response.json()
|
|
|
|
|
|
@pytest.fixture(autouse=True, scope="module")
|
|
def setup(on_disk_vectors, collection_name):
|
|
basic_collection_setup(collection_name=collection_name, on_disk_vectors=on_disk_vectors)
|
|
upsert_points(collection_name=collection_name, amount=total_points)
|
|
create_payload_index(
|
|
collection_name=collection_name, field_name="city", field_schema="keyword"
|
|
)
|
|
create_payload_index(
|
|
collection_name=collection_name, field_name="is_middle_split", field_schema="bool"
|
|
)
|
|
create_payload_index(collection_name=collection_name, field_name="price", field_schema="float")
|
|
create_payload_index(
|
|
collection_name=collection_name, field_name="maybe_repeated_float", field_schema="float"
|
|
)
|
|
create_payload_index(
|
|
collection_name=collection_name, field_name="payload_id", field_schema="integer"
|
|
)
|
|
create_payload_index(
|
|
collection_name=collection_name, field_name="multi_id", field_schema="integer"
|
|
)
|
|
create_payload_index(
|
|
collection_name=collection_name, field_name="date_rfc3339", field_schema="datetime"
|
|
)
|
|
create_payload_index(
|
|
collection_name=collection_name, field_name="date_simple", field_schema="datetime"
|
|
)
|
|
yield
|
|
drop_collection(collection_name=collection_name)
|
|
|
|
|
|
def test_order_by_int_ascending(collection_name):
|
|
response = request_with_validation(
|
|
api="/collections/{collection_name}/points/scroll",
|
|
method="POST",
|
|
path_params={"collection_name": collection_name},
|
|
body={
|
|
"order_by": {"key": "payload_id", "direction": "asc"},
|
|
"limit": 5,
|
|
},
|
|
)
|
|
assert response.ok, response.json()
|
|
|
|
result = response.json()["result"]
|
|
assert len(result["points"]) == 5
|
|
|
|
ids = [x["id"] for x in result["points"]]
|
|
assert [0, 1, 2, 3, 4] == ids
|
|
# Offset is not supported for order_by
|
|
assert result["next_page_offset"] == None
|
|
|
|
|
|
def test_order_by_int_descending(collection_name):
|
|
response = request_with_validation(
|
|
api="/collections/{collection_name}/points/scroll",
|
|
method="POST",
|
|
path_params={"collection_name": collection_name},
|
|
body={
|
|
"order_by": {"key": "payload_id", "direction": "desc"},
|
|
"limit": 5,
|
|
},
|
|
)
|
|
assert response.ok, response.json()
|
|
|
|
result = response.json()["result"]
|
|
assert len(result["points"]) == 5
|
|
|
|
ids = [x["id"] for x in result["points"]]
|
|
|
|
# We expect the last ids
|
|
expected_ids = [total_points - (i + 1) for i in range(5)]
|
|
assert expected_ids == ids
|
|
# Offset is not supported for order_by
|
|
assert result["next_page_offset"] == None
|
|
|
|
|
|
def paginate_whole_collection(collection_name, key, direction, must=None):
|
|
limit = 23
|
|
pages = 0
|
|
points_count = 0
|
|
points_set = set()
|
|
last_value = None
|
|
last_value_ids = set()
|
|
start_from = None
|
|
|
|
# Get filtered total points
|
|
response = request_with_validation(
|
|
api="/collections/{collection_name}/points/count",
|
|
method="POST",
|
|
path_params={"collection_name": collection_name},
|
|
body={
|
|
"filter": {"must": must},
|
|
"exact": True,
|
|
},
|
|
)
|
|
assert response.ok, response.json()
|
|
|
|
expected_points = response.json()["result"]["count"]
|
|
|
|
while True:
|
|
response = request_with_validation(
|
|
api="/collections/{collection_name}/points/scroll",
|
|
method="POST",
|
|
path_params={"collection_name": collection_name},
|
|
body={
|
|
"order_by": {"key": key, "direction": direction, "start_from": start_from},
|
|
"limit": limit,
|
|
"filter": {
|
|
"must": must,
|
|
"must_not": [{"has_id": [id_ for id_ in last_value_ids]}],
|
|
},
|
|
},
|
|
)
|
|
assert response.ok, response.json()
|
|
|
|
points = response.json()["result"]["points"]
|
|
|
|
if len(points) > 0:
|
|
last_value = points[-1]["payload"][key]
