zulip/tools/check-templates

#!/usr/bin/env python3
import argparse
import logging
import sys

# check for the venv
from lib import sanity_check
from lib.html_branches import build_id_dict
from lib.pretty_print import validate_indent_html
from lib.template_parser import validate

sanity_check.check_venv(__file__)

from typing import Dict, Iterable, List

from zulint import lister

EXCLUDED_FILES = [
    ## Test data Files for testing modules in tests
    "tools/tests/test_template_data",
    # Our parser doesn't handle the way its conditionals are layered
    "templates/zerver/emails/missed_message.source.html",
    # Previously unchecked and our parser doesn't like its indentation
    "static/assets/icons/template.hbs",
]


def check_our_files(modified_only: bool, all_dups: bool, fix: bool, targets: List[str]) -> None:
    by_lang = lister.list_files(
        targets=targets,
        modified_only=args.modified,
        ftypes=["hbs", "html"],
        group_by_ftype=True,
        exclude=EXCLUDED_FILES,
    )

    check_handlebar_templates(by_lang["hbs"], fix)
    check_html_templates(by_lang["html"], all_dups, fix)


def check_html_templates(templates: Iterable[str], all_dups: bool, fix: bool) -> None:
    # Our files with .html extensions are usually for Django, but we also
    # have a few static .html files.
    logging.basicConfig(format="%(levelname)s:%(message)s")
    templates = sorted(fn for fn in templates)
    # Use of lodash templates <%= %>.
    if "templates/zerver/team.html" in templates:
        templates.remove("templates/zerver/team.html")

    def check_for_duplicate_ids(templates: List[str]) -> Dict[str, List[str]]:
        template_id_dict = build_id_dict(templates)
        # TODO: Clean up these cases of duplicate ids in the code
        IGNORE_IDS = [
            "api-example-tabs",
            "errors",
            "error-message-box",
            "email",
            "messages",
            "registration",
            "pw_strength",
            "id_password",
            "top_navbar",
            "id_email",
            "id_terms",
            "send_confirm",
            "register",
            "footer",
            "charged_amount",
            "change-plan-status",
            # Temporary while we have searchbox forked
            "search_exit",
            "search_query",
            "message_view_header",
            "search_arrows",
            "searchbox_form",
            "searchbox",
        ]
        bad_ids_dict = {
            ids: fns
            for ids, fns in template_id_dict.items()
            if (ids not in IGNORE_IDS) and len(fns) > 1
        }

        if all_dups:
            ignorable_ids_dict = {
                ids: fns
                for ids, fns in template_id_dict.items()
                if ids in IGNORE_IDS and len(fns) > 1
            }

            for ids, fns in ignorable_ids_dict.items():
                logging.warning(
                    "Duplicate ID(s) detected :Id '" + ids + "' present at following files:"
                )
                for fn in fns:
                    print(fn)

        for ids, fns in bad_ids_dict.items():
            logging.error("Duplicate ID(s) detected :Id '" + ids + "' present at following files:")
            for fn in fns:
                print(fn)
        return bad_ids_dict

    bad_ids_list: List[str] = []
    archive_templates = [fn for fn in templates if "templates/zerver/archive" in fn]
    templates = [fn for fn in templates if "templates/zerver/archive" not in fn]

    bad_ids_list += list(check_for_duplicate_ids(archive_templates).keys())
    bad_ids_list += list(check_for_duplicate_ids(templates).keys())

    if bad_ids_list:
        print("Exiting--please clean up all duplicates before running this again.")
        sys.exit(1)

    for fn in templates:
        # Many of our Django templates have strange indentation.  The
        # indentation errors are often harmless, even stylistically
        # harmless, but they tend to be in files that might be old
        # and might eventually require more scrutiny for things like
        # localization.  See GitHub #1236.
        bad_files = [
            # These use various whitespace-dependent formatting that
            # prevent cleaning them.
            "templates/corporate/zephyr-mirror.html",
        ]
        validate(fn=fn, check_indent=(fn not in bad_files))

    # Ignore these files since these have not been cleaned yet :/
    IGNORE_FILES = [
        # zephyr-mirror.html has some whitespace-dependent formatting
        # for code blocks that prevent cleaning it.  Might make sense
        # to convert it to a /help/ Markdown article.
        "templates/corporate/zephyr-mirror.html",
    ]
    # TODO: Clean these files
    for fn in templates:
        if fn not in IGNORE_FILES:
            if not validate_indent_html(fn, fix):
                sys.exit(1)


def check_handlebar_templates(templates: Iterable[str], fix: bool) -> None:
    # Check all our handlebars templates.
    templates = [fn for fn in templates if fn.endswith(".hbs")]

    IGNORE_FILES = [
        # TODO: Add some exclude mechanism for the line-wrapping issue here.
        "static/templates/recipient_row.hbs",
    ]

    for fn in templates:
        if fn in IGNORE_FILES:
            continue
        validate(fn=fn, check_indent=True)

    for fn in templates:
        if fn in IGNORE_FILES:
            continue
        if not validate_indent_html(fn, fix):
            sys.exit(1)


if __name__ == "__main__":
    parser = argparse.ArgumentParser()
    parser.add_argument("-m", "--modified", action="store_true", help="only check modified files")
    parser.add_argument(
        "--all-dups",
        action="store_true",
        help="Run lint tool to detect duplicate ids on ignored files as well",
    )
    parser.add_argument(
        "--fix", action="store_true", help="Automatically fix indentation problems."
    )
    parser.add_argument("targets", nargs=argparse.REMAINDER)
    args = parser.parse_args()
    check_our_files(args.modified, args.all_dups, args.fix, args.targets)