esp-idf/docs/build_docs.py

#!/usr/bin/env python3
# coding=utf-8
#
# Top-level docs builder
#
# This is just a front-end to sphinx-build that can call it multiple times for different language/target combinations
#
# Will build out to _build/LANG/TARGET by default
#
# Specific custom docs functionality should be added in conf_common.py or in a Sphinx extension, not here.
#
# Copyright 2020 Espressif Systems (Shanghai) PTE LTD
#
# 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.
#
from __future__ import print_function
import argparse
import locale
import math
import multiprocessing
import os
import os.path
import subprocess
import sys
import re
from collections import namedtuple

LANGUAGES = ["en", "zh_CN"]
TARGETS = ["esp32", "esp32s2"]

SPHINX_WARN_LOG = "sphinx-warning-log.txt"
SPHINX_SANITIZED_LOG = "sphinx-warning-log-sanitized.txt"
SPHINX_KNOWN_WARNINGS = os.path.join(os.environ["IDF_PATH"], "docs", "sphinx-known-warnings.txt")

DXG_WARN_LOG = "doxygen-warning-log.txt"
DXG_SANITIZED_LOG = "doxygen-warning-log-sanitized.txt"
DXG_KNOWN_WARNINGS = os.path.join(os.environ["IDF_PATH"], "docs", "doxygen-known-warnings.txt")

LogMessage = namedtuple("LogMessage", "original_text sanitized_text")

languages = LANGUAGES
targets = TARGETS


def main():
    # check Python dependencies for docs
    try:
        subprocess.check_call([sys.executable,
                               os.path.join(os.environ["IDF_PATH"],
                                            "tools",
                                            "check_python_dependencies.py"),
                               "-r",
                               "{}/docs/requirements.txt".format(os.environ["IDF_PATH"])
                               ])
    except subprocess.CalledProcessError:
        raise SystemExit(2)  # stdout will already have these errors

    # This is not the only way to make sure that all files opened by Python are treated as UTF-8, but the other way is passing encoding='utf-8' to all open()
    # functions and this way makes Python 2 compatibility really tough if there is any code that assumes text files contain strings (kconfiglib assumes this).
    # The reason for that is that you need to import io.open() to support the encoding argument on Python 2, and this function always uses Py2's unicode
    # type not the str type.
    if 'UTF-8' not in locale.getlocale():
        raise RuntimeError("build_docs.py requires the default locale's encoding to be UTF-8. " +
                           "Setting environment variable LC_ALL=C.UTF-8 when running build_docs.py may be enough to fix this.")

    parser = argparse.ArgumentParser(description='build_docs.py: Build IDF docs', prog='build_docs.py')

    parser.add_argument("--language", "-l", choices=LANGUAGES, required=False)
    parser.add_argument("--target", "-t", choices=TARGETS, required=False)
    parser.add_argument("--build-dir", "-b", type=str, default="_build")
    parser.add_argument("--builders", "-bs", nargs='+', type=str, default=["html"],
                        help="List of builders for Sphinx, e.g. html or latex, for latex a PDF is also generated")
    parser.add_argument("--sphinx-parallel-builds", "-p", choices=["auto"] + [str(x) for x in range(8)],
                        help="Parallel Sphinx builds - number of independent Sphinx builds to run", default="auto")
    parser.add_argument("--sphinx-parallel-jobs", "-j", choices=["auto"] + [str(x) for x in range(8)],
                        help="Sphinx parallel jobs argument - number of threads for each Sphinx build to use", default="1")
    parser.add_argument("--input-docs", "-i", nargs='+', default=[""],
                        help="List of documents to build relative to the doc base folder, i.e. the language folder. Defaults to all documents")

    action_parsers = parser.add_subparsers(dest='action')

    build_parser = action_parsers.add_parser('build', help='Build documentation')
    build_parser.add_argument("--check-warnings-only", "-w", action='store_true')

    action_parsers.add_parser('linkcheck', help='Check links (a current IDF revision should be uploaded to GitHub)')

    action_parsers.add_parser('gh-linkcheck', help='Checking for hardcoded GitHub links')

    args = parser.parse_args()

    global languages
    if args.language is None:
        print("Building all languages")
        languages = LANGUAGES
    else:
        languages = [args.language]

