def test_s3_iter_moto(self):
"""Are S3 files iterated over correctly?"""
# a list of strings to test with
expected = [b"*" * 5 * 1024**2] + [b'0123456789'] * 1024 + [b"test"]
# create fake bucket and fake key
conn = boto.connect_s3()
conn.create_bucket("mybucket")
# lower the multipart upload size, to speed up these tests
smart_open_lib.S3_MIN_PART_SIZE = 5 * 1024**2
with smart_open.smart_open("s3://mybucket/mykey", "wb") as fout:
# write a single huge line (=full multipart upload)
fout.write(expected[0] + b'\n')
# write lots of small lines
for lineno, line in enumerate(expected[1:-1]):
fout.write(line + b'\n')
# ...and write the last line too, no newline at the end
fout.write(expected[-1])
# connect to fake s3 and read from the fake key we filled above
smart_open_object = smart_open.S3OpenRead(
smart_open.ParseUri("s3://mybucket/mykey"))
output = [line.rstrip(b'\n') for line in smart_open_object]
self.assertEqual(output, expected)
# same thing but using a context manager
with smart_open.S3OpenRead(smart_open.ParseUri(
"s3://mybucket/mykey")) as smart_open_object:
output = [line.rstrip(b'\n') for line in smart_open_object]
self.assertEqual(output, expected)
def test_hdfs(self, mock_subprocess):
"""Is HDFS write called correctly"""
smart_open_object = smart_open.HdfsOpenWrite(smart_open.ParseUri("hdfs:///tmp/test.txt"))
smart_open_object.write("test")
# called with the correct params?
mock_subprocess.Popen.assert_called_with(["hdfs","dfs","-put","-f","-","/tmp/test.txt"], stdin=mock_subprocess.PIPE)
# second possibility of schema
smart_open_object = smart_open.HdfsOpenWrite(smart_open.ParseUri("hdfs://tmp/test.txt"))
smart_open_object.write("test")
mock_subprocess.Popen.assert_called_with(["hdfs","dfs","-put","-f","-","/tmp/test.txt"], stdin=mock_subprocess.PIPE)
def test_webhdfs_uri(self):
"""Do webhdfs URIs parse correctly"""
# valid uri, no query
parsed_uri = smart_open.ParseUri("webhdfs://host:port/path/file")
self.assertEqual(parsed_uri.scheme, "webhdfs")
self.assertEqual(parsed_uri.uri_path, "host:port/webhdfs/v1/path/file")
# valid uri, with query
parsed_uri = smart_open.ParseUri("webhdfs://host:port/path/file?query_part_1&query_part2")
self.assertEqual(parsed_uri.scheme, "webhdfs")
self.assertEqual(parsed_uri.uri_path, "host:port/webhdfs/v1/path/file?query_part_1&query_part2")
def test_hdfs(self, mock_subprocess):
"""Is HDFS line iterator called correctly?"""
mock_subprocess.PIPE.return_value = "test"
smart_open_object = smart_open.HdfsOpenRead(smart_open.ParseUri("hdfs:///tmp/test.txt"))
smart_open_object.__iter__()
# called with the correct params?
mock_subprocess.Popen.assert_called_with(["hdfs", "dfs", "-cat", "/tmp/test.txt"], stdout=mock_subprocess.PIPE)
# second possibility of schema
smart_open_object = smart_open.HdfsOpenRead(smart_open.ParseUri("hdfs://tmp/test.txt"))
smart_open_object.__iter__()
mock_subprocess.Popen.assert_called_with(["hdfs", "dfs", "-cat", "/tmp/test.txt"], stdout=mock_subprocess.PIPE)
def test_scheme(self):
"""Do URIs schemes parse correctly?"""
