X-Git-Url: https://git.arvados.org/arvados.git/blobdiff_plain/cd708c67b92f2c094e33584bb1fc2c6b4640fb90..5ca261ee945cfea279f6df0364c8311e83cebb44:/sdk/python/tests/test_collections.py diff --git a/sdk/python/tests/test_collections.py b/sdk/python/tests/test_collections.py index a1285608e6..24fedea7f7 100644 --- a/sdk/python/tests/test_collections.py +++ b/sdk/python/tests/test_collections.py @@ -1,7 +1,4 @@ from __future__ import absolute_import -# usage example: -# -# ARVADOS_API_TOKEN=abc ARVADOS_API_HOST=arvados.local python -m unittest discover from builtins import object import arvados @@ -10,6 +7,7 @@ import mock import os import pprint import re +import sys import tempfile import unittest @@ -523,8 +521,8 @@ class CollectionTestMixin(tutil.ApiClientMock): @tutil.skip_sleep class CollectionReaderTestCase(unittest.TestCase, CollectionTestMixin): - def mock_get_collection(self, api_mock, code, body): - body = self.API_COLLECTIONS.get(body) + def mock_get_collection(self, api_mock, code, fixturename): + body = self.API_COLLECTIONS.get(fixturename) self._mock_api_call(api_mock.collections().get, code, body) def api_client_mock(self, status=200): @@ -648,7 +646,7 @@ class CollectionReaderTestCase(unittest.TestCase, CollectionTestMixin): def test_open_collection_file_one_argument(self): client = self.api_client_mock(200) reader = arvados.CollectionReader(self.DEFAULT_UUID, api_client=client) - cfile = reader.open('./foo') + cfile = reader.open('./foo', 'rb') self.check_open_file(cfile, '.', 'foo', 3) def test_open_deep_file(self): @@ -657,7 +655,7 @@ class CollectionReaderTestCase(unittest.TestCase, CollectionTestMixin): self.mock_get_collection(client, 200, coll_name) reader = arvados.CollectionReader( self.API_COLLECTIONS[coll_name]['uuid'], api_client=client) - cfile = reader.open('./subdir2/subdir3/file2_in_subdir3.txt') + cfile = reader.open('./subdir2/subdir3/file2_in_subdir3.txt', 'rb') self.check_open_file(cfile, './subdir2/subdir3', 'file2_in_subdir3.txt', 32) @@ -806,6 +804,38 @@ class CollectionWriterTestCase(unittest.TestCase, CollectionTestMixin): self.assertRaises(arvados.errors.AssertionError, writer.open, 'two') +class CollectionOpenModes(run_test_server.TestCaseWithServers): + + def test_open_binary_modes(self): + c = Collection() + for mode in ['wb', 'wb+', 'ab', 'ab+']: + with c.open('foo', 'wb') as f: + f.write(b'foo') + + def test_open_invalid_modes(self): + c = Collection() + for mode in ['+r', 'aa', '++', 'r+b', 'beer', '', None]: + with self.assertRaises(Exception): + c.open('foo', mode) + + def test_open_text_modes(self): + c = Collection() + with c.open('foo', 'wb') as f: + f.write('foo') + for mode in ['r', 'rt', 'r+', 'rt+', 'w', 'wt', 'a', 'at']: + if sys.version_info >= (3, 0): + with self.assertRaises(NotImplementedError): + c.open('foo', mode) + else: + with c.open('foo', mode) as f: + if mode[0] == 'r' and '+' not in mode: + self.assertEqual('foo', f.read(3)) + else: + f.write('bar') + f.seek(-3, os.SEEK_CUR) + self.assertEqual('bar', f.read(3)) + + class NewCollectionTestCase(unittest.TestCase, CollectionTestMixin): def test_replication_desired_kept_on_load(self): @@ -933,11 +963,15 @@ class NewCollectionTestCase(unittest.TestCase, CollectionTestMixin): c1 = Collection('. 