Skip to content

GitLab

Switch branch/tag
Find file Normal viewHistoryPermalink
test_s3.py 3.38 KB
Newer Older
icecraft's avatar
feat: add [figure | table] match [caption | footnote] match algorithm v2  
icecraft committed
1
2
3
4
5
6
7
8
9
10
11 
import json
import os

import pytest

from magic_pdf.data.data_reader_writer import S3DataReader, S3DataWriter


@pytest.mark.skipif(
    os.getenv('S3_ACCESS_KEY', None) is None, reason='need s3 config!'
)
icecraft's avatar
feat: add more unittest  
icecraft committed
12 
def test_s3_reader_writer():
icecraft's avatar
feat: add [figure | table] match [caption | footnote] match algorithm v2  
icecraft committed
13
14
15
16
17
18
19
20 
    """test multi bucket s3 reader writer must config s3 config in the
    environment export S3_BUCKET=xxx export S3_ACCESS_KEY=xxx export
    S3_SECRET_KEY=xxx export S3_ENDPOINT=xxx."""
    bucket = os.getenv('S3_BUCKET', '')
    ak = os.getenv('S3_ACCESS_KEY', '')
    sk = os.getenv('S3_SECRET_KEY', '')
    endpoint_url = os.getenv('S3_ENDPOINT', '')
icecraft's avatar
feat: add more unittest  
icecraft committed
21
22 
    reader = S3DataReader('', bucket, ak, sk, endpoint_url)
    writer = S3DataWriter('', bucket, ak, sk, endpoint_url)
icecraft's avatar
feat: add [figure | table] match [caption | footnote] match algorithm v2  
icecraft committed
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53 

    bits = reader.read('meta-index/scihub/v001/scihub/part-66210c190659-000026.jsonl')

    assert bits == reader.read(
        f's3://{bucket}/meta-index/scihub/v001/scihub/part-66210c190659-000026.jsonl'
    )

    bits = reader.read(
        'meta-index/scihub/v001/scihub/part-66210c190659-000026.jsonl?bytes=566,713'
    )
    assert bits == reader.read_at(
        'meta-index/scihub/v001/scihub/part-66210c190659-000026.jsonl', 566, 713
    )
    assert len(json.loads(bits)) > 0

    writer.write_string(
        'unittest/data/data_reader_writer/multi_bucket_s3_data/test01.txt', 'abc'
    )

    assert 'abc'.encode() == reader.read(
        'unittest/data/data_reader_writer/multi_bucket_s3_data/test01.txt'
    )

    writer.write(
        f'{bucket}/unittest/data/data_reader_writer/multi_bucket_s3_data/test02.txt',
        '123'.encode(),
    )

    assert '123'.encode() == reader.read(
        'unittest/data/data_reader_writer/multi_bucket_s3_data/test02.txt'
    )
icecraft's avatar
feat: add more unittest  
icecraft committed
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106 


@pytest.mark.skipif(
    os.getenv('S3_ACCESS_KEY', None) is None, reason='need s3 config!'
)
def test_s3_reader_writer_with_prefix():
    """test multi bucket s3 reader writer must config s3 config in the
    environment export S3_BUCKET=xxx export S3_ACCESS_KEY=xxx export
    S3_SECRET_KEY=xxx export S3_ENDPOINT=xxx."""
    bucket = os.getenv('S3_BUCKET', '')
    ak = os.getenv('S3_ACCESS_KEY', '')
    sk = os.getenv('S3_SECRET_KEY', '')
    endpoint_url = os.getenv('S3_ENDPOINT', '')

    prefix = 'meta-index'

    reader = S3DataReader(prefix, bucket, ak, sk, endpoint_url)
    writer = S3DataWriter(prefix, bucket, ak, sk, endpoint_url)

    bits = reader.read('scihub/v001/scihub/part-66210c190659-000026.jsonl')

    assert bits == reader.read(
        f's3://{bucket}/{prefix}/scihub/v001/scihub/part-66210c190659-000026.jsonl'
    )

    bits = reader.read(
        'scihub/v001/scihub/part-66210c190659-000026.jsonl?bytes=566,713'
    )
    assert bits == reader.read_at(
        'scihub/v001/scihub/part-66210c190659-000026.jsonl', 566, 713
    )
    assert len(json.loads(bits)) > 0

    writer.write_string(
        'unittest/data/data_reader_writer/multi_bucket_s3_data/test01.txt', 'abc'
    )

    assert 'abc'.encode() == reader.read(
        'unittest/data/data_reader_writer/multi_bucket_s3_data/test01.txt'
    )

    assert 'abc'.encode() == reader.read(
        f's3://{bucket}/{prefix}/unittest/data/data_reader_writer/multi_bucket_s3_data/test01.txt'
    )

    writer.write(
        f'{bucket}/{prefix}/unittest/data/data_reader_writer/multi_bucket_s3_data/test02.txt',
        '123'.encode(),
    )

    assert '123'.encode() == reader.read(
        'unittest/data/data_reader_writer/multi_bucket_s3_data/test02.txt'
    )