tools/ctexplain/ctexplain.py

# Lint as: python3
# Copyright 2020 The Bazel Authors. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#    http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""ctexplain: how does configuration affect build graphs?

This is a swiss army knife tool that tries to explain why build graphs are the
size they are and how build flags, configuration transitions, and dependency
structures affect that.

This can help developers use flags and transitions with minimal memory and
maximum build speed.

Usage:

  $ ctexplain [--analysis=...] -b "<targets_to_build> [build flags]"

Example:

  $ ctexplain -b "//mypkg:mybinary --define MY_FEATURE=1"

Relevant terms in https://docs.bazel.build/versions/main/glossary.html:
  "target", "configuration", "analysis phase", "configured target",
  "configuration trimming", "transition"

TODO(gregce): link to proper documentation for full details.
"""
from typing import Callable
from typing import Tuple

# Do not edit this line. Copybara replaces it with PY2 migration helper.
from absl import app
from absl import flags
from dataclasses import dataclass

# Do not edit this line. Copybara replaces it with PY2 migration helper..third_party.bazel.tools.ctexplain.analyses.summary as summary
from tools.ctexplain.bazel_api import BazelApi
# Do not edit this line. Copybara replaces it with PY2 migration helper..third_party.bazel.tools.ctexplain.lib as lib
from tools.ctexplain.types import ConfiguredTarget
# Do not edit this line. Copybara replaces it with PY2 migration helper..third_party.bazel.tools.ctexplain.util as util

FLAGS = flags.FLAGS


@dataclass(frozen=True)
class Analysis():
  """Supported analysis type."""
  # The value in --analysis=<value> that triggers this analysis.
  key: str
  # The function that invokes this analysis.
  exec: Callable[[Tuple[ConfiguredTarget, ...]], None]
  # User-friendly analysis description.
  description: str

available_analyses = [
    Analysis(
        "summary",
        lambda x: summary.report(summary.analyze(x)),
        "summarizes build graph size and how trimming could help"
    ),
    Analysis(
        "culprits",
        lambda x: print("this analysis not yet implemented"),
        "shows which flags unnecessarily fork configured targets. These\n"
        + "are conceptually mergeable."
    ),
    Analysis(
        "forked_targets",
        lambda x: print("this analysis not yet implemented"),
        "ranks targets by how many configured targets they\n"
        + "create. These may be legitimate forks (because they behave "
        + "differently with\n different flags) or identical clones that are "
        + "conceptually mergeable."
    ),
    Analysis(
        "cloned_targets",
        lambda x: print("this analysis not yet implemented"),
        "ranks targets by how many behavior-identical configured\n targets "
        + "they produce. These are conceptually mergeable."
    )
]

# Available analyses, keyed by --analysis=<value> triggers.
analyses = {analysis.key: analysis for analysis in available_analyses}


# Command-line flag registration:


def _render_analysis_help_text() -> str:
  """Pretty-prints help text for available analyses."""
  return "\n".join(f'- "{name}": {analysis.description}'
                   for name, analysis in analyses.items())

flags.DEFINE_list("analysis", ["summary"], f"""
Analyses to run. May be any comma-separated combination of

{_render_analysis_help_text()}
""")

flags.register_validator(
    "analysis",
    lambda flag_value: all(name in analyses for name in flag_value),
    message=f'available analyses: {", ".join(analyses.keys())}')

flags.DEFINE_multi_string(
    "build", [],
    """command-line invocation of the build to analyze. For example:
"//foo --define a=b". If listed multiple times, this is a "multi-build
analysis" that measures how much distinct builds can share subgraphs""",
    short_name="b")


# Core program logic:


def _get_build_flags(cmdline: str) -> Tuple[Tuple[str, ...], Tuple[str, ...]]:
  """Parses a build invocation command line.

  Args:
    cmdline: raw build invocation string. For example: "//foo --cpu=x86"

  Returns:
    Tuple of ((target labels to build), (build flags))
  """
  cmdlist = cmdline.split()
  labels = [arg for arg in cmdlist if arg.startswith("//")]
  build_flags = [arg for arg in cmdlist if not arg.startswith("//")]
  return (tuple(labels), tuple(build_flags))


def main(argv):
  del argv  # Satisfy py linter's "unused" warning.
  if not FLAGS.build:
    exit("ctexplain: build efficiency measurement tool. Add --help "
         + "for usage.")
  elif len(FLAGS.build) > 1:
    exit("TODO(gregce): support multi-build shareability analysis")

  (labels, build_flags) = _get_build_flags(FLAGS.build[0])
  build_desc = ",".join(labels)
  with util.ProgressStep(f"Collecting configured targets for {build_desc}"):
    cts = lib.analyze_build(BazelApi(), labels, build_flags)
  for analysis in FLAGS.analysis:
    analyses[analysis].exec(cts)


if __name__ == "__main__":
  app.run(main)