# supported schemes
for scheme in ("s3", "s3n", "hdfs", "file", "http", "https"):
parsed_uri = smart_open.ParseUri(scheme + "://mybucket/mykey")
self.assertEqual(parsed_uri.scheme, scheme)
# unsupported scheme => NotImplementedError
self.assertRaises(NotImplementedError, smart_open.ParseUri, "foobar://mybucket/mykey")
# unknown scheme => default_scheme
parsed_uri = smart_open.ParseUri("blah blah")
self.assertEqual(parsed_uri.scheme, "file")
def convert(self, value, param, ctx):
uri = smart_open.ParseUri(value)
if uri.scheme == 'file':
return super(SmartFile, self).convert(uri.uri_path, param, ctx)
else:
return smart_open.smart_open(value, self.mode)
def test_write(self):
def request_callback(request):
resp_body = ""
headers = {'location': 'http://127.0.0.1:8440/file'}
return (307, headers, resp_body)
responses.add_callback(responses.PUT,
"http://127.0.0.1:8440/webhdfs/v1/path/file",
callback=request_callback)
responses.add(responses.PUT, "http://127.0.0.1:8440/file", status=201)
smart_open_object = smart_open.WebHdfsOpenWrite(
smart_open.ParseUri("webhdfs://127.0.0.1:8440/path/file"))
def write_callback(request):
assert request.body == u"žluťoučký koníček".encode('utf8')
headers = {}
return (200, headers, "")
test_string = u"žluťoučký koníček".encode('utf8')
responses.add_callback(responses.POST,
"http://127.0.0.1:8440/webhdfs/v1/path/file",
callback=request_callback)
responses.add_callback(responses.POST,
"http://127.0.0.1:8440/file",
callback=write_callback)
smart_open_object.write(test_string)
smart_open_object.close()
assert len(responses.calls) == 4
assert responses.calls[
2].request.url == "http://127.0.0.1:8440/webhdfs/v1/path/file?op=APPEND"
assert responses.calls[3].request.url == "http://127.0.0.1:8440/file"
def test_webhdfs(self):
"""Is webhdfs line iterator called correctly"""
responses.add(responses.GET, "http://127.0.0.1:8440/webhdfs/v1/path/file", body='line1\nline2')
smart_open_object = smart_open.WebHdfsOpenRead(smart_open.ParseUri("webhdfs://127.0.0.1:8440/path/file"))
iterator = iter(smart_open_object)
self.assertEqual(next(iterator).decode("utf-8"), "line1")
self.assertEqual(next(iterator).decode("utf-8"), "line2")
def test_rw_encoding(self):
"""Should read and write text, respecting encodings, etc."""
conn = boto.connect_s3()
conn.create_bucket("bucket")
uri = smart_open.ParseUri("s3://bucket/key")
text = u"расцветали яблони и груши"
with smart_open.s3_open_uri(uri, "w", encoding="koi8-r") as fout:
fout.write(text)
with smart_open.s3_open_uri(uri, "r", encoding="koi8-r") as fin:
self.assertEqual(text, fin.read())
with smart_open.s3_open_uri(uri, "rb") as fin:
self.assertEqual(text.encode("koi8-r"), fin.read())
with smart_open.s3_open_uri(uri, "r", encoding="euc-jp") as fin:
self.assertRaises(UnicodeDecodeError, fin.read)
with smart_open.s3_open_uri(uri,
"r",
encoding="euc-jp",
errors="replace") as fin:
fin.read()
def test_https_readline(self):
"""Does https readline method work correctly"""
responses.add(responses.GET,
"https://127.0.0.1/index.html",
body='line1\nline2')
smart_open_object = smart_open.HttpOpenRead(
smart_open.ParseUri("https://127.0.0.1/index.html"))
self.assertEqual(smart_open_object.readline().decode("utf-8"), "line1")
def test_http_pass(self):
"""Does http authentication work correctly"""
responses.add(responses.GET, "http://127.0.0.1/index.html", body='line1\nline2')
_ = smart_open.HttpOpenRead(smart_open.ParseUri("http://127.0.0.1/index.html"), user='******', password='******')
self.assertEquals(len(responses.calls), 1)
actual_request = responses.calls[0].request
self.assert_('Authorization' in actual_request.headers)
self.assert_(actual_request.headers['Authorization'].startswith('Basic '))
def test_gzip_write_mode(self):
"""Should always open in binary mode when writing through a codec."""
s3 = boto3.resource('s3')
s3.create_bucket(Bucket='bucket')
uri = smart_open.ParseUri("s3://bucket/key.gz")
with mock.patch('smart_open.smart_open_s3.open') as mock_open:
smart_open.s3_open_uri(uri, "wb")
mock_open.assert_called_with('bucket', 'key.gz', 'wb')
def test_s3_boto(self, mock_s3_iter_lines, mock_boto):
"""Is S3 line iterator called correctly?"""
