almost...alive

Author: wesm

gitbench/git.py

@@ -121,8 +121,33 @@ class BenchRepo(object):
     """
     Manage an isolated copy of a repository for benchmarking
     """
-    def __init__(self, conf):
-        self.conf = conf
+    def __init__(self, source_dir, target_dir, build_cmds):
+        self.source_dir = source_dir
+        self.target_dir = target_dir
+        self.build_cmds = build_cmds
+        self._copy_repo()
+
+    def _copy_repo(self):
+        if os.path.exists(self.target_dir):
+            print 'Deleting %s first' % self.target_dir
+            # response = raw_input('%s exists, delete? y/n' % self.target_dir)
+            # if response == 'n':
+            #     raise Exception('foo')
+            cmd = 'rm -rf %s' % self.target_dir
+            print cmd
+            os.system(cmd)
+        cmd = 'cp -r %s %s' % (self.source_dir, self.target_dir)
+        print cmd
+        os.system(cmd)
+
+        self._copy_benchmark_script()
+
+    def _copy_benchmark_script(self):
+        pth, _ = os.path.split(os.path.abspath(__file__))
+        script_path = os.path.join(pth, 'scripts/gb_run_benchmarks.py')
+        cmd = 'cp %s %s' % (script_path, self.target_dir)
+        print cmd
+        os.system(cmd)
 
     def switch_to_revision(self, rev):
         """
@@ -132,10 +157,25 @@ def switch_to_revision(self, rev):
         self._build()
 
     def _checkout(self, rev):
-        pass
+        git = _git_command(self.target_dir)
+        rest = 'checkout -f %s' % rev
+        args = git.split() + rest.split()
+
+        print ' '.join(args)
+
+        proc = subprocess.Popen(args, stdout=subprocess.PIPE)
+        proc.wait()
+        print proc.stdout.read()
 
     def _build(self):
-        pass
+        cmd = ';'.join([x for x in self.build_cmds.split('\n')
+                        if len(x.strip()) > 0])
+
+        print cmd
+        proc = subprocess.Popen(cmd, stdout=subprocess.PIPE, shell=True,
+                                cwd=self.target_dir)
+        stdout, stderr = proc.communicate()
+        print stdout
 
 
 def _convert_timezones(stamps):

gitbench/runner.py

@@ -1,5 +1,6 @@
 import cPickle as pickle
 import os
+import subprocess
 
 from gitbench.git import GitRepo, BenchRepo
 from gitbench.db import BenchmarkDB
@@ -49,7 +50,7 @@ def run(self):
             results = self._run_revision(rev)
 
             for checksum, timing in results.iteritems():
-                self.db.write_result(checksum, revision,
+                self.db.write_result(checksum, rev,
                                      timing.get('loops'),
                                      timing.get('timing'),
                                      timing.get('traceback'))
@@ -73,15 +74,25 @@ def _run_revision(self, rev):
 
         print 'Running %d benchmarks for revision %s' % (len(need_to_run), rev)
 
-        self._prepare_code(rev)
+        self.bench_repo.switch_to_revision(rev)
 
-        pickle_path = pjoin(self.tmp_dir, 'benchmarks.pickle')
-        results_path = pjoin(self.tmp_dir, 'results.pickle')
+        pickle_path = os.path.join(self.tmp_dir, 'benchmarks.pickle')
+        results_path = os.path.join(self.tmp_dir, 'results.pickle')
         if os.path.exists(results_path):
             os.remove(results_path)
         pickle.dump(need_to_run, open(pickle_path, 'w'))
 
-        _run_benchmark_subproc(pickle_path, results_path)
+        # run the process
+        cmd = 'python run_benchmarks.py --input=%s --output=%s' % (pickle_path,
+                                                                   results_path)
+        proc = subprocess.Popen(cmd, stdout=subprocess.PIPE,
+                                cwd=self.tmp_dir)
+        stdout, stderr = proc.communicate()
+
+        print stdout
+
+        if stderr:
+            raise Exception(stderr)
 
         results = pickle.load(open(results_path, 'r'))
         os.remove(pickle_path)
@@ -110,12 +121,3 @@ def _get_revisions_to_run(self):
             raise Exception('unrecognized run_method %s' % self.run_method)
 
         return revs_to_run
-
-    def _prepare_code(self, rev):
-        self.bench_repo.checkout(rev)
-        self.bench_repo.build()
-
-def _run_benchmark_subproc(input_path, result_path):
-    from subprocess import Popen, PIPE
-
-

gitbench/scripts/run_benchmarks.py renamed to gitbench/scripts/gb_run_benchmarks.py

@@ -3,22 +3,31 @@
 import gitbench.git as git
 reload(git)
 
-repo_path = '/home/wesm/code/pandas'
-repo = git.GitRepo(repo_path)
+# repo_path = '/home/wesm/code/pandas'
+# repo = git.GitRepo(repo_path)
 
-hists = repo.messages
+# hists = repo.messages
 
-def churn_graph(repo):
-    omit_paths = [path for path in churn.major_axis
-                  if not path.endswith('.pyx') or not path.endswith('.py')]
-    omit_shas = [sha for sha in churn.minor_axis
-                 if 'LF' in hists[sha]]
-    omit_shas.append('dcf3490')
+# def churn_graph(repo):
+#     omit_paths = [path for path in churn.major_axis
+#                   if not path.endswith('.pyx') or not path.endswith('.py')]
+#     omit_shas = [sha for sha in churn.minor_axis
+#                  if 'LF' in hists[sha]]
+#     omit_shas.append('dcf3490')
 
-    by_date = repo.get_churn(omit_shas=omit_shas, omit_paths=omit_paths)
-    by_date = by_date.drop([datetime(2011, 6, 10)])
+#     by_date = repo.get_churn(omit_shas=omit_shas, omit_paths=omit_paths)
+#     by_date = by_date.drop([datetime(2011, 6, 10)])
 
-    # clean out days where I touched Cython
-    by_date = by_date[by_date < 5000]
-    return by_date
+#     # clean out days where I touched Cython
+#     by_date = by_date[by_date < 5000]
+#     return by_date
 
+REPO_PATH = '/home/wesm/code/pandas'
+TMP_DIR = '/home/wesm/tmp/gb_pandas'
+BUILD = """
+python setup.py build_ext --inplace
+"""
+
+repo = git.GitRepo(REPO_PATH)
+
+burp = git.BenchRepo(REPO_PATH, TMP_DIR, BUILD)