path: root/src/hydrilla/server/serve.py

# SPDX-License-Identifier: AGPL-3.0-or-later

# Main repository logic.
#
# This file is part of Hydrilla
#
# Copyright (C) 2021, 2022 Wojtek Kosior
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU Affero General Public License as
# published by the Free Software Foundation, either version 3 of the
# License, or (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU Affero General Public License for more details.
#
# You should have received a copy of the GNU Affero General Public License
# along with this program.  If not, see <https://www.gnu.org/licenses/>.
#
#
# I, Wojtek Kosior, thereby promise not to sue for violation of this
# file's license. Although I request that you do not make use this code
# in a proprietary program, I am not going to enforce this in court.

# Enable using with Python 3.7.
from __future__ import annotations

import re
import os
import pathlib
import json
import logging

from pathlib import Path
from hashlib import sha256
from abc import ABC, abstractmethod
from typing import Optional, Union, Iterable, TypeVar, Generic

import click
import flask

from werkzeug import Response

from .. import _version, versions, json_instances
from ..item_infos import ResourceInfo, MappingInfo, VersionedItemInfo
from ..translations import smart_gettext as _, translation as make_translation
#from ..url_patterns import PatternTree
from . import config

here = Path(__file__).resolve().parent

generated_by = {
    'name': 'hydrilla.server',
    'version': _version.version
}

    # def as_query_result(self) -> dict[str, Union[str, list[int]]]:
    #     """
    #     Produce a json.dump()-able object describing this mapping as one of a
    #     collection of query results.
    #     """
    #     return {
    #         'version':    self.version,
    #         'identifier': self.identifier,
    #         'long_name':  self.long_name
    #     }

class Malcontent:
    """
    Represent a directory with files that can be loaded and served by Hydrilla.
    """
    def __init__(self, malcontent_dir_path: Path):
        """
        When an instance of Malcontent is constructed, it searches
        malcontent_dir_path for serveable site-modifying packages and loads
        them into its data structures.
        """
        self.resource_infos: dict[str, VersionedItemInfo[ResourceInfo]] = {}
        self.mapping_infos:  dict[str, VersionedItemInfo[MappingInfo]]  = {}

        self.pattern_tree: PatternTree[MappingInfo] = PatternTree()

        self.malcontent_dir_path = malcontent_dir_path

        if not self.malcontent_dir_path.is_dir():
            raise ValueError(_('malcontent_dir_path_not_dir_{}')
                             .format(malcontent_dir_path))

        for item_type in ('mapping', 'resource'):
            type_path = self.malcontent_dir_path / item_type
            if not type_path.is_dir():
                continue

            for subpath in type_path.iterdir():
                if not subpath.is_dir():
                    continue

                for ver_file in subpath.iterdir():
                    try:
                        self._load_item(item_type, ver_file)
                    except Exception as e:
                        if flask.current_app._hydrilla_werror:
                            raise e from None

                        msg = _('couldnt_load_item_from_{}').format(ver_file)
                        logging.error(msg, exc_info=True)

        self._report_missing()
        self._finalize()

    @staticmethod
    def _register_info(infos: dict[str, VersionedItemInfo[VersionedType]],
                       identifier: str, item_info: VersionedType) -> None:
        """
        ...........
        """
        infos.setdefault(identifier, VersionedItemInfo())\
             .register(item_info)

    def _load_item(self, item_type: str, ver_file: Path) -> None:
        """
        Reads, validates and autocompletes serveable mapping/resource
        definition, then registers information from it in data structures.
        """
        version    = versions.parse_version(ver_file.name)
        identifier = ver_file.parent.name

        item_json, major = util.load_instance_from_file(ver_file)

        util.validator_for(f'api_{item_type}_description-{major}.schema.json')\
            .validate(item_json)

        # Assertion needed for mypy. If validation passed, this should not fail.
        assert major is not None

        item_info: ItemInfo = ResourceInfo(item_json, major) \
            if item_type == 'resource' else MappingInfo(item_json, major)

        if item_info.identifier != identifier:
            msg = _('item_{item}_in_file_{file}')\
                .format({'item': item_info.identifier, 'file': ver_file})
            raise ValueError(msg)

        if item_info.version != version:
            ver_str = util.version_string(item_info.version)
            msg = _('item_version_{ver}_in_file_{file}')\
                .format({'ver': ver_str, 'file': ver_file})
            raise ValueError(msg)

        if isinstance(item_info, ResourceInfo):
            self._register_info(self.resource_infos, identifier, item_info)
        elif isinstance(item_info, MappingInfo):
            self._register_info(self.mapping_infos, identifier, item_info)