781e5e245d69b566979b86e28d23f2c7+10 0:10:count1.txt\n') c2 = Collection('. 5348b82a029fd9e971a811ce1f71360b+43 0:10:count2.txt\n') d = c2.diff(c1) - self.assertEqual(d, [('del', './count2.txt', c2["count2.txt"]), - ('add', './count1.txt', c1["count1.txt"])]) + self.assertEqual(sorted(d), [ + ('add', './count1.txt', c1["count1.txt"]), + ('del', './count2.txt', c2["count2.txt"]), + ]) d = c1.diff(c2) - self.assertEqual(d, [('del', './count1.txt', c1["count1.txt"]), - ('add', './count2.txt', c2["count2.txt"])]) + self.assertEqual(sorted(d), [ + ('add', './count2.txt', c2["count2.txt"]), + ('del', './count1.txt', c1["count1.txt"]), + ]) self.assertNotEqual(c1.portable_manifest_text(), c2.portable_manifest_text()) c1.apply(d) self.assertEqual(c1.portable_manifest_text(), c2.portable_manifest_text()) @@ -970,11 +1004,15 @@ class NewCollectionTestCase(unittest.TestCase, CollectionTestMixin): c1 = Collection('. 781e5e245d69b566979b86e28d23f2c7+10 0:10:count1.txt\n') c2 = Collection('. 781e5e245d69b566979b86e28d23f2c7+10 5348b82a029fd9e971a811ce1f71360b+43 0:10:count1.txt 10:20:count2.txt\n') d = c2.diff(c1) - self.assertEqual(d, [('del', './count2.txt', c2["count2.txt"]), - ('tok', './count1.txt', c2["count1.txt"], c1["count1.txt"])]) + self.assertEqual(sorted(d), [ + ('del', './count2.txt', c2["count2.txt"]), + ('tok', './count1.txt', c2["count1.txt"], c1["count1.txt"]), + ]) d = c1.diff(c2) - self.assertEqual(d, [('add', './count2.txt', c2["count2.txt"]), - ('tok', './count1.txt', c2["count1.txt"], c1["count1.txt"])]) + self.assertEqual(sorted(d), [ + ('add', './count2.txt', c2["count2.txt"]), + ('tok', './count1.txt', c2["count1.txt"], c1["count1.txt"]), + ]) self.assertNotEqual(c1.portable_manifest_text(), c2.portable_manifest_text()) c1.apply(d) @@ -984,12 +1022,15 @@ class NewCollectionTestCase(unittest.TestCase, CollectionTestMixin): c1 = Collection('. 781e5e245d69b566979b86e28d23f2c7+10 0:10:count1.txt\n') c2 = Collection('. 781e5e245d69b566979b86e28d23f2c7+10 0:10:count1.txt\n./foo 5348b82a029fd9e971a811ce1f71360b+43 0:10:count2.txt\n') d = c2.diff(c1) - self.assertEqual(d, [('del', './foo', c2["foo"]), - ('tok', './count1.txt', c2["count1.txt"], c1["count1.txt"])]) + self.assertEqual(sorted(d), [ + ('del', './foo', c2["foo"]), + ('tok', './count1.txt', c2["count1.txt"], c1["count1.txt"]), + ]) d = c1.diff(c2) - self.assertEqual(d, [('add', './foo', c2["foo"]), - ('tok', './count1.txt', c2["count1.txt"], c1["count1.txt"])]) - + self.assertEqual(sorted(d), [ + ('add', './foo', c2["foo"]), + ('tok', './count1.txt', c2["count1.txt"], c1["count1.txt"]), + ]) self.assertNotEqual(c1.portable_manifest_text(), c2.portable_manifest_text()) c1.apply(d) self.assertEqual(c1.portable_manifest_text(), c2.portable_manifest_text()) @@ -997,15 +1038,18 @@ class NewCollectionTestCase(unittest.TestCase, CollectionTestMixin): def test_diff_del_add_in_subcollection(self): c1 = Collection('. 