Merge pull request #3216 from google:use-pyink

PiperOrigin-RevId: 550029828

.git-blame-ignore-revs

@@ -0,0 +1,2 @@
+# apply pyink
+40a6e074e5224d733f964be00e21e0a1cb98bd2e

.github/analytics/get_repo_metrics.py

@@ -25,17 +25,18 @@
 import matplotlib.dates as mdates
 
 
-token = os.environ["GITHUB_TOKEN"]
-endpoint = r"https://api.github.com/graphql"
-headers = {"Authorization": f"bearer {token}"}
+token = os.environ['GITHUB_TOKEN']
+endpoint = r'https://api.github.com/graphql'
+headers = {'Authorization': f'bearer {token}'}
 
-#------------------------------------------------------------------------------
+# ------------------------------------------------------------------------------
 # GraphQL
-#------------------------------------------------------------------------------
+# ------------------------------------------------------------------------------
 # NOTE: This GraphQL logic was ported and adapted from this script:
 # https://github.com/scientific-python/devstats-data/blob/4c022961abc4ca6061f8719d9c3387e98734b90c/query.py
 # It contains style differences from Google's style guide.
 
+
 def load_query_from_file(fname, repo_owner, repo_name) -> str:
  with open(fname) as fh:
  query = fh.read()
@@ -75,9 +76,9 @@ def send_query(query, query_type, cursor=None):
  # TODO: Expand this, either by parsing the query type from the query
  # directly or manually adding more query_types to the set
  if query_type not in {'issues', 'pullRequests'}:
-  raise ValueError(
-  'Only \'issues\' and \'pullRequests\' queries are currently supported'
-  )
+ raise ValueError(
+ "Only 'issues' and 'pullRequests' queries are currently supported"
+ )
  # TODO: Generalize this
  # WARNING: The cursor injection depends on the specific structure of the
  # query, this is the main reason why query types are limited to issues/PRs
@@ -86,12 +87,13 @@ def send_query(query, query_type, cursor=None):
  cursor_ind = query.find(cursor_insertion_key) + len(cursor_insertion_key)
  query = query[:cursor_ind] + f'after:"{cursor}", ' + query[cursor_ind:]
  # Build request payload
- payload = {'query' : query}
+ payload = {'query': query}
  response = requests.post(endpoint, json=payload, headers=headers)
  return json.loads(response.content)
 
+
 def get_all_responses(query, query_type):
- "Helper function to bypass GitHub GraphQL API node limit."
+ 'Helper function to bypass GitHub GraphQL API node limit.'
  # Get data from a single response
  initial_data = send_query(query, query_type)
  data, last_cursor, total_count = parse_single_query(initial_data, query_type)
@@ -105,6 +107,7 @@ def get_all_responses(query, query_type):
  print('Done.')
  return data
 
+
 def parse_single_query(data, query_type):
  """
  Parses the data returned by `send_query`
@@ -160,19 +163,22 @@ def __init__(self, query_fname, query_type, repo_owner, repo_name):
 
  def load_query(self):
  self.query = load_query_from_file(
- self.query_fname, self.repo_owner, self.repo_name
+  self.query_fname, self.repo_owner, self.repo_name
  )
 
  def get(self):
  self.raw_data = get_all_responses(self.query, self.query_type)
 
-#------------------------------------------------------------------------------
+
+# ------------------------------------------------------------------------------
 # metrics helpers
-#------------------------------------------------------------------------------
+# ------------------------------------------------------------------------------
+
 
 def _to_datetime(date_str: str) -> datetime:
  return datetime.fromisoformat(date_str.replace('Z', ''))
 
+
 def _get_issues_features(issues):
  for issue in issues:
  issue = issue['node']
@@ -191,12 +197,13 @@ def _get_issues_features(issues):
  time_issue_closed = _to_datetime(event['createdAt'])
 
  yield {
- 'created_at': created_at,
- 'time_labeled_or_converted': time_labeled_or_converted,
- 'time_issue_closed': time_issue_closed,
- 'issue_closed': issue['state'] == 'CLOSED',
+  'created_at': created_at,
+  'time_labeled_or_converted': time_labeled_or_converted,
+  'time_issue_closed': time_issue_closed,
+  'issue_closed': issue['state'] == 'CLOSED',
  }
 
+
 def _get_pr_features(prs):
  for pr in prs:
  pr = pr['node']
@@ -207,23 +214,23 @@ def _get_pr_features(prs):
  time_merged_or_closed = None
  time_review = None
 