# no credentials
smart_open_object = smart_open.S3OpenRead(smart_open.ParseUri("s3://mybucket/mykey"))
smart_open_object.__iter__()
mock_boto.connect_s3.assert_called_with(aws_access_key_id=None, aws_secret_access_key=None)
# with credential
smart_open_object = smart_open.S3OpenRead(smart_open.ParseUri("s3://access_id:access_secret@mybucket/mykey"))
smart_open_object.__iter__()
mock_boto.connect_s3.assert_called_with(aws_access_key_id="access_id", aws_secret_access_key="access_secret")
# lookup bucket, key; call s3_iter_lines
smart_open_object = smart_open.S3OpenRead(smart_open.ParseUri("s3://access_id:access_secret@mybucket/mykey"))
smart_open_object.__iter__()
mock_boto.connect_s3().get_bucket.assert_called_with("mybucket")
mock_boto.connect_s3().get_bucket().lookup.assert_called_with("mykey")
self.assertTrue(mock_s3_iter_lines.called)
def test_gzip_write_mode(self):
"""Should always open in binary mode when writing through a codec."""
conn = boto.connect_s3()
conn.create_bucket("bucket")
uri = smart_open.ParseUri("s3://bucket/key.gz")
with mock.patch('smart_open.smart_open_s3.open') as mock_open:
smart_open.s3_open_uri(uri, "wb")
mock_open.assert_called_with('bucket', 'key.gz', 'wb')
def test_webhdfs_read(self):
"""Does webhdfs read method work correctly"""
responses.add(responses.GET,
"http://127.0.0.1:8440/webhdfs/v1/path/file",
body='line1\nline2')
smart_open_object = smart_open.WebHdfsOpenRead(
smart_open.ParseUri("webhdfs://127.0.0.1:8440/path/file"))
self.assertEqual(smart_open_object.read().decode("utf-8"),
"line1\nline2")
def test_http_gz(self):
"""Can open gzip via http?"""
fpath = os.path.join(CURR_DIR, 'test_data/crlf_at_1k_boundary.warc.gz')
data = open(fpath, 'rb').read()
responses.add(responses.GET, "http://127.0.0.1/data.gz", body=data)
smart_open_object = smart_open.HttpOpenRead(
smart_open.ParseUri("http://127.0.0.1/data.gz"))
m = hashlib.md5(smart_open_object.read())
# decompress the gzip and get the same md5 hash
self.assertEqual(m.hexdigest(), '18473e60f8c7c98d29d65bf805736a0d')
def test_s3_uri(self):
"""Do S3 URIs parse correctly?"""
