Fixed odb performance tests

8 files changed, 102 insertions, 94 deletions

diff --git a/git/test/performance/__init__.py b/git/test/performance/__init__.py
new file mode 100644
index 00000000..6bd117b9
--- /dev/null
+++ b/git/test/performance/__init__.py
@@ -0,0 +1,2 @@
+"""Note: This init file makes the performance tests an integral part of the test suite
+as nose will now pick them up. Previously the init file was intentionally omitted"""
diff --git a/git/test/performance/db/__init__.py b/git/test/performance/db/__init__.py
new file mode 100644
index 00000000..8b137891
--- /dev/null
+++ b/git/test/performance/db/__init__.py
@@ -0,0 +1 @@
+
diff --git a/git/test/performance/db/odb_impl.py b/git/test/performance/db/odb_impl.py
new file mode 100644
index 00000000..50ee37e1
--- /dev/null
+++ b/git/test/performance/db/odb_impl.py
@@ -0,0 +1,69 @@
+"""Performance tests for object store"""
+
+from time import time
+import sys
+import stat
+import copy
+
+from git.test.performance.lib import (
+	TestBigRepoR,
+	GlobalsItemDeletorMetaCls
+	)
+
+class PerfBaseDeletorMetaClass(GlobalsItemDeletorMetaCls):
+	ModuleToDelete = 'TestObjDBPerformanceBase'
+	
+
+class TestObjDBPerformanceBase(TestBigRepoR):
+	__metaclass__ = PerfBaseDeletorMetaClass
+	
+	def test_random_access_test(self):
+		repo = self.rorepo
+	
+		# GET COMMITS
+		st = time()
+		root_commit = repo.commit(self.head_sha_2k)
+		commits = list(root_commit.traverse())
+		nc = len(commits)
+		elapsed = time() - st
+		
+		print >> sys.stderr, "%s: Retrieved %i commits from ObjectStore in %g s ( %f commits / s )" % (type(repo.odb), nc, elapsed, nc / elapsed)
+			
+		# GET TREES
+		# walk all trees of all commits
+		st = time()
+		blobs_per_commit = list()
+		nt = 0
+		for commit in commits:
+			tree = commit.tree
+			blobs = list()
+			for item in tree.traverse():
+				nt += 1
+				if item.type == 'blob':
+					blobs.append(item)
+				# direct access for speed
+			# END while trees are there for walking
+			blobs_per_commit.append(blobs)
+		# END for each commit
+		elapsed = time() - st
+		
+		print >> sys.stderr, "%s: Retrieved %i objects from %i commits in %g s ( %f objects / s )" % (type(repo.odb), nt, len(commits), elapsed, nt / elapsed)
+		
+		# GET BLOBS
+		st = time()
+		nb = 0
+		too_many = 15000
+		data_bytes = 0
+		for blob_list in blobs_per_commit:
+			for blob in blob_list:
+				data_bytes += len(blob.data_stream.read())
+			# END for each blobsha
+			nb += len(blob_list)
+			if nb > too_many:
+				break
+		# END for each bloblist
+		elapsed = time() - st
+		
+		print >> sys.stderr, "%s: Retrieved %i blob (%i KiB) and their data in %g s ( %f blobs / s, %f KiB / s )" % (type(repo.odb), nb, data_bytes/1000, elapsed, nb / elapsed, (data_bytes / 1000) / elapsed)
+		
+		
diff --git a/git/test/performance/db/test_odb_cmd.py b/git/test/performance/db/test_odb_cmd.py
new file mode 100644
index 00000000..acd55cc9
--- /dev/null
+++ b/git/test/performance/db/test_odb_cmd.py
@@ -0,0 +1,6 @@
+from git.db.complex import CmdCompatibilityGitDB
+from odb_impl import TestObjDBPerformanceBase
+
+class TestCmdDB(TestObjDBPerformanceBase):
+	RepoCls = CmdCompatibilityGitDB
+	
diff --git a/git/test/performance/db/test_odb_pure.py b/git/test/performance/db/test_odb_pure.py
new file mode 100644
index 00000000..6ed3585d
--- /dev/null
+++ b/git/test/performance/db/test_odb_pure.py
@@ -0,0 +1,6 @@
+from git.db.complex import PureCompatibilityGitDB
+from odb_impl import TestObjDBPerformanceBase
+
+class TestPureDB(TestObjDBPerformanceBase):
+	RepoCls = PureCompatibilityGitDB
+	
diff --git a/git/test/performance/lib.py b/git/test/performance/lib.py
index 08c0e91c..cf80a0de 100644
--- a/git/test/performance/lib.py
+++ b/git/test/performance/lib.py
@@ -1,18 +1,12 @@
 """Contains library functions"""
 import os
-from git.test.lib import TestBase
+from git.test.lib import (
+						TestBase, 
+						GlobalsItemDeletorMetaCls
+						)
 import shutil
 import tempfile
 
-from git.db import (
-						CmdGitDB,
-						GitDB
-					)
-
-from git import (
-	Repo
-	)
-
 #{ Invvariants
 k_env_git_repo = "GIT_PYTHON_TEST_GIT_REPO_BASE"
 #} END invariants
@@ -38,11 +32,7 @@ class TestBigRepoR(TestBase):
 	
 	* gitrorepo
 	
-	 * Read-Only git repository - actually the repo of git itself
-	 
-    * puregitrorepo
-    
-     * As gitrepo, but uses pure python implementation
+	 * a big read-only git repository
     """
 	 