    global targets
    if args.target is None:
        print("Building all targets")
        targets = TARGETS
    else:
        targets = [args.target]

    if args.action == "build" or args.action is None:
        if args.action is None:
            args.check_warnings_only = False
        sys.exit(action_build(args))

    if args.action == "linkcheck":
        sys.exit(action_linkcheck(args))

    if args.action == "gh-linkcheck":
        sys.exit(action_gh_linkcheck(args))


def parallel_call(args, callback):
    num_sphinx_builds = len(languages) * len(targets)
    num_cpus = multiprocessing.cpu_count()

    if args.sphinx_parallel_builds == "auto":
        # at most one sphinx build per CPU, up to the number of CPUs
        args.sphinx_parallel_builds = min(num_sphinx_builds, num_cpus)
    else:
        args.sphinx_parallel_builds = int(args.sphinx_parallel_builds)

    # Force -j1 because sphinx works incorrectly
    args.sphinx_parallel_jobs = 1
    if args.sphinx_parallel_jobs == "auto":
        # N CPUs per build job, rounded up - (maybe smarter to round down to avoid contention, idk)
        args.sphinx_parallel_jobs = int(math.ceil(num_cpus / args.sphinx_parallel_builds))
    else:
        args.sphinx_parallel_jobs = int(args.sphinx_parallel_jobs)

    print("Will use %d parallel builds and %d jobs per build" % (args.sphinx_parallel_builds, args.sphinx_parallel_jobs))
    pool = multiprocessing.Pool(args.sphinx_parallel_builds)

    if args.sphinx_parallel_jobs > 1:
        print("WARNING: Sphinx parallel jobs currently produce incorrect docs output with Sphinx 1.8.5")

    # make a list of all combinations of build_docs() args as tuples
    #
    # there's probably a fancy way to do this with itertools but this way is actually readable
    entries = []
    for target in targets:
        for language in languages:
            build_dir = os.path.realpath(os.path.join(args.build_dir, language, target))
            entries.append((language, target, build_dir, args.sphinx_parallel_jobs, args.builders, args.input_docs))

    print(entries)
    errcodes = pool.map(callback, entries)
    print(errcodes)

    is_error = False
    for ret in errcodes:
        if ret != 0:
            print("\nThe following language/target combinations failed to build:")
            is_error = True
            break
    if is_error:
        for ret, entry in zip(errcodes, entries):
            if ret != 0:
                print("language: %s, target: %s, errcode: %d" % (entry[0], entry[1], ret))
        # Don't re-throw real error code from each parallel process
        return 1
    else:
        return 0


def sphinx_call(language, target, build_dir, sphinx_parallel_jobs, buildername, input_docs):
    # Note: because this runs in a multiprocessing Process, everything which happens here should be isolated to a single process
    # (ie it doesn't matter if Sphinx is using global variables, as they're it's own copy of the global variables)

    # wrap stdout & stderr in a way that lets us see which build_docs instance they come from
    #
    # this doesn't apply to subprocesses, they write to OS stdout & stderr so no prefix appears
    prefix = "%s/%s: " % (language, target)

    print("Building in build_dir: %s" % (build_dir))
    try:
        os.makedirs(build_dir)
    except OSError:
        pass

    environ = {}
    environ.update(os.environ)
    environ['BUILDDIR'] = build_dir

    args = [sys.executable, "-u", "-m", "sphinx.cmd.build",
            "-j", str(sphinx_parallel_jobs),
            "-b", buildername,
            "-d", os.path.join(build_dir, "doctrees"),
            "-w", SPHINX_WARN_LOG,
            "-t", target,
            "-D", "idf_target={}".format(target),
            "-D", "docs_to_build={}".format(",". join(input_docs)),
            os.path.join(os.path.abspath(os.path.dirname(__file__)), language),  # srcdir for this language
            os.path.join(build_dir, buildername)                    # build directory
            ]

    saved_cwd = os.getcwd()
    os.chdir(build_dir)  # also run sphinx in the build directory
    print("Running '%s'" % (" ".join(args)))

