From 4e1c89ec97ec90037583e85d0e9e71e9c845a19b Mon Sep 17 00:00:00 2001 From: Sebastian Thiel Date: Wed, 2 Jun 2010 16:13:32 +0200 Subject: Added performance testing foundation library, reworked existing performance tests to work on larger repositories --- test/git/performance/lib.py | 46 +++++++++++++++++++++++++ test/git/performance/test_commit.py | 68 +++++++++++++++++++++++++++++++++++++ test/git/test_performance.py | 68 ------------------------------------- 3 files changed, 114 insertions(+), 68 deletions(-) create mode 100644 test/git/performance/lib.py create mode 100644 test/git/performance/test_commit.py delete mode 100644 test/git/test_performance.py diff --git a/test/git/performance/lib.py b/test/git/performance/lib.py new file mode 100644 index 00000000..4b552b20 --- /dev/null +++ b/test/git/performance/lib.py @@ -0,0 +1,46 @@ +"""Contains library functions""" +import os +from test.testlib import * + +from git import ( + Repo + ) + +#{ Invvariants +k_env_git_repo = "GIT_PYTHON_TEST_GIT_REPO_BASE" +#} END invariants + + +#{ Utilities +def resolve_or_fail(env_var): + """:return: resolved environment variable or raise EnvironmentError""" + try: + return os.environ[env_var] + except KeyError: + raise EnvironmentError("Please set the %r envrionment variable and retry" % env_var) + # END exception handling + +#} END utilities + + +#{ Base Classes + +class TestBigRepoReadOnly(TestBase): + """TestCase providing access to readonly 'big' repositories using the following + member variables: + + * gitrepo + + * Read-Only git repository - actually the repo of git itself""" + + #{ Invariants + head_sha_2k = '235d521da60e4699e5bd59ac658b5b48bd76ddca' + head_sha_50 = '32347c375250fd470973a5d76185cac718955fd5' + #} END invariants + + @classmethod + def setUpAll(cls): + super(TestBigRepoReadOnly, cls).setUpAll() + cls.gitrepo = Repo(resolve_or_fail(k_env_git_repo)) + +#} END base classes diff --git a/test/git/performance/test_commit.py b/test/git/performance/test_commit.py new file mode 100644 index 00000000..c1f8ce59 --- /dev/null +++ b/test/git/performance/test_commit.py @@ -0,0 +1,68 @@ +# test_performance.py +# Copyright (C) 2008, 2009 Michael Trier (mtrier@gmail.com) and contributors +# +# This module is part of GitPython and is released under +# the BSD License: http://www.opensource.org/licenses/bsd-license.php + +from test.testlib import * +from git import * +from time import time +import sys + +class TestPerformance(TestBase): + + # ref with about 100 commits in its history + ref_100 = '0.1.6' + + def _query_commit_info(self, c): + c.author + c.authored_date + c.author_tz_offset + c.committer + c.committed_date + c.committer_tz_offset + c.message + c.parents + + def test_iteration(self): + no = 0 + nc = 0 + + # find the first commit containing the given path - always do a full + # iteration ( restricted to the path in question ), but in fact it should + # return quite a lot of commits, we just take one and hence abort the operation + + st = time() + for c in self.rorepo.iter_commits(self.ref_100): + nc += 1 + self._query_commit_info(c) + for obj in c.tree.traverse(): + obj.size + no += 1 + # END for each object + # END for each commit + elapsed_time = time() - st + print >> sys.stderr, "Traversed %i Trees and a total of %i unchached objects in %s [s] ( %f objs/s )" % (nc, no, elapsed_time, no/elapsed_time) + + def test_commit_traversal(self): + # bound to cat-file parsing performance + nc = 0 + st = time() + for c in self.rorepo.commit(self.ref_100).traverse(branch_first=False): + nc += 1 + self._query_commit_info(c) + # END for each traversed commit + elapsed_time = time() - st + print >> sys.stderr, "Traversed %i Commits in %s [s] ( %f commits/s )" % (nc, elapsed_time, nc/elapsed_time) + + def test_commit_iteration(self): + # bound to stream parsing performance + nc = 0 + st = time() + for c in Commit.iter_items(self.rorepo, self.ref_100): + nc += 1 + self._query_commit_info(c) + # END for each traversed commit + elapsed_time = time() - st + print >> sys.stderr, "Iterated %i Commits in %s [s] ( %f commits/s )" % (nc, elapsed_time, nc/elapsed_time) + diff --git a/test/git/test_performance.py b/test/git/test_performance.py deleted file mode 100644 index c1f8ce59..00000000 --- a/test/git/test_performance.py +++ /dev/null @@ -1,68 +0,0 @@ -# test_performance.py -# Copyright (C) 2008, 2009 Michael Trier (mtrier@gmail.com) and contributors -# -# This module is part of GitPython and is released under -# the BSD License: http://www.opensource.org/licenses/bsd-license.php - -from test.testlib import * -from git import * -from time import time -import sys - -class TestPerformance(TestBase): - - # ref with about 100 commits in its history - ref_100 = '0.1.6' - - def _query_commit_info(self, c): - c.author - c.authored_date - c.author_tz_offset - c.committer - c.committed_date - c.committer_tz_offset - c.message - c.parents - - def test_iteration(self): - no = 0 - nc = 0 - - # find the first commit containing the given path - always do a full - # iteration ( restricted to the path in question ), but in fact it should - # return quite a lot of commits, we just take one and hence abort the operation - - st = time() - for c in self.rorepo.iter_commits(self.ref_100): - nc += 1 - self._query_commit_info(c) - for obj in c.tree.traverse(): - obj.size - no += 1 - # END for each object - # END for each commit - elapsed_time = time() - st - print >> sys.stderr, "Traversed %i Trees and a total of %i unchached objects in %s [s] ( %f objs/s )" % (nc, no, elapsed_time, no/elapsed_time) - - def test_commit_traversal(self): - # bound to cat-file parsing performance - nc = 0 - st = time() - for c in self.rorepo.commit(self.ref_100).traverse(branch_first=False): - nc += 1 - self._query_commit_info(c) - # END for each traversed commit - elapsed_time = time() - st - print >> sys.stderr, "Traversed %i Commits in %s [s] ( %f commits/s )" % (nc, elapsed_time, nc/elapsed_time) - - def test_commit_iteration(self): - # bound to stream parsing performance - nc = 0 - st = time() - for c in Commit.iter_items(self.rorepo, self.ref_100): - nc += 1 - self._query_commit_info(c) - # END for each traversed commit - elapsed_time = time() - st - print >> sys.stderr, "Iterated %i Commits in %s [s] ( %f commits/s )" % (nc, elapsed_time, nc/elapsed_time) - -- cgit v1.2.1