 	#{ Invariants
@@ -50,29 +40,33 @@ class TestBigRepoR(TestBase):
 	head_sha_50 = '32347c375250fd470973a5d76185cac718955fd5'
 	#} END invariants 
 	
+	#{ Configuration
+	RepoCls = None
+	#} END configuration
+	
 	@classmethod
 	def setUpAll(cls):
 		super(TestBigRepoR, cls).setUpAll()
-		repo_path = resolve_or_fail(k_env_git_repo)
-		cls.gitrorepo = Repo(repo_path, odbt=CmdGitDB)
-		cls.puregitrorepo = Repo(repo_path, odbt=GitDB)
+		if cls.RepoCls is None:
+			raise AssertionError("Require RepoCls in class %s to be set" % cls)
+		#END assert configuration
+		cls.rorepo = cls.RepoCls(resolve_or_fail(k_env_git_repo))
 
 
 class TestBigRepoRW(TestBigRepoR):
 	"""As above, but provides a big repository that we can write to.
 	
-	Provides ``self.gitrwrepo`` and ``self.puregitrwrepo``"""
+	Provides ``self.rwrepo``"""
 	
 	@classmethod
 	def setUpAll(cls):
 		super(TestBigRepoRW, cls).setUpAll()
 		dirname = tempfile.mktemp()
 		os.mkdir(dirname)
-		cls.gitrwrepo = cls.gitrorepo.clone(dirname, shared=True, bare=True, odbt=CmdGitDB)
-		cls.puregitrwrepo = Repo(dirname, odbt=GitDB)
+		cls.rwrepo = cls.gitrorepo.clone(dirname, shared=True, bare=True)
 	
 	@classmethod
 	def tearDownAll(cls):
-		shutil.rmtree(cls.gitrwrepo.working_dir)
+		shutil.rmtree(cls.rwrepo.working_dir)
 		
 #} END base classes
diff --git a/git/test/performance/test_commit.py b/git/test/performance/test_commit.py
index 561e073f..72755e05 100644
--- a/git/test/performance/test_commit.py
+++ b/git/test/performance/test_commit.py
@@ -70,9 +70,9 @@ class TestPerformance(TestBigRepoRW):
 		print >> sys.stderr, "Iterated %i Commits in %s [s] ( %f commits/s )" % (nc, elapsed_time, nc/elapsed_time)
 		
 	def test_commit_serialization(self):
-		assert_commit_serialization(self.gitrwrepo, self.head_sha_2k, True)
+		assert_commit_serialization(self.rwrepo, self.head_sha_2k, True)
 		
-		rwrepo = self.gitrwrepo
+		rwrepo = self.rwrepo
 		make_object = rwrepo.odb.store
 		# direct serialization - deserialization can be tested afterwards
 		# serialization is probably limited on IO
diff --git a/git/test/performance/test_odb.py b/git/test/performance/test_odb.py
deleted file mode 100644
index 32b70f69..00000000
--- a/git/test/performance/test_odb.py
+++ /dev/null
@@ -1,70 +0,0 @@
-"""Performance tests for object store"""
-
-from time import time
-import sys
-import stat
-
-from lib import (
-	TestBigRepoR
-	)
-
-
-class TestObjDBPerformance(TestBigRepoR):
-	
-	def test_random_access(self):
-		results = [ ["Iterate Commits"], ["Iterate Blobs"], ["Retrieve Blob Data"] ]
-		for repo in (self.gitrorepo, self.puregitrorepo):
-			# GET COMMITS
-			st = time()
-			root_commit = repo.commit(self.head_sha_2k)
-			commits = list(root_commit.traverse())
-			nc = len(commits)
-			elapsed = time() - st
-			
-			print >> sys.stderr, "%s: Retrieved %i commits from ObjectStore in %g s ( %f commits / s )" % (type(repo.odb), nc, elapsed, nc / elapsed)
-			results[0].append(elapsed)
-				
-			# GET TREES
-			# walk all trees of all commits
-			st = time()
-			blobs_per_commit = list()
-			nt = 0
-			for commit in commits:
-				tree = commit.tree
-				blobs = list()
-				for item in tree.traverse():
-					nt += 1
-					if item.type == 'blob':
-						blobs.append(item)
-					# direct access for speed
-				# END while trees are there for walking
-				blobs_per_commit.append(blobs)
-			# END for each commit
-			elapsed = time() - st
-			
-			print >> sys.stderr, "%s: Retrieved %i objects from %i commits in %g s ( %f objects / s )" % (type(repo.odb), nt, len(commits), elapsed, nt / elapsed)
-			results[1].append(elapsed)
-			
-			# GET BLOBS
-			st = time()
-			nb = 0
-			too_many = 15000
-			data_bytes = 0
-			for blob_list in blobs_per_commit:
-				for blob in blob_list:
-					data_bytes += len(blob.data_stream.read())
-				# END for each blobsha
-				nb += len(blob_list)
-				if nb > too_many:
-					break
-			# END for each bloblist
-			elapsed = time() - st
-			
-			print >> sys.stderr, "%s: Retrieved %i blob (%i KiB) and their data in %g s ( %f blobs / s, %f KiB / s )" % (type(repo.odb), nb, data_bytes/1000, elapsed, nb / elapsed, (data_bytes / 1000) / elapsed)
-			results[2].append(elapsed)
-		# END for each repo type
-		
-		# final results
-		for test_name, a, b in results:
-			print >> sys.stderr, "%s: %f s vs %f s, pure is %f times slower" % (test_name, a, b, b / a)
-		# END for each result