Refactor LargeFile

This commit is contained in:
Andras Schmelczer 2022-06-03 10:44:01 +02:00
parent 01510ee835
commit c6c236c493
18 changed files with 583 additions and 396 deletions

View file

View file

@ -0,0 +1,26 @@
import unittest
from src.great_ai.large_file.helper import human_readable_to_byte
class TestHumanReadableToByte(unittest.TestCase):
def test_simple_cases(self) -> None:
self.assertEqual(human_readable_to_byte("1KB"), 1024)
self.assertEqual(human_readable_to_byte("2KB"), 2048)
def test_fractions(self) -> None:
self.assertEqual(human_readable_to_byte("0.5KB"), 512)
self.assertEqual(human_readable_to_byte("20.5KB"), 1024 * 20 + 512)
def test_formating(self) -> None:
self.assertEqual(human_readable_to_byte(" 1MB"), 1024 * 1024)
self.assertEqual(human_readable_to_byte(" 2 MB"), 1024 * 1024 * 2)
self.assertEqual(human_readable_to_byte(" 4 MB "), 1024 * 1024 * 4)
self.assertEqual(human_readable_to_byte("8MB "), 1024 * 1024 * 8)
self.assertEqual(human_readable_to_byte(" 1.5 MB "), 1024 * 1024 * 1.5)
def test_casing(self) -> None:
self.assertEqual(human_readable_to_byte("0.5GB"), 0.5 * 1024 * 1024 * 1024)
self.assertEqual(human_readable_to_byte("0.5gB"), 0.5 * 1024 * 1024 * 1024)
self.assertEqual(human_readable_to_byte("0.5Gb"), 0.5 * 1024 * 1024 * 1024)
self.assertEqual(human_readable_to_byte("0.5gb"), 0.5 * 1024 * 1024 * 1024)

View file

@ -0,0 +1,120 @@
import unittest
from pathlib import Path
from typing import Any
from unittest.mock import Mock, patch
import boto3
PATH = Path(__file__).parent.resolve()
from src.great_ai import LargeFileS3
credentials = {
"aws_region_name": "your_region_like_eu-west-2",
"aws_access_key_id": "YOUR_ACCESS_KEY_ID",
"aws_secret_access_key": "YOUR_VERY_SECRET_ACCESS_KEY",
"large_files_bucket_name": "create_a_bucket_and_put_its_name_here",
"aws_endpoint_url": "this is optional, for backblaze, use this: https://s3.us-west-002.backblazeb2.com",
}
class TestLargeFileS3(unittest.TestCase):
def test_uninitialized(self) -> None:
self.assertRaises(ValueError, LargeFileS3, "test-file")
def test_bad_file_modes(self) -> None:
self.assertRaises(ValueError, LargeFileS3, "test-file", "w", version=3)
self.assertRaises(ValueError, LargeFileS3, "test-file", "wb", version=3)
self.assertRaises(ValueError, LargeFileS3, "test-file", "w+r")
self.assertRaises(ValueError, LargeFileS3, "test-file", "test")
@patch.object(boto3, "client")
def test_initialized_with_dict(self, client: Any) -> None:
s3 = Mock()
s3.list_objects_v2 = Mock(
return_value={
"Contents": [
{
"Key": "test-file/0",
"Size": 30,
},
{
"Key": "test-file/1",
"Size": 300,
},
{
"Key": "test-file/2",
"Size": 187,
},
]
}
)
boto3.client = Mock(return_value=s3)
LargeFileS3.configure_credentials(
aws_region_name=credentials["aws_region_name"],
aws_access_key_id=credentials["aws_access_key_id"],
aws_secret_access_key=credentials["aws_secret_access_key"],
large_files_bucket_name=credentials["large_files_bucket_name"],
aws_endpoint_url=credentials["aws_endpoint_url"],
)
lf = LargeFileS3("test-file")
boto3.client.assert_called_once_with(
"s3",
aws_access_key_id=credentials["aws_access_key_id"],
aws_secret_access_key=credentials["aws_secret_access_key"],
region_name=credentials["aws_region_name"],
endpoint_url=credentials["aws_endpoint_url"],
)
s3.list_objects_v2.assert_called_once_with(
Bucket=credentials["large_files_bucket_name"], Prefix="test-file"
)
self.assertEqual(lf._version, 2)
self.assertEqual(lf._local_name, "test-file-2")
@patch.object(boto3, "client")
def test_initialized_with_file(self, client: Any) -> None:
s3 = Mock()
s3.list_objects_v2 = Mock(
return_value={
"Contents": [
{
"Key": "test-file/0",
"Size": 30,
},
{
"Key": "test-file/1",
"Size": 300,
},
{
"Key": "test-file/2",
"Size": 187,
},
]
}
)
boto3.client = Mock(return_value=s3)
LargeFileS3.configure_credentials_from_file(PATH / "../../example_secrets.ini")
lf = LargeFileS3("test-file")
boto3.client.assert_called_once_with(
"s3",
aws_access_key_id=credentials["aws_access_key_id"],
aws_secret_access_key=credentials["aws_secret_access_key"],
region_name=credentials["aws_region_name"],
endpoint_url=credentials["aws_endpoint_url"],
)
assert s3.list_objects_v2.called
s3.list_objects_v2.assert_called_once_with(
Bucket=credentials["large_files_bucket_name"], Prefix="test-file"
)
self.assertEqual(lf._version, 2)
self.assertEqual(lf._local_name, "test-file-2")