    @staticmethod
    def _all_infos(infos: dict[str, VersionedItemInfo[VersionedType]]) \
        -> Iterable[VersionedType]:
        """
        ...........
        """
        for versioned_info in infos.values():
            for item_info in versioned_info.by_version.values():
                yield item_info

    def _report_missing(self) -> None:
        """
        Use logger to print information about items that are referenced but
        were not loaded.
        """
        def report_missing_dependency(info: ResourceInfo, dep: str) -> None:
            msg = _('no_dep_{resource}_{ver}_{dep}')\
                .format(dep=dep, resource=info.identifier,
                        ver=util.version_string(info.version))
            logging.error(msg)

        for resource_info in self._all_infos(self.resource_infos):
            for dep in resource_info.dependencies:
                if dep not in self.resource_infos:
                    report_missing_dependency(resource_info, dep)

        def report_missing_payload(info: MappingInfo, payload: str) -> None:
            msg = _('no_payload_{mapping}_{ver}_{payload}')\
                .format(mapping=info.identifier, payload=payload,
                        ver=util.version_string(info.version))
            logging.error(msg)

        for mapping_info in self._all_infos(self.mapping_infos):
            for payload in mapping_info.payloads.values():
                if payload not in self.resource_infos:
                    report_missing_payload(mapping_info, payload)

        def report_missing_mapping(info: ItemInfo,
                                   required_mapping: str) -> None:
            msg = _('no_mapping_{required_by}_{ver}_{required}')\
                .format(required_by=info.identifier, required=required_mapping,
                        ver=util.version_string(info.version))
            logging.error(msg)

        for item_info in (*self._all_infos(self.mapping_infos),
                          *self._all_infos(self.resource_infos)):
            for required in item_info.required_mappings:
                if required not in self.mapping_infos:
                    report_missing_mapping(item_info, required)

    def _finalize(self):
        """
        Initialize structures needed to serve queries. Called once after all
        data gets loaded.
        """
        for versioned_info in (*self.mapping_infos.values(),
                               *self.resource_infos.values()):
                versioned_info.known_versions.sort()

        for info in self._all_infos(self.mapping_infos):
            for pattern in info.payloads:
                try:
                    self.pattern_tree = \
                        self.pattern_tree.register(pattern, info)
                except Exception as e:
                    if flask.current_app._hydrilla_werror:
                        raise e from None
                    msg = _('couldnt_register_{mapping}_{ver}_{pattern}')\
                        .format(mapping=info.identifier, pattern=pattern,
                                ver=util.version_string(info.version))
                    logging.error(msg)

    def query(self, url: str) -> list[MappingInfo]:
        """
        Return a list of registered mappings that match url.

        If multiple versions of a mapping are applicable, only the most recent
        is included in the result.
        """
        collected: dict[str, MappingInfo] = {}
        for result_set in self.pattern_tree.search(url):
            for wrapped_mapping_info in result_set:
                info = wrapped_mapping_info.item
                previous = collected.get(info.identifier)
                if previous and previous.version > info.version:
                    continue

                collected[info.identifier] = info

        return list(collected.values())

bp = flask.Blueprint('bp', __package__)

class HydrillaApp(flask.Flask):
    """Flask app that implements a Hydrilla server."""
    def __init__(self, hydrilla_config: dict, flask_config: dict={}):
        """Create the Flask instance according to the configuration"""
        super().__init__(__package__, static_url_path='/',
                         static_folder=hydrilla_config['malcontent_dir'])
        self.config.update(flask_config)

        # https://stackoverflow.com/questions/9449101/how-to-stop-flask-from-initialising-twice-in-debug-mode
        if self.debug and os.environ.get('WERKZEUG_RUN_MAIN') != 'true':
            return

        self.jinja_options = {
            **self.jinja_options,
            'extensions': [
                *self.jinja_options.get('extensions', []),
                'jinja2.ext.i18n'
            ]
        }

        self._hydrilla_project_url = hydrilla_config['hydrilla_project_url']
        self._hydrilla_port = hydrilla_config['port']
        self._hydrilla_werror = hydrilla_config.get('werror', False)

        if 'hydrilla_parent' in hydrilla_config:
            raise ValueError("Option 'hydrilla_parent' is not implemented.")

        malcontent_dir = Path(hydrilla_config['malcontent_dir']).resolve()
        with self.app_context():
            self._hydrilla_malcontent = Malcontent(malcontent_dir)

        self.register_blueprint(bp)

    def create_jinja_environment(self, *args, **kwargs) \
        -> flask.templating.Environment:
        """
        Flask's create_jinja_environment(), but tweaked to always include the
        'hydrilla_project_url' global variable and to install proper
        translations.
        """
        env = super().create_jinja_environment(*args, **kwargs) # type: ignore
        env.install_gettext_translations(make_translation())
        env.globals['hydrilla_project_url'] = self._hydrilla_project_url