- if pr["reviews"]["nodes"]:
- review = pr["reviews"]["nodes"][0]
- time_review = _to_datetime(review["createdAt"])
+ if pr['reviews']['nodes']:
+ review = pr['reviews']['nodes'][0]
+ time_review = _to_datetime(review['createdAt'])
 
  for event in pr['timelineItems']['edges']:
  event = event['node']
 
  if (
- time_labeled_or_assigned is None
- and event['__typename'] == 'LabeledEvent'
- and 'cla:' not in event['label']['name']
+  time_labeled_or_assigned is None
+  and event['__typename'] == 'LabeledEvent'
+  and 'cla:' not in event['label']['name']
  ):
  time_labeled_or_assigned = _to_datetime(event['createdAt'])
 
  if (
- time_labeled_or_assigned is None
- and event['__typename'] == 'AssignedEvent'
+  time_labeled_or_assigned is None
+  and event['__typename'] == 'AssignedEvent'
  ):
  time_labeled_or_assigned = _to_datetime(event['createdAt'])
 
@@ -234,17 +241,19 @@ def _get_pr_features(prs):
  ready_for_review_at = _to_datetime(event['createdAt'])
 
  yield {
- 'created_at': created_at,
- 'ready_for_review_at': ready_for_review_at,
- 'time_labeled_or_assigned': time_labeled_or_assigned,
- 'time_merged_or_closed': time_merged_or_closed,
- 'time_review': time_review,
- 'pr_closed': pr['state'] != 'OPEN',
+  'created_at': created_at,
+  'ready_for_review_at': ready_for_review_at,
+  'time_labeled_or_assigned': time_labeled_or_assigned,
+  'time_merged_or_closed': time_merged_or_closed,
+  'time_review': time_review,
+  'pr_closed': pr['state'] != 'OPEN',
  }
 
+
 def _start_of_month(date: datetime) -> datetime:
  return date.replace(day=1, hour=0, minute=0, second=0, microsecond=0)
 
+
 def _shift_n_months(date: datetime, n: int) -> datetime:
  month = ((date.month + n - 1) % 12) + 1
 
@@ -258,14 +267,14 @@ def _shift_n_months(date: datetime, n: int) -> datetime:
 
 
 def _rolling_window(
- df: pd.DataFrame,
- f: Callable[[pd.DataFrame], pd.Series],
- window_size: int = 6,
- step: int = 1,
+  df: pd.DataFrame,
+  f: Callable[[pd.DataFrame], pd.Series],
+  window_size: int = 6,
+  step: int = 1,
 ) -> pd.DataFrame:
  # start of month of the first issue
  start: datetime = df.iloc[0]['created_at'].replace(
- day=1, hour=0, minute=0, second=0, microsecond=0
+  day=1, hour=0, minute=0, second=0, microsecond=0
  )
  end = _shift_n_months(start, window_size)
 
@@ -286,56 +295,66 @@ def _rolling_window(
 
  return df
 
+
 def _process_prs(df: pd.DataFrame) -> pd.Series:
  return pd.Series({
- 'pr_response_time': df['pr_response_time'].dt.days.mean(),
- 'pr_resolution_time': df['pr_resolution_time'].dt.days.mean(),
+  'pr_response_time': df['pr_response_time'].dt.days.mean(),
+  'pr_resolution_time': df['pr_resolution_time'].dt.days.mean(),
  })
 
+
 def _process_issues(df: pd.DataFrame) -> pd.Series:
  return pd.Series({
- 'issue_response_time': df['issue_response_time'].dt.days.mean(),
- 'issue_resolution_time': df['issue_resolution_time'].dt.days.mean(),
+  'issue_response_time': df['issue_response_time'].dt.days.mean(),
+  'issue_resolution_time': df['issue_resolution_time'].dt.days.mean(),
  })
 
-#-----------------------------------------------------------------------------
+
+# -----------------------------------------------------------------------------
 # main
-#-----------------------------------------------------------------------------
+# -----------------------------------------------------------------------------
 FLAGS = flags.FLAGS
 flags.DEFINE_string('repo_owner', 'google', 'User name or organization')
 flags.DEFINE_string('repo_name', 'flax', 'Name of the repository')
 
+
 def main(_):
  repo_owner: str = FLAGS.repo_owner
  repo_name: str = FLAGS.repo_name
 
  # Download issue data
  issues = GithubGrabber(
- '.github/analytics/issue_activity_since_date.gql',
- 'issues',
- repo_owner=repo_owner,
- repo_name=repo_name,
+  '.github/analytics/issue_activity_since_date.gql',
+  'issues',
+  repo_owner=repo_owner,
+  repo_name=repo_name,
  )
  issues.get()
 
  df_issues = df_issues0 = pd.DataFrame(list(_get_issues_features(issues.raw_data)))
- df_issues['issue_response_time'] = df_issues['time_labeled_or_converted'] - df_issues['created_at']
- df_issues['issue_resolution_time'] = df_issues['time_issue_closed'] - df_issues['created_at']
+ df_issues['issue_response_time'] = (
+ df_issues['time_labeled_or_converted'] - df_issues['created_at']
+ )
+ df_issues['issue_resolution_time'] = (
+ df_issues['time_issue_closed'] - df_issues['created_at']
+ )
 
  df_issues = _rolling_window(df_issues, _process_issues)
 