# correct uri without credentials
parsed_uri = smart_open.ParseUri("s3://mybucket/mykey")
self.assertEqual(parsed_uri.scheme, "s3")
self.assertEqual(parsed_uri.bucket_id, "mybucket")
self.assertEqual(parsed_uri.key_id, "mykey")
self.assertEqual(parsed_uri.access_id, None)
self.assertEqual(parsed_uri.access_secret, None)
# correct uri, key contains slash
parsed_uri = smart_open.ParseUri("s3://mybucket/mydir/mykey")
self.assertEqual(parsed_uri.scheme, "s3")
self.assertEqual(parsed_uri.bucket_id, "mybucket")
self.assertEqual(parsed_uri.key_id, "mydir/mykey")
self.assertEqual(parsed_uri.access_id, None)
self.assertEqual(parsed_uri.access_secret, None)
# correct uri with credentials
parsed_uri = smart_open.ParseUri(
"s3://ACCESSID456:acces/sse_cr-et@mybucket/mykey")
self.assertEqual(parsed_uri.scheme, "s3")
self.assertEqual(parsed_uri.bucket_id, "mybucket")
self.assertEqual(parsed_uri.key_id, "mykey")
self.assertEqual(parsed_uri.access_id, "ACCESSID456")
self.assertEqual(parsed_uri.access_secret, "acces/sse_cr-et")
# correct uri, contains credentials
parsed_uri = smart_open.ParseUri(
"s3://accessid:access/secret@mybucket/mykey")
self.assertEqual(parsed_uri.scheme, "s3")
self.assertEqual(parsed_uri.bucket_id, "mybucket")
self.assertEqual(parsed_uri.key_id, "mykey")
self.assertEqual(parsed_uri.access_id, "accessid")
self.assertEqual(parsed_uri.access_secret, "access/secret")
# incorrect uri - only one '@' in uri is allowed
self.assertRaises(RuntimeError, smart_open.ParseUri,
"s3://access_id@access_secret@mybucket/mykey")
def test_initialize_write(self):
def request_callback(request):
resp_body = ""
headers = {'location': 'http://127.0.0.1:8440/file'}
return (307, headers, resp_body)
responses.add_callback(responses.PUT, "http://127.0.0.1:8440/webhdfs/v1/path/file", callback=request_callback)
responses.add(responses.PUT, "http://127.0.0.1:8440/file", status=201)
smart_open_object = smart_open.WebHdfsOpenWrite(smart_open.ParseUri("webhdfs://127.0.0.1:8440/path/file"))
assert len(responses.calls) == 2
path, params = responses.calls[0].request.url.split("?")
assert path == "http://127.0.0.1:8440/webhdfs/v1/path/file"
assert params == "overwrite=True&op=CREATE" or params == "op=CREATE&overwrite=True"
assert responses.calls[1].request.url == "http://127.0.0.1:8440/file"
def test_gzip_read_mode(self):
"""Should always open in binary mode when reading through a codec."""
conn = boto.connect_s3()
conn.create_bucket("bucket")
uri = smart_open.ParseUri("s3://bucket/key.gz")
text = u"если-б я был султан и имел трёх жён, то тройной красотой был бы окружён"
with smart_open.s3_open_uri(uri, "wb") as fout:
fout.write(text.encode("utf-8"))
with mock.patch('smart_open.smart_open_s3.open') as mock_open:
smart_open.s3_open_uri(uri, "r")
mock_open.assert_called_with('bucket', 'key.gz', 'rb')
def stream_other(self, path, fn):
"""stream from local folders call a function
Args:
folders (list): list of folders
fn (function): function to call
"""
parsed_uri = smart_open.ParseUri(path)
if parsed_uri.scheme in ("file", ) and os.path.isdir(
path): # stream each file from directory
for f in glob.glob(path + "/*"):
for line in smart_open.smart_open(f):
fn(line)
else: #let smart_open handle streaming of file from location
for line in smart_open.smart_open(path):
fn(line)
def test_r(self):
"""Reading a UTF string should work."""
conn = boto.connect_s3()
conn.create_bucket("bucket")
bucket = conn.get_bucket("bucket")
key = boto.s3.key.Key(bucket)
key.key = "key"
text = u"физкульт-привет!"
key.set_contents_from_string(text.encode("utf-8"))
with smart_open.s3_open_key(key, "r") as fin:
self.assertEqual(fin.read(), u"физкульт-привет!")
parsed_uri = smart_open.ParseUri("s3://bucket/key")
with smart_open.s3_open_uri(parsed_uri, "r") as fin:
self.assertEqual(fin.read(), u"физкульт-привет!")
def test_r(self):
"""Reading a UTF string should work."""
text = u"физкульт-привет!"