|
|
|
|
# Exclude the ids we've already seen for the start_from value.
|
|
# This is what we expect the users to do in order to paginate with order_by
|
|
if start_from != last_value:
|
|
last_value_ids.clear()
|
|
start_from = last_value
|
|
|
|
same_value_points = [
|
|
point["id"] for point in points if point["payload"][key] == last_value
|
|
]
|
|
last_value_ids.update(same_value_points)
|
|
|
|
points_len = len(response.json()["result"]["points"])
|
|
|
|
points_count += points_len
|
|
pages += 1
|
|
|
|
# Check no duplicates
|
|
for point in points:
|
|
assert point["id"] not in points_set
|
|
points_set.add(point["id"])
|
|
|
|
if points_len < limit:
|
|
break
|
|
|
|
try:
|
|
assert math.ceil(expected_points / limit) == pages
|
|
assert expected_points == points_count
|
|
except AssertionError:
|
|
# Check which points we're missing
|
|
response = request_with_validation(
|
|
api="/collections/{collection_name}/points/scroll",
|
|
method="POST",
|
|
path_params={"collection_name": collection_name},
|
|
body={
|
|
"limit": total_points,
|
|
"filter": {"must": must},
|
|
},
|
|
)
|
|
assert response.ok, response.json()
|
|
filtered_points = set([point["id"] for point in response.json()["result"]["points"]])
|
|
|
|
assert filtered_points == points_set, f"Missing points: {filtered_points - points_set}"
|
|
|
|
|
|
@pytest.mark.parametrize(
|
|
"key, direction",
|
|
[
|
|
("payload_id", "asc"),
|
|
("payload_id", "desc"),
|
|
("price", "asc"),
|
|
("price", "desc"),
|
|
("maybe_repeated_float", "asc"),
|
|
("maybe_repeated_float", "desc"),
|
|
("date_rfc3339", "asc"),
|
|
("date_rfc3339", "desc"),
|
|
("date_simple", "asc"),
|
|
("date_simple", "desc"),
|
|
],
|
|
)
|
|
@pytest.mark.timeout(60) # possibly break of an infinite loop
|
|
def test_paginate_whole_collection(collection_name, key, direction):
|
|
paginate_whole_collection(collection_name, key, direction)
|
|
|
|
|
|
@pytest.mark.parametrize(
|
|
"key, direction",
|
|
[
|
|
("payload_id", "asc"),
|
|
("payload_id", "desc"),
|
|
("price", "asc"),
|
|
("price", "desc"),
|
|
("maybe_repeated_float", "asc"),
|
|
("maybe_repeated_float", "desc"),
|
|
("date_rfc3339", "asc"),
|
|
("date_rfc3339", "desc"),
|
|
("date_simple", "asc"),
|
|
("date_simple", "desc"),
|
|
],
|
|
)
|
|
@pytest.mark.timeout(60) # possibly break of an infinite loop
|
|
def test_order_by_pagination_with_filters(collection_name, key, direction):
|
|
musts = [
|
|
[
|
|
{
|
|
"key": "city",
|
|
"match": {
|
|
"value": "London",
|
|
},
|
|
}
|
|
],
|
|
[
|
|
{
|
|
"key": "is_middle_split",
|
|
"match": {
|
|
"value": False,
|
|
},
|
|
}
|
|
],
|
|
]
|
|
|
|
for must in musts:
|
|
paginate_whole_collection(collection_name, key, direction, must)
|
|
|
|
|
|
def test_multi_values_appear_multiple_times(collection_name):
|
|
limit = total_points * 2
|
|
|
|
response = request_with_validation(
|
|
api="/collections/{collection_name}/points/scroll",
|
|
method="POST",
|
|
path_params={"collection_name": collection_name},
|
|
body={
|
|
"order_by": "multi_id",
|
|
"limit": limit,
|
|
},
|
|
)
|
|
assert response.ok, response.json()
|
|
|
|
points = response.json()["result"]["points"]
|
|
assert len(points) == limit
|
|
|
|
freqs = {}
|
|
for point in points:
|
|
id_ = point["id"]
|
|
if id_ in freqs:
|
|
freqs[id_] += 1
|
|
else:
|
|
freqs[id_] = 1
|
|
|
|
assert all([count == 2 for count in freqs.values()])
|
|
|
|
|
|
def test_cannot_use_offset_with_order_by(collection_name):
|
|
response = request_with_validation(
|
|
api="/collections/{collection_name}/points/scroll",
|
|
method="POST",
|
|
path_params={"collection_name": collection_name},
|
|
body={
|
|
"order_by": "payload_id",
|
|
"offset": 10,
|
|
"limit": 10,
|
|
},
|
|
)
|
|
assert not response.ok
|
|
assert response.status_code == 400
|