    ret = 1
    try:
        # Note: we can't call sphinx.cmd.build.main() here as multiprocessing doesn't est >1 layer deep
        # and sphinx.cmd.build() also does a lot of work in the calling thread, especially for j ==1,
        # so using a Pyhthon thread for this part is  a poor option (GIL)
        p = subprocess.Popen(args, stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
        for c in iter(lambda: p.stdout.readline(), b''):
            sys.stdout.write(prefix)
            sys.stdout.write(c.decode('utf-8'))
        ret = p.wait()
        assert (ret is not None)
        sys.stdout.flush()
    except KeyboardInterrupt:  # this seems to be the only way to get Ctrl-C to kill everything?
        p.kill()
        os.chdir(saved_cwd)
        return 130  # FIXME It doesn't return this errorcode, why? Just prints stacktrace
    os.chdir(saved_cwd)
    return ret


def action_build(args):
    if not args.check_warnings_only:
        ret = parallel_call(args, call_build_docs)
        if ret != 0:
            return ret


def call_build_docs(entry):
    (language, target, build_dir, sphinx_parallel_jobs, builders, input_docs) = entry
    for buildername in builders:
        ret = sphinx_call(language, target, build_dir, sphinx_parallel_jobs, buildername, input_docs)

        # Warnings are checked after each builder as logs are overwritten
        # check Doxygen warnings:
        ret += check_docs(language, target,
                          log_file=os.path.join(build_dir, DXG_WARN_LOG),
                          known_warnings_file=DXG_KNOWN_WARNINGS,
                          out_sanitized_log_file=os.path.join(build_dir, DXG_SANITIZED_LOG))
        # check Sphinx warnings:
        ret += check_docs(language, target,
                          log_file=os.path.join(build_dir, SPHINX_WARN_LOG),
                          known_warnings_file=SPHINX_KNOWN_WARNINGS,
                          out_sanitized_log_file=os.path.join(build_dir, SPHINX_SANITIZED_LOG))

        if ret != 0:
            return ret

    # Build PDF from tex
    if 'latex' in builders:
        latex_dir = os.path.join(build_dir, "latex")
        ret = build_pdf(language, target, latex_dir)

    return ret


def build_pdf(language, target, latex_dir):
    # Note: because this runs in a multiprocessing Process, everything which happens here should be isolated to a single process

    # wrap stdout & stderr in a way that lets us see which build_docs instance they come from
    #
    # this doesn't apply to subprocesses, they write to OS stdout & stderr so no prefix appears
    prefix = "%s/%s: " % (language, target)

    print("Building PDF in latex_dir: %s" % (latex_dir))

    saved_cwd = os.getcwd()
    os.chdir(latex_dir)

    # Based on read the docs PDFBuilder
    rcfile = 'latexmkrc'

    cmd = [
        'latexmk',
        '-r',
        rcfile,
        '-pdf',
        # When ``-f`` is used, latexmk will continue building if it
        # encounters errors. We still receive a failure exit code in this
        # case, but the correct steps should run.
        '-f',
        '-dvi-',    # dont generate dvi
        '-ps-',     # dont generate ps
        '-interaction=nonstopmode',
        '-quiet',
        '-outdir=build',
    ]

    try:
        p = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
        for c in iter(lambda: p.stdout.readline(), b''):
            sys.stdout.write(prefix)
            sys.stdout.write(c.decode('utf-8'))
        ret = p.wait()
        assert (ret is not None)
        sys.stdout.flush()
    except KeyboardInterrupt:  # this seems to be the only way to get Ctrl-C to kill everything?
        p.kill()
        os.chdir(saved_cwd)
        return 130  # FIXME It doesn't return this errorcode, why? Just prints stacktrace
    os.chdir(saved_cwd)

    return ret


SANITIZE_FILENAME_REGEX = re.compile("[^:]*/([^/:]*)(:.*)")
SANITIZE_LINENUM_REGEX = re.compile("([^:]*)(:[0-9]+:)(.*)")


def sanitize_line(line):
    """
    Clear a log message from insignificant parts

    filter:
        - only filename, no path at the beginning
        - no line numbers after the filename
    """

    line = re.sub(SANITIZE_FILENAME_REGEX, r'\1\2', line)
    line = re.sub(SANITIZE_LINENUM_REGEX, r'\1:line:\3', line)
    return line


def check_docs(language, target, log_file, known_warnings_file, out_sanitized_log_file):
    """
    Check for Documentation warnings in `log_file`: should only contain (fuzzy) matches to `known_warnings_file`

