[Findit] Add local_git_repository

appengine/findit/common/local_git_repository.py

+# Copyright 2014 The Chromium Authors. All rights reserved.
+# Use of this source code is governed by a BSD-style license that can be
+# found in the LICENSE file.
+
+import logging
+import os
+import subprocess
+import threading
+
+from common import local_git_parsers
+from common import repo_util
+from common.repository import Repository
+
+_CHANGELOG_FORMAT_STRING = ('commit %H%n'
+                            'author %an%n'
+                            'author-mail %ae%n'
+                            'author-time %at%n%n'
+                            'committer %cn%n'
+                            'committer-mail %ce%n'
+                            'committer-time %ct%n%n'
+                            '--Message start--%n%B%n--Message end--%n')
+_CHANGELOGS_FORMAT_STRING = ('**Changelog start**%%n%s' %
+                             _CHANGELOG_FORMAT_STRING)
+CHECKOUT_ROOT_DIR = os.path.join(os.path.expanduser('~'), '.local_checkouts')
+
+
+class LocalGitRepository(Repository):
+  """Represents local checkout of git repository on chromium host.
+
+  Note, to checkout internal repos automatically which you have access to,
+  follow the instructions in ('https://g3doc.corp.google.com/company/teams/
+  chrome/chrome_build_instructions.md?
+  cl=head#authentication-to-git-servers-chrome-internalgooglesourcecom') first.
+  """
+  lock = threading.Lock()
+  # Keep track all the updated repos, so every repo only get updated once.
+  updated_repos = set()
+  def __init__(self, repo_url=None):
+    self._repo_path = None
+    self._repo_url = None
+    self.repo_url = repo_url
+
+  @property
+  def repo_path(self):
+    return self._repo_path
+
+  @property
+  def repo_url(self):
+    return self._repo_url
+
+  @repo_url.setter
+  def repo_url(self, repo_url):
+    if not repo_url or (hasattr(self, '_repo_url') and

[stgao, 2016/10/22 00:43:35]

Why hasattr is needed? Isn't self._repo_url already added in the constructor?

[Sharu Jiang, 2016/10/25 06:45:16]

Oops, should have deleted.

+                        self._repo_url == repo_url):
+      return
+
+    self._repo_url = repo_url
+    http_trimmed_url = repo_url.split('https://', 1)[1]
+    host_trimmed_url = '/'.join(http_trimmed_url.split('/')[1:])
+
+    self._repo_path = host_trimmed_url
+    self._CloneOrUpdateRepoIfNeeded()
+
+  def _CloneOrUpdateRepoIfNeeded(self):
+    """Clones repo, or update it if it didn't got updated before."""
+    if self.repo_url in LocalGitRepository.updated_repos:
+      return
+
+    real_repo_path = os.path.join(CHECKOUT_ROOT_DIR, self.repo_path)
+    with LocalGitRepository.lock:
+      # Clone the repo if needed.
+      if not os.path.exists(real_repo_path):
+        subprocess.call(['git', 'clone', self.repo_url, real_repo_path])
+      # Update repo if it's already cloned.
+      else:
+        # Disable verbose of git pull.
+        with open(os.devnull, 'w') as null_handle:
+          subprocess.check_call(
+              'cd %s; git pull' % real_repo_path,
+              stdout=null_handle,
+              stderr=null_handle,
+              shell=True)
+
+      LocalGitRepository.updated_repos.add(self.repo_url)
+
+  def _GetLocalGitCommandOutput(self, command):
+    """Gets the output stream of a git command for a local git repo.
+
+    If the corresponding git repo_path is not downloaded yet, download it from
+    https://chromium.googlesource.com.
+
+    Args:
+      command (str): Command to execute at this repo to get output.
+
+    Return:
+      Output steam of the git command.
+    """
+    real_repo_path = os.path.join(CHECKOUT_ROOT_DIR, self.repo_path)
+    command = 'cd %s; %s' % (real_repo_path, command)
+    p = subprocess.Popen(
+        command, stdout=subprocess.PIPE, stderr=subprocess.STDOUT, shell=True)
+    return p.communicate()[0]
+
+  def GetChangeLog(self, revision):
+    """Returns the change log of the given revision."""
+    command = ('git log --pretty=format:"%s" --max-count=1 --raw '
+               '--no-abbrev %s' % (_CHANGELOG_FORMAT_STRING, revision))
+    output = self._GetLocalGitCommandOutput(command)
+    change_log = local_git_parsers.GitChangeLogParser()(output)
+    change_log.commit_url = '%s/+/%s' % (self.repo_url, change_log.revision)
+    return change_log
+
+  def GetChangeLogs(self, start_revision, end_revision):  # pylint: disable=W
+    """Returns change log list in (start_revision, end_revision]."""
+    command = ('git log --pretty=format:"%s" --raw '
+               '--no-abbrev %s' % (_CHANGELOGS_FORMAT_STRING,
+                                   '%s..%s' % (start_revision, end_revision)))
+    output = self._GetLocalGitCommandOutput(command)
+    return local_git_parsers.GitChangeLogsParser()(output)
+
+  def GetChangeDiff(self, revision, path=None):  # pylint: disable=W
+    """Returns the diff of the given revision."""
+    command = 'git log --format="" --max-count=1 %s' % revision
+    if path:
+      command += ' -p %s' % path
+    output = self._GetLocalGitCommandOutput(command)
+    return local_git_parsers.GitDiffParser()(output)
+
+  def GetBlame(self, path, revision):
+    """Returns blame of the file at ``path`` of the given revision."""
+    command = 'git blame --porcelain %s %s' % (path, revision)
+    output = self._GetLocalGitCommandOutput(command, path)
+    return local_git_parsers.GitBlameParser()(output)
+
+  def GetSource(self, path, revision):
+    """Returns source code of the file at ``path`` of the given revision."""
+    # Check whether the requested file exist or not.
+    if not os.path.isfile(os.path.join(
+        os.path.join(CHECKOUT_ROOT_DIR, self.repo_path),
+        path)):
+      return None
+
+    command = 'git show %s:%s' % (revision, path)
+    output = self._GetLocalGitCommandOutput(command, path)
+    return local_git_parsers.GitSourceParser()(output)