        return env

    def run(self, *args, **kwargs):
        """
        Flask's run(), but tweaked to use the port from hydrilla configuration
        by default.
        """
        return super().run(*args, port=self._hydrilla_port, **kwargs)

def malcontent():
    return flask.current_app._hydrilla_malcontent

@bp.route('/')
def index():
    return flask.render_template('index.html')

identifier_json_re = re.compile(r'^([-0-9a-z.]+)\.json$')

def get_resource_or_mapping(item_type: str, identifier: str) -> Response:
    """
    Strip '.json' from 'identifier', look the item up and send its JSON
    description.
    """
    match = identifier_json_re.match(identifier)
    if not match:
        flask.abort(404)

    identifier = match.group(1)

    if item_type == 'resource':
        infos = malcontent().resource_infos
    else:
        infos = malcontent().mapping_infos

    versioned_info = infos.get(identifier)

    info = versioned_info and versioned_info.get_by_ver()
    if info is None:
        flask.abort(404)

    # no need for send_from_directory(); path is safe, constructed by us
    file_path = malcontent().malcontent_dir_path / item_type / info.path()
    return flask.send_file(open(file_path, 'rb'), mimetype='application/json')

@bp.route('/mapping/<string:identifier_dot_json>')
def get_newest_mapping(identifier_dot_json: str) -> Response:
    return get_resource_or_mapping('mapping', identifier_dot_json)

@bp.route('/resource/<string:identifier_dot_json>')
def get_newest_resource(identifier_dot_json: str) -> Response:
    return get_resource_or_mapping('resource', identifier_dot_json)

@bp.route('/query')
def query():
    url = flask.request.args['url']

    mapping_refs = [i.as_query_result() for i in malcontent().query(url)]
    result = {
        '$schema': 'https://hydrilla.koszko.org/schemas/api_query_result-1.schema.json',
        'mappings': mapping_refs,
        'generated_by': generated_by
    }

    return Response(json.dumps(result), mimetype='application/json')

@bp.route('/--help')
def mm_help():
    return start.get_help(click.Context(start_wsgi)) + '\n'

@bp.route('/--version')
def mm_version():
    prog_info = {'prog': 'Hydrilla', 'version': _version.version}
    return _('%(prog)s_%(version)s_license') % prog_info + '\n'

default_config_path = Path('/etc/hydrilla/config.json')
default_malcontent_dir = '/var/lib/hydrilla/malcontent'
default_project_url = 'https://hydrillabugs.koszko.org/projects/hydrilla/wiki'

@click.command(help=_('serve_hydrilla_packages_explain_wsgi_considerations'))
@click.option('-m', '--malcontent-dir',
              type=click.Path(exists=True, file_okay=False),
              help=_('directory_to_serve_from_overrides_config'))
@click.option('-h', '--hydrilla-project-url', type=click.STRING,
              help=_('project_url_to_display_overrides_config'))
@click.option('-p', '--port', type=click.INT,
              help=_('tcp_port_to_listen_on_overrides_config'))
@click.option('-c', '--config', 'config_path',
              type=click.Path(exists=True, dir_okay=False, resolve_path=True),
              help=_('path_to_config_file_explain_default'))
@click.version_option(version=_version.version, prog_name='Hydrilla',
                      message=_('%(prog)s_%(version)s_license'),
                      help=_('version_printing'))
def start(malcontent_dir: Optional[str], hydrilla_project_url: Optional[str],
          port: Optional[int], config_path: Optional[str]) -> None:
    """
    Run a development Hydrilla server.

    This command is meant to be the entry point of hydrilla command exported by
    this package.
    """
    if config_path is None:
        hydrilla_config = config.load()
    else:
        hydrilla_config = config.load(config_paths=[Path(config_path)])

    if malcontent_dir is not None:
        hydrilla_config['malcontent_dir'] = str(Path(malcontent_dir).resolve())

    if hydrilla_project_url is not None:
        hydrilla_config['hydrilla_project_url'] = hydrilla_project_url

    if port is not None:
        hydrilla_config['port'] = port

    for opt in ('malcontent_dir', 'hydrilla_project_url', 'port'):
        if opt not in hydrilla_config:
            raise ValueError(_('config_option_{}_not_supplied').format(opt))

    HydrillaApp(hydrilla_config).run()

@click.command(help=_('serve_hydrilla_packages_wsgi_help'),
               context_settings={
                   'ignore_unknown_options': True,
                   'allow_extra_args': True
               })
@click.version_option(version=_version.version, prog_name='Hydrilla',
                      message=_('%(prog)s_%(version)s_license'),
                      help=_('version_printing'))
def start_wsgi() -> flask.Flask:
    """
    Create application object for use in WSGI deployment.

    This command Also handles --help and --version options in case it gets
    called outside WSGI environment.
    """
    return HydrillaApp(click.get_current_context().obj or config.load())