    It prints all unknown messages with `target`/`language` prefix
    It leaves `out_sanitized_log_file` file for observe and debug
    """

    # Sanitize all messages
    all_messages = list()
    with open(log_file) as f, open(out_sanitized_log_file, 'w') as o:
        for line in f:
            sanitized_line = sanitize_line(line)
            all_messages.append(LogMessage(line, sanitized_line))
            o.write(sanitized_line)

    known_messages = list()
    with open(known_warnings_file) as k:
        for known_line in k:
            known_messages.append(known_line)

    if "doxygen" in known_warnings_file:
        # Clean a known Doxygen limitation: it's expected to always document anonymous
        # structs/unions but we don't do this in our docs, so filter these all out with a regex
        # (this won't match any named field, only anonymous members -
        # ie the last part of the field is is just <something>::@NUM not <something>::name)
        RE_ANONYMOUS_FIELD = re.compile(r".+:line: warning: parameters of member [^:\s]+(::[^:\s]+)*(::@\d+)+ are not \(all\) documented")
        all_messages = [msg for msg in all_messages if not re.match(RE_ANONYMOUS_FIELD, msg.sanitized_text)]

    # Collect all new messages that are not match with the known messages.
    # The order is an important.
    new_messages = list()
    known_idx = 0
    for msg in all_messages:
        try:
            known_idx = known_messages.index(msg.sanitized_text, known_idx)
        except ValueError:
            new_messages.append(msg)

    if new_messages:
        print("\n%s/%s: Build failed due to new/different warnings (%s):\n" % (language, target, log_file))
        for msg in new_messages:
            print("%s/%s: %s" % (language, target, msg.original_text), end='')
        print("\n%s/%s: (Check files %s and %s for full details.)" % (language, target, known_warnings_file, log_file))
        return 1

    return 0


def action_linkcheck(args):
    return parallel_call(args, call_linkcheck)


def call_linkcheck(entry):
    # Remove the last entry which the buildername, since the linkcheck builder is not supplied through the builder list argument
    return sphinx_call(*entry[:4], buildername="linkcheck")


# https://github.com/espressif/esp-idf/tree/
# https://github.com/espressif/esp-idf/blob/
# https://github.com/espressif/esp-idf/raw/
GH_LINK_RE = r"https://github.com/espressif/esp-idf/(?:tree|blob|raw)/[^\s]+"

# we allow this one link, because we always want users to see the latest support policy
GH_LINK_ALLOWED = ["https://github.com/espressif/esp-idf/blob/master/SUPPORT_POLICY.md"]


def action_gh_linkcheck(args):
    print("Checking for hardcoded GitHub links\n")

    github_links = []

    docs_dir = os.path.relpath(os.path.dirname(__file__))
    for root, _, files in os.walk(docs_dir):
        if "_build" in root:
            continue
        files = [os.path.join(root, f) for f in files if f.endswith(".rst")]
        for path in files:
            with open(path, "r") as f:
                for link in re.findall(GH_LINK_RE, f.read()):
                    if link not in GH_LINK_ALLOWED:
                        github_links.append((path, link))

    if github_links:
        for path, link in github_links:
            print("%s: %s" % (path, link))
        print("WARNING: Some .rst files contain hardcoded Github links.")
        print("Please check above output and replace links with one of the following:")
        print("- :idf:`dir` - points to directory inside ESP-IDF")
        print("- :idf_file:`file` - points to file inside ESP-IDF")
        print("- :idf_raw:`file` - points to raw view of the file inside ESP-IDF")
        print("- :component:`dir` - points to directory inside ESP-IDF components dir")
        print("- :component_file:`file` - points to file inside ESP-IDF components dir")
        print("- :component_raw:`file` - points to raw view of the file inside ESP-IDF components dir")
        print("- :example:`dir` - points to directory inside ESP-IDF examples dir")
        print("- :example_file:`file` - points to file inside ESP-IDF examples dir")
        print("- :example_raw:`file` - points to raw view of the file inside ESP-IDF examples dir")
        print("These link types will point to the correct GitHub version automatically")
        return 1
    else:
        print("No hardcoded links found")
        return 0


if __name__ == "__main__":
    main()