781e5e245d69b566979b86e28d23f2c7+10 0:10:count1.txt\n./foo 5348b82a029fd9e971a811ce1f71360b+43 0:10:count2.txt\n') c2 = Collection('. 781e5e245d69b566979b86e28d23f2c7+10 0:10:count1.txt\n./foo 5348b82a029fd9e971a811ce1f71360b+43 0:3:count3.txt\n') - d = c2.diff(c1) - self.assertEqual(d, [('del', './foo/count3.txt', c2.find("foo/count3.txt")), - ('add', './foo/count2.txt', c1.find("foo/count2.txt")), - ('tok', './count1.txt', c2["count1.txt"], c1["count1.txt"])]) + self.assertEqual(sorted(d), [ + ('add', './foo/count2.txt', c1.find("foo/count2.txt")), + ('del', './foo/count3.txt', c2.find("foo/count3.txt")), + ('tok', './count1.txt', c2["count1.txt"], c1["count1.txt"]), + ]) d = c1.diff(c2) - self.assertEqual(d, [('del', './foo/count2.txt', c1.find("foo/count2.txt")), - ('add', './foo/count3.txt', c2.find("foo/count3.txt")), - ('tok', './count1.txt', c2["count1.txt"], c1["count1.txt"])]) + self.assertEqual(sorted(d), [ + ('add', './foo/count3.txt', c2.find("foo/count3.txt")), + ('del', './foo/count2.txt', c1.find("foo/count2.txt")), + ('tok', './count1.txt', c2["count1.txt"], c1["count1.txt"]), + ]) self.assertNotEqual(c1.portable_manifest_text(), c2.portable_manifest_text()) c1.apply(d) @@ -1015,11 +1059,15 @@ class NewCollectionTestCase(unittest.TestCase, CollectionTestMixin): c1 = Collection('. 781e5e245d69b566979b86e28d23f2c7+10 0:10:count1.txt\n./foo 5348b82a029fd9e971a811ce1f71360b+43 0:10:count2.txt\n') c2 = Collection('. 781e5e245d69b566979b86e28d23f2c7+10 0:10:count1.txt 0:3:foo\n') d = c2.diff(c1) - self.assertEqual(d, [('mod', './foo', c2["foo"], c1["foo"]), - ('tok', './count1.txt', c2["count1.txt"], c1["count1.txt"])]) + self.assertEqual(sorted(d), [ + ('mod', './foo', c2["foo"], c1["foo"]), + ('tok', './count1.txt', c2["count1.txt"], c1["count1.txt"]), + ]) d = c1.diff(c2) - self.assertEqual(d, [('mod', './foo', c1["foo"], c2["foo"]), - ('tok', './count1.txt', c2["count1.txt"], c1["count1.txt"])]) + self.assertEqual(sorted(d), [ + ('mod', './foo', c1["foo"], c2["foo"]), + ('tok', './count1.txt', c2["count1.txt"], c1["count1.txt"]), + ]) self.assertNotEqual(c1.portable_manifest_text(), c2.portable_manifest_text()) c1.apply(d) @@ -1029,9 +1077,11 @@ class NewCollectionTestCase(unittest.TestCase, CollectionTestMixin): c1 = Collection('. 781e5e245d69b566979b86e28d23f2c7+10 0:10:count1.txt\n') c2 = Collection('. 5348b82a029fd9e971a811ce1f71360b+43 0:10:count2.txt\n') d = c1.diff(c2) - self.assertEqual(d, [('del', './count1.txt', c1["count1.txt"]), - ('add', './count2.txt', c2["count2.txt"])]) - f = c1.open("count1.txt", "w") + self.assertEqual(sorted(d), [ + ('add', './count2.txt', c2["count2.txt"]), + ('del', './count1.txt', c1["count1.txt"]), + ]) + f = c1.open("count1.txt", "wb") f.write(b"zzzzz") # c1 changed, so it should not be deleted. @@ -1043,7 +1093,7 @@ class NewCollectionTestCase(unittest.TestCase, CollectionTestMixin): c2 = Collection('. 