appengine/findit/model/flake/flake_analysis_request.py

# Copyright 2016 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.

from google.appengine.ext import ndb

from gae_libs.model.versioned_model import VersionedModel
from model.flake.master_flake_analysis import MasterFlakeAnalysis


class BuildStep(ndb.Model):
  """Represents a build step on Waterfall or Commit Queue.

  For a build step on Commit Queue, the matching Waterfall build step could be
  added if found.
  """

  # The build step in which a flake actually occurred.
  master_name = ndb.StringProperty(indexed=False)
  builder_name = ndb.StringProperty(indexed=False)
  build_number = ndb.IntegerProperty(indexed=False)
  step_name = ndb.StringProperty(indexed=False)

  # When the flake was reported on this step.
  reported_time = ndb.DateTimeProperty(indexed=False)

  # The matching build step on the Waterfall of the matching test configuration.
  wf_master_name = ndb.StringProperty(indexed=False)
  wf_builder_name = ndb.StringProperty(indexed=False)
  wf_build_number = ndb.IntegerProperty(indexed=False)
  wf_step_name = ndb.StringProperty(indexed=False)

  # Indicate whether the flake is run on Swarming.
  swarmed = ndb.BooleanProperty(indexed=False, default=False)

  # Indicate whether analysis on the step is supported.
  supported = ndb.BooleanProperty(indexed=False, default=False)

  # Indicate whether the analysis of flake on this configuration is scheduled.
  scheduled = ndb.BooleanProperty(indexed=False, default=False)

  # Step metadata of the step.
  step_metadata = ndb.JsonProperty(compressed=True)

  @staticmethod
  def _StripMasterPrefix(name):
    master_prefix = 'master.'
    if name.startswith(master_prefix):
      return name[len(master_prefix):]
    return name

  @staticmethod
  def Create(master_name, builder_name, build_number, step_name, reported_time):
    return BuildStep(
        master_name=BuildStep._StripMasterPrefix(master_name),
        builder_name=builder_name,
        build_number=build_number,
        step_name=step_name,
        reported_time=reported_time)

  @property
  def has_matching_waterfall_step(self):
    return None not in (self.wf_master_name, self.wf_builder_name,
                        self.wf_build_number, self.wf_step_name)

  @property
  def on_cq(self):
    """Returns True if the build step of the flake is on Commit Queue."""
    return self.master_name.startswith('tryserver.')


class FlakeAnalysisRequest(VersionedModel):
  """Represents a request to analyze a flake.

  The name of the flake will be the key, and the model is versioned.
  """

  # Name of the flake. Could be a step name, or a test name.
  # Assume there are no step and test with the same name.
  name = ndb.StringProperty(indexed=True)

  # Indicate whether the flake is a step or a test.
  is_step = ndb.BooleanProperty(indexed=True, default=True)

  # Indicate whether the flake is run on Swarming for some configuration.
  swarmed = ndb.BooleanProperty(indexed=False, default=False)

  # Indicate whether analysis on this flake is supported.
  supported = ndb.BooleanProperty(indexed=False, default=False)

  # The bug id for this flake on Monorail.
  bug_id = ndb.IntegerProperty(indexed=False)

  # The reporter of this bug, use triggering_source as an emun to define this
  # value. The triggering source FINDIT_PIPELINES should be used for when
  # Findit itself reports a bug.
  bug_reported_by = ndb.IntegerProperty(indexed=False)

  # The emails of users who request analysis of this flake.
  user_emails = ndb.StringProperty(indexed=False, repeated=True)

  # Whether the user emails have been obscured.
  user_emails_obscured = ndb.BooleanProperty(indexed=True, default=False)

  # When was the last edit of the email list.
  user_emails_last_edit = ndb.DateTimeProperty(indexed=True)

  # The build steps in which the flake occurred.
  build_steps = ndb.LocalStructuredProperty(
      BuildStep, compressed=True, repeated=True)

  # Executed analyses on different test configurations.
  analyses = ndb.KeyProperty(MasterFlakeAnalysis, repeated=True)

  # Arguments number differs from overridden method - pylint: disable=W0221
  @classmethod
  def Create(cls, name, is_step, bug_id):
    instance = super(cls, FlakeAnalysisRequest).Create(key=name)
    instance.name = name
    instance.is_step = is_step
    instance.bug_id = bug_id
    return instance

  def AddBuildStep(self, master_name, builder_name, build_number, step_name,
                   reported_time):
    """Adds a build step in which the flake is found."""
    for s in self.build_steps:
      if s.master_name == master_name and s.builder_name == builder_name:
        # For the same builder/tester, only analyze the earliest build.
        # TODO: re-evaluate cases that flakes might be re-introduced in between.
        if s.build_number <= build_number:
          return False
        s.build_number = build_number
        s.reported_time = reported_time
        return True

    self.build_steps.append(
        BuildStep.Create(master_name, builder_name, build_number, step_name,
                         reported_time))
    return True

  def CopyFrom(self, other):
    """Copies all states from the given request."""
    assert isinstance(other, FlakeAnalysisRequest)
    self.is_step = other.is_step
    self.bug_id = other.bug_id
    self.user_emails = other.user_emails
    self.build_steps = other.build_steps
    self.analyses = other.analyses

  @property
  def on_cq(self):
    """Returns True if the flake is on Commit Queue."""
    return any(step.on_cq for step in self.build_steps)

  def _GetNormalizedConfigurationNames(self, master_name, builder_name):
    for build_step in self.build_steps:
      if ((build_step.master_name == master_name and
           build_step.builder_name == builder_name) or
          (build_step.wf_master_name == master_name and
           build_step.wf_builder_name == builder_name)):
        return build_step.wf_master_name, build_step.wf_builder_name
    return None, None

  def FindMatchingAnalysisForConfiguration(self, master_name, builder_name):
    # Returns the analysis that corresponds to the requested master and builder.
    normalized_master_name, normalized_builder_name = (
        self._GetNormalizedConfigurationNames(master_name, builder_name))

    if not normalized_master_name or not normalized_builder_name:
      return None

    for analysis_key in self.analyses:
      analysis_master_name, analysis_builder_name = (
          MasterFlakeAnalysis.GetBuildConfigurationFromKey(analysis_key))

      if (analysis_master_name == normalized_master_name and
          analysis_builder_name == normalized_builder_name):
        return analysis_key.get()

    return None

  def Update(self, **kwargs):
    """Updates fields according to what's specified in kwargs.

      Fields specified in kwargs will be updated accordingly, while those not
      present in kwargs will be untouched.

    Args:
      **kwargs (dict): The keys in kwargs should match the name of the field you
          want to update, and the values should be the value you update it to.
    """
    # TODO(crbug.com/772156): Refactor this into a base model.
    any_changes = False

    for arg, value in kwargs.iteritems():
      current_value = getattr(self, arg, None)
      if current_value != value:
        setattr(self, arg, value)
        any_changes = True

    if any_changes:
      self.put()