s3 = boto3.resource('s3')
s3.create_bucket(Bucket='bucket')
key = s3.Object('bucket', 'key')
key.put(Body=text.encode('utf-8'))
with smart_open.s3_open_key(key, "rb") as fin:
self.assertEqual(fin.read(), text.encode('utf-8'))
with smart_open.s3_open_key(key, "r", encoding='utf-8') as fin:
self.assertEqual(fin.read(), text)
uri = smart_open.ParseUri("s3://bucket/key")
with smart_open.s3_open_uri(uri, "r", encoding='utf-8') as fin:
self.assertEqual(fin.read(), text)
def test_s3_read_moto(self):
"""Are S3 files read correctly?"""
conn = boto.connect_s3()
conn.create_bucket("mybucket")
# write some bogus key so we can check it below
content = "hello wořld\nhow are you?"
with smart_open.smart_open("s3://mybucket/mykey", "wb") as fout:
fout.write(content)
smart_open_object = smart_open.S3OpenRead(smart_open.ParseUri("s3://mybucket/mykey"))
self.assertEqual(content[:6], smart_open_object.read(6))
self.assertEqual(content[6:14], smart_open_object.read(8)) # ř is 2 bytes
# make sure iteration does not affect read()
for line in smart_open_object:
pass
self.assertEqual(content[14:], smart_open_object.read()) # read the rest
def test_s3_seek_moto(self):
"""Does seeking in S3 files work correctly?"""
conn = boto.connect_s3()
conn.create_bucket("mybucket")
# write some bogus key so we can check it below
content = "hello wořld\nhow are you?"
with smart_open.smart_open("s3://mybucket/mykey", "wb") as fout:
fout.write(content)
smart_open_object = smart_open.S3OpenRead(smart_open.ParseUri("s3://mybucket/mykey"))
self.assertEqual(content[:6], smart_open_object.read(6))
self.assertEqual(content[6:14], smart_open_object.read(8)) # ř is 2 bytes
smart_open_object.seek(0)
self.assertEqual(content, smart_open_object.read()) # no size given => read whole file
smart_open_object.seek(0)
self.assertEqual(content, smart_open_object.read(-1)) # same thing
def test_rw_gzip(self):
"""Should read/write gzip files, implicitly and explicitly."""
conn = boto.connect_s3()
conn.create_bucket("bucket")
uri = smart_open.ParseUri("s3://bucket/key.gz")
text = u"не слышны в саду даже шорохи"
with smart_open.s3_open_uri(uri, "wb") as fout:
fout.write(text.encode("utf-8"))
#
# Check that what we've created is a gzip.
#
with smart_open.s3_open_uri(uri, "rb", ignore_extension=True) as fin:
gz = gzip.GzipFile(fileobj=fin)
self.assertEqual(gz.read().decode("utf-8"), text)
#
# We should be able to read it back as well.
#
with smart_open.s3_open_uri(uri, "rb") as fin:
self.assertEqual(fin.read().decode("utf-8"), text)
def test_http_bz2(self):
"""Can open bz2 via http?"""
test_string = b'Hello World Compressed.'
test_file = tempfile.NamedTemporaryFile('wb',
suffix='.bz2',
delete=False).name
with smart_open.smart_open(test_file, 'wb') as outfile:
outfile.write(test_string)
with open(test_file, 'rb') as infile:
compressed_data = infile.read()
if os.path.isfile(test_file):
os.unlink(test_file)
responses.add(responses.GET,
"http://127.0.0.1/data.bz2",
body=compressed_data)
smart_open_object = smart_open.HttpOpenRead(
smart_open.ParseUri("http://127.0.0.1/data.bz2"))
# decompress the gzip and get the same md5 hash
self.assertEqual(smart_open_object.read(), test_string)
def test_bad_mode(self):
"""Bad mode should raise and exception."""
uri = smart_open.ParseUri("s3://bucket/key")
self.assertRaises(NotImplementedError, smart_open.s3_open_uri, uri,
"x")
def test_webhdfs_uri(self):
"""Do webhdfs URIs parse correctly"""
parsed_uri = smart_open.ParseUri("webhdfs://host:port/path/file")
self.assertEqual(parsed_uri.scheme, "webhdfs")
self.assertEqual(parsed_uri.uri_path, "host:port/webhdfs/v1/path/file")