5348b82a029fd9e971a811ce1f71360b+43 0:10:count1.txt') d = c1.diff(c2) self.assertEqual(d, [('mod', './count1.txt', c1["count1.txt"], c2["count1.txt"])]) - f = c1.open("count1.txt", "w") + f = c1.open("count1.txt", "wb") f.write(b"zzzzz") # c1 changed, so c2 mod will go to a conflict file @@ -1056,9 +1106,11 @@ class NewCollectionTestCase(unittest.TestCase, CollectionTestMixin): c1 = Collection('. 781e5e245d69b566979b86e28d23f2c7+10 0:10:count2.txt\n') c2 = Collection('. 5348b82a029fd9e971a811ce1f71360b+43 0:10:count1.txt\n') d = c1.diff(c2) - self.assertEqual(d, [('del', './count2.txt', c1["count2.txt"]), - ('add', './count1.txt', c2["count1.txt"])]) - f = c1.open("count1.txt", "w") + self.assertEqual(sorted(d), [ + ('add', './count1.txt', c2["count1.txt"]), + ('del', './count2.txt', c1["count2.txt"]), + ]) + f = c1.open("count1.txt", "wb") f.write(b"zzzzz") # c1 added count1.txt, so c2 add will go to a conflict file @@ -1084,22 +1136,22 @@ class NewCollectionTestCase(unittest.TestCase, CollectionTestMixin): c1 = Collection() events = [] c1.subscribe(lambda event, collection, name, item: events.append((event, collection, name, item))) - f = c1.open("foo.txt", "w") + f = c1.open("foo.txt", "wb") self.assertEqual(events[0], (arvados.collection.ADD, c1, "foo.txt", f.arvadosfile)) def test_open_w(self): c1 = Collection(". 781e5e245d69b566979b86e28d23f2c7+10 0:10:count1.txt\n") self.assertEqual(c1["count1.txt"].size(), 10) - c1.open("count1.txt", "w").close() + c1.open("count1.txt", "wb").close() self.assertEqual(c1["count1.txt"].size(), 0) class NewCollectionTestCaseWithServers(run_test_server.TestCaseWithServers): def test_get_manifest_text_only_committed(self): c = Collection() - with c.open("count.txt", "w") as f: + with c.open("count.txt", "wb") as f: # One file committed - with c.open("foo.txt", "w") as foo: + with c.open("foo.txt", "wb") as foo: foo.write(b"foo") foo.flush() # Force block commit f.write(b"0123456789") @@ -1121,21 +1173,79 @@ class NewCollectionTestCaseWithServers(run_test_server.TestCaseWithServers): def test_only_small_blocks_are_packed_together(self): c = Collection() - # Write a couple of small files, - f = c.open("count.txt", "w") + # Write a couple of small files, + f = c.open("count.txt", "wb") f.write(b"0123456789") f.close(flush=False) - foo = c.open("foo.txt", "w") + foo = c.open("foo.txt", "wb") foo.write(b"foo") foo.close(flush=False) # Then, write a big file, it shouldn't be packed with the ones above - big = c.open("bigfile.txt", "w") + big = c.open("bigfile.txt", "wb") big.write(b"x" * 1024 * 1024 * 33) # 33 MB > KEEP_BLOCK_SIZE/2 big.close(flush=False) self.assertEqual( c.manifest_text("."), '. 