  prs = GithubGrabber(
- '.github/analytics/pr_data_query.gql',
- 'pullRequests',
- repo_owner=repo_owner,
- repo_name=repo_name,
+  '.github/analytics/pr_data_query.gql',
+  'pullRequests',
+  repo_owner=repo_owner,
+  repo_name=repo_name,
  )
  prs.get()
 
  df_prs = df_prs0 = pd.DataFrame(list(_get_pr_features(prs.raw_data)))
  time_response = df_prs[['time_labeled_or_assigned', 'time_review']].min(axis=1)
  df_prs['pr_response_time'] = time_response - df_prs['ready_for_review_at']
- df_prs['pr_resolution_time'] = df_prs['time_merged_or_closed'] - df_prs['ready_for_review_at']
+ df_prs['pr_resolution_time'] = (
+ df_prs['time_merged_or_closed'] - df_prs['ready_for_review_at']
+ )
 
  df_prs = _rolling_window(df_prs, _process_prs)
 
@@ -367,7 +386,6 @@ def main(_):
  plt.gca().xaxis.set_major_locator(plt.MaxNLocator(5))
  plt.gca().xaxis.set_major_formatter(mdates.DateFormatter('%b %Y'))
 
-
  # plot for isssue_response_time
  plt.figure()
  plt.plot(df_issues['period_end'], df_issues['issue_response_time'])
@@ -411,5 +429,6 @@ def main(_):
  # show plots
  plt.show()
 
+
 if __name__ == '__main__':
  app.run(main)

.pre-commit-config.yaml

@@ -18,7 +18,7 @@ repos:
  hooks:
  - id: check-toml
  - id: trailing-whitespace
- exclude: ^docs/.*\.md$
+ exclude: ^docs/.*\.md$|^dev/.*\.py$
 - repo: https://github.com/kynan/nbstripout
  rev: 0.6.1
  hooks:
@@ -29,3 +29,7 @@ repos:
  --extra-keys,
  "metadata.kernelspec metadata.vscode metadata.colab cell.metadata.executionInfo.user cell.metadata.executionInfo.user_tz cell.metadata.colab",
  ]
+- repo: https://github.com/google/pyink
+ rev: 23.5.0
+ hooks:
+ - id: pyink

dev/update_requirements.py

@@ -58,7 +58,8 @@
  '`--version="$(pip freeze | sed s/==/-/g) flax-0.3.6"` '
  '(note the flax version "override") '
  'or from the "install dependencies" step in the github build action '
- 'https://github.com/google/flax/actions/workflows/build.yml')
+ 'https://github.com/google/flax/actions/workflows/build.yml',
+)
 flags.mark_flag_as_required('versions')
 flags.DEFINE_bool('verbose', False, 'enables verbose output.')
 flags.DEFINE_list('ignore', ['jax'], 'packages not to add to requirements.')
@@ -67,22 +68,28 @@
 import_re = re.compile(r'(?:from|import)\s+(\w+)')
 # maps `import cv2` to `pip install opencv-python`
 pkg_map = {
- 'absl': 'absl-py',
- 'atari_py': 'atari-py',
- 'cv2': 'opencv-python',
- 'ml_collections': 'ml-collections',
- 'PIL': 'Pillow',
- 'tensorflow_datasets': 'tensorflow-datasets',
- 'tensorflow_text': 'tensorflow-text',
+  'absl': 'absl-py',
+  'atari_py': 'atari-py',
+  'cv2': 'opencv-python',
+  'ml_collections': 'ml-collections',
+  'PIL': 'Pillow',
+  'tensorflow_datasets': 'tensorflow-datasets',
+  'tensorflow_text': 'tensorflow-text',
 }
-standard_libs = set('codecs collections dataclasses datetime enum functools math multiprocessing itertools os pathlib random re sys tempfile time typing unicodedata warnings'.split(' '))
+standard_libs = set(
+ 'codecs collections dataclasses datetime enum functools math'
+ ' multiprocessing itertools os pathlib random re sys tempfile time typing'
+ ' unicodedata warnings'.split(' ')
+)
 
 
 def main(argv):
  del argv
 
  versions = {
- pkg_version[:pkg_version.rindex('-')]: pkg_version[pkg_version.rindex('-') + 1:]
+ pkg_version[: pkg_version.rindex('-')]: pkg_version[
+ pkg_version.rindex('-') + 1 :
+ ]
  for pkg_version in FLAGS.versions.replace('\n', ' ').split(' ')
  if '-' in pkg_version
  }
@@ -117,7 +124,8 @@ def main(argv):
  print(f'{requirements} -', end=' ')
  with requirements.open('w') as f:
  for pkg in sorted(pkgs, key=str.casefold):
- if pkg in ignore: continue
+ if pkg in ignore:
+ continue
  pkg = pkg_map.get(pkg, pkg)
  print(f'{pkg}-{versions[pkg]}', end=' ')
  f.write(f'{pkg}=={versions[pkg]}\n')