# Copyright (C) 2022 The Android Open Source Project # # 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. import csv import dataclasses import datetime import enum import functools import json import logging import os import re import subprocess import sys from datetime import date from pathlib import Path import textwrap from typing import Callable from typing import Final from typing import Generator from typing import TypeVar INDICATOR_FILE: Final[str] = "build/soong/soong_ui.bash" # metrics.csv is written to but not read by this tool. # It's supposed to be viewed as a spreadsheet that compiles data from multiple # builds to be analyzed by other external tools. METRICS_TABLE: Final[str] = "metrics.csv" RUN_DIR_PREFIX: Final[str] = "run" BUILD_INFO_JSON: Final[str] = "build_info.json" @functools.cache def _is_important(column) -> bool: patterns = { "actions", r"build_ninja_(?:hash|size)", "build_type", "cquery_out_size", "description", "log", r"mixed\.enabled", "targets", # the following are time-based values "bp2build", r"kati/kati (?:build|package)", "ninja/ninja", "soong/soong", r"soong_build/\*(?:\.bazel)?", "symlink_forest", "time", } for pattern in patterns: if re.fullmatch(pattern, column): return True return False class BuildResult(enum.Enum): SUCCESS = enum.auto() FAILED = enum.auto() TEST_FAILURE = enum.auto() class BuildType(enum.Enum): # see https://docs.python.org/3/library/enum.html#enum.Enum._ignore_ _ignore_ = "_soong_cmd" # _soong_cmd_ will not be listed as an enum constant because of `_ignore_` _soong_cmd = ["build/soong/soong_ui.bash", "--make-mode", "--skip-soong-tests"] SOONG_ONLY = [*_soong_cmd, "BUILD_BROKEN_DISABLE_BAZEL=true"] MIXED_PROD = [*_soong_cmd, "--bazel-mode"] MIXED_STAGING = [*_soong_cmd, "--bazel-mode-staging"] B_BUILD = ["build/bazel/bin/b", "build"] B_ANDROID = [*B_BUILD, "--config=android"] @staticmethod def from_flag(s: str) -> list["BuildType"]: chosen: list[BuildType] = [] for e in BuildType: if s.lower() in e.name.lower(): chosen.append(e) if len(chosen) == 0: raise RuntimeError(f"no such build type: {s}") return chosen def to_flag(self): return self.name.lower() CURRENT_BUILD_TYPE: BuildType """global state capturing what the current build type is""" @dataclasses.dataclass class BuildInfo: actions: int bp_size_total: int build_ninja_hash: str # hash build_ninja_size: int build_result: BuildResult build_type: BuildType bz_size_total: int cquery_out_size: int description: str product: str targets: tuple[str, ...] time: datetime.timedelta tag: str = None rebuild: bool = False warmup: bool = False class CustomEncoder(json.JSONEncoder): def default(self, obj): if isinstance(obj, BuildInfo): return self.default(dataclasses.asdict(obj)) if isinstance(obj, dict): return {k: v for k, v in obj.items() if v is not None} if isinstance(obj, datetime.timedelta): return hhmmss(obj, decimal_precision=True) if isinstance(obj, enum.Enum): return obj.name return json.JSONEncoder.default(self, obj) def get_csv_columns_cmd(d: Path) -> str: """ :param d: the log directory :return: a quick shell command to view columns in metrics.csv """ csv_file = d.joinpath(METRICS_TABLE).resolve() return f'head -n 1 "{csv_file}" | sed "s/,/\\n/g" | less -N' def get_cmd_to_display_tabulated_metrics(d: Path, ci_mode: bool) -> str: """ :param d: the log directory :param ci_mode: if true all top-level events are displayed :return: a quick shell command to view some collected metrics """ csv_file = d.joinpath(METRICS_TABLE) headers: list[str] = [] if csv_file.exists(): with open(csv_file) as r: reader = csv.DictReader(r) headers = reader.fieldnames or [] cols: list[int] = [i + 1 for i, h in enumerate(headers) if _is_important(h)] if ci_mode: # ci mode contains all information about the top level events for i, h in enumerate(headers): if re.match(r"^\w+/[^.]