2d303c138c118af809f39319e5d507e9+34603008 a8430a058b8fbf408e1931b794dbd6fb+13 0:34603008:bigfile.txt 34603008:10:count.txt 34603018:3:foo.txt\n') + def test_flush_after_small_block_packing(self): + c = Collection() + # Write a couple of small files, + f = c.open("count.txt", "wb") + f.write(b"0123456789") + f.close(flush=False) + foo = c.open("foo.txt", "wb") + foo.write(b"foo") + foo.close(flush=False) + + self.assertEqual( + c.manifest_text(), + '. a8430a058b8fbf408e1931b794dbd6fb+13 0:10:count.txt 10:3:foo.txt\n') + + f = c.open("count.txt", "rb+") + f.close(flush=True) + + self.assertEqual( + c.manifest_text(), + '. a8430a058b8fbf408e1931b794dbd6fb+13 0:10:count.txt 10:3:foo.txt\n') + + def test_write_after_small_block_packing2(self): + c = Collection() + # Write a couple of small files, + f = c.open("count.txt", "wb") + f.write(b"0123456789") + f.close(flush=False) + foo = c.open("foo.txt", "wb") + foo.write(b"foo") + foo.close(flush=False) + + self.assertEqual( + c.manifest_text(), + '. a8430a058b8fbf408e1931b794dbd6fb+13 0:10:count.txt 10:3:foo.txt\n') + + f = c.open("count.txt", "rb+") + f.write(b"abc") + f.close(flush=False) + + self.assertEqual( + c.manifest_text(), + '. 900150983cd24fb0d6963f7d28e17f72+3 a8430a058b8fbf408e1931b794dbd6fb+13 0:3:count.txt 6:7:count.txt 13:3:foo.txt\n') + + + def test_small_block_packing_with_overwrite(self): + c = Collection() + c.open("b1", "wb").close() + c["b1"].writeto(0, b"b1", 0) + + c.open("b2", "wb").close() + c["b2"].writeto(0, b"b2", 0) + + c["b1"].writeto(0, b"1b", 0) + + self.assertEquals(c.manifest_text(), ". ed4f3f67c70b02b29c50ce1ea26666bd+4 0:2:b1 2:2:b2\n") + self.assertEquals(c["b1"].manifest_text(), ". ed4f3f67c70b02b29c50ce1ea26666bd+4 0:2:b1\n") + self.assertEquals(c["b2"].manifest_text(), ". ed4f3f67c70b02b29c50ce1ea26666bd+4 2:2:b2\n") + class CollectionCreateUpdateTest(run_test_server.TestCaseWithServers): MAIN_SERVER = {} @@ -1150,7 +1260,7 @@ class CollectionCreateUpdateTest(run_test_server.TestCaseWithServers): self.assertEqual(c.portable_data_hash(), "d41d8cd98f00b204e9800998ecf8427e+0") self.assertEqual(c.api_response()["portable_data_hash"], "d41d8cd98f00b204e9800998ecf8427e+0" ) - with c.open("count.txt", "w") as f: + with c.open("count.txt", "wb") as f: f.write(b"0123456789") self.assertEqual(c.portable_manifest_text(), ". 781e5e245d69b566979b86e28d23f2c7+10 0:10:count.txt\n") @@ -1176,7 +1286,7 @@ class CollectionCreateUpdateTest(run_test_server.TestCaseWithServers): c1.save() c2 = Collection(c1.manifest_locator()) - with c2.open("count.txt", "w") as f: + with c2.open("count.txt", "wb") as f: f.write(b"abcdefg") diff = c1.diff(c2) @@ -1204,7 +1314,7 @@ class CollectionCreateUpdateTest(run_test_server.TestCaseWithServers): c1.save() c2 = arvados.collection.Collection(c1.manifest_locator()) - with c2.open("count.txt", "w") as f: + with c2.open("count.txt", "wb") as f: f.write(b"abcdefg") c2.save() @@ -1218,11 +1328,11 @@ class CollectionCreateUpdateTest(run_test_server.TestCaseWithServers): c1 = self.create_count_txt() c1.save() - with c1.open("count.txt", "w") as f: + with c1.open("count.txt", "wb") as f: f.write(b"XYZ") c2 = arvados.collection.Collection(c1.manifest_locator()) - with c2.open("count.txt", "w") as f: + with c2.open("count.txt", "wb") as f: f.write(b"abcdefg") c2.save()