+$", h) and i not in cols: cols.append(i) if len(cols) == 0: # syntactically correct command even if the file doesn't exist cols.append(1) f = ",".join(str(i) for i in cols) # the sed invocations are to account for # https://man7.org/linux/man-pages/man1/column.1.html#BUGS # example: if a row were `,,,hi,,,,` # the successive sed conversions would be # `,,,hi,,,,` => # `,--,,hi,--,,--,` => # `,--,--,hi,--,--,--,` => # `--,--,--,hi,--,--,--,` => # `--,--,--,hi,--,--,--,--` # Note sed doesn't support lookahead or lookbehinds return textwrap.dedent( f"""\ grep -v "WARMUP\\|rebuild-" "{csv_file}" | \\ sed "s/,,/,--,/g" | \\ sed "s/,,/,--,/g" | \\ sed "s/^,/--,/" | \\ sed "s/,$/,--/" | \\ cut -d, -f{f} | column -t -s,""" ) @functools.cache def get_top_dir(d: Path = Path(".").resolve()) -> Path: """Get the path to the root of the Android source tree""" top_dir = os.environ.get("ANDROID_BUILD_TOP") if top_dir: logging.info("ANDROID BUILD TOP = %s", d) return Path(top_dir).resolve() logging.debug("Checking if Android source tree root is %s", d) if d.parent == d: sys.exit( "Unable to find ROOT source directory, specifically," f"{INDICATOR_FILE} not found anywhere. " "Try `m nothing` and `repo sync`" ) if d.joinpath(INDICATOR_FILE).is_file(): logging.info("ANDROID BUILD TOP assumed to be %s", d) return d return get_top_dir(d.parent) @functools.cache def get_out_dir() -> Path: out_dir = os.environ.get("OUT_DIR") return Path(out_dir).resolve() if out_dir else get_top_dir().joinpath("out") @functools.cache def get_default_log_dir() -> Path: return get_top_dir().parent.joinpath(f'timing-{date.today().strftime("%b%d")}') def is_interactive_shell() -> bool: return ( sys.__stdin__.isatty() and sys.__stdout__.isatty() and sys.__stderr__.isatty() ) # see test_next_path_helper() for examples def _next_path_helper(basename: str) -> str: def padded_suffix(i: int) -> str: return f"{i:03d}" # find the sequence digits following "-" and preceding the file extension name = re.sub( r"(?<=-)(?P\d+)$", lambda d: padded_suffix(int(d.group("suffix")) + 1), basename, ) if name == basename: # basename didn't have any numeric suffix name = f"{name}-{padded_suffix(1)}" return name def next_path(path: Path) -> Generator[Path, None, None]: """ Generator for indexed paths :returns a new Path with an increasing number suffix to the name e.g. _to_file('a.txt') = a-5.txt (if a-4.txt already exists) """ while True: name = _next_path_helper(path.name) path = path.parent.joinpath(name) if not path.exists(): yield path def has_uncommitted_changes() -> bool: """ effectively a quick 'repo status' that fails fast if any project has uncommitted changes """ for cmd in ["diff", "diff --staged"]: diff = subprocess.run( f"repo forall -c git {cmd} --quiet --exit-code", cwd=get_top_dir(), shell=True, text=True, capture_output=True, ) if diff.returncode != 0: logging.error(diff.stderr) return True return False def hhmmss(t: datetime.timedelta, decimal_precision: bool = False) -> str: """pretty prints time periods, prefers mm:ss.sss and resorts to hh:mm:ss.sss only if t >= 1 hour. Examples(non_decimal_precision): 02:12, 1:12:13 Examples(decimal_precision): 02:12.231, 00:00.512, 00:01:11.321, 1:12:13.121 See unit test for more examples.""" h, f = divmod(t.seconds, 60 * 60) m, f = divmod(f, 60) s = f + t.microseconds / 1000_000 if decimal_precision: return f"{h}:{m:02d}:{s:06.3f}" if h else f"{m:02d}:{s:06.3f}" else: return f"{h}:{m:02}:{s:02.0f}" if h else f"{m:02}:{s:02.0f}" def period_to_seconds(s: str) -> float: """converts a time period into seconds. The input is expected to be in the format used by hhmmss(). Example: 02:04 -> 125 See unit test for more examples.""" if s == "": return 0.0 acc = 0.0 while True: [left, *right] = s.split(":", 1) acc = acc * 60 + float(left) if right: s = right[0] else: return acc R = TypeVar("R") def groupby(xs: list[R], key: Callable[[R], str]) -> dict[str, list[R]]: grouped: dict[str, list[R]] = {} for x in xs: grouped.setdefault(key(x), []).append(x) return grouped