# -*- coding: utf-8 -*-

"""

kallithea.model.changeset_status

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

This file was forked by the Kallithea project in July 2014.

Original author and date, and relevant copyright and licensing information is below:

:created_on: Apr 30, 2012

:author: marcink

:copyright: (c) 2013 RhodeCode GmbH, and others.

:license: GPLv3, see LICENSE.md for more details.

"""

# This program is free software: you can redistribute it and/or modify

# it under the terms of the GNU 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 General Public License for more details.

#

# You should have received a copy of the GNU General Public License

# along with this program.  If not, see <http://www.gnu.org/licenses/>.

import logging

from collections import  defaultdict

from sqlalchemy.orm import joinedload

from kallithea.model import BaseModel

from kallithea.model.db import ChangesetStatus, PullRequest

from kallithea.lib.exceptions import StatusChangeOnClosedPullRequestError

log = logging.getLogger(__name__)

class ChangesetStatusModel(BaseModel):

    cls = ChangesetStatus

    def __get_changeset_status(self, changeset_status):

        return self._get_instance(ChangesetStatus, changeset_status)

    def __get_pull_request(self, pull_request):

        return self._get_instance(PullRequest, pull_request)

    def _get_status_query(self, repo, revision, pull_request,

                          with_revisions=False):

        repo = self._get_repo(repo)

        q = ChangesetStatus.query()\

:copyright: (c) 2013 RhodeCode GmbH, and others.

:license: GPLv3, see LICENSE.md for more details.

"""

import os

import time

import logging

import datetime

import traceback

import hashlib

import collections

import functools

from sqlalchemy import *

from sqlalchemy.ext.hybrid import hybrid_property

from sqlalchemy.orm import relationship, joinedload, class_mapper, validates

from beaker.cache import cache_region, region_invalidate

from webob.exc import HTTPNotFound

from pylons.i18n.translation import lazy_ugettext as _

from kallithea import DB_PREFIX

from kallithea.lib.vcs import get_backend

from kallithea.lib.vcs.utils.helpers import get_scm

from kallithea.lib.vcs.exceptions import VCSError

from kallithea.lib.vcs.utils.lazy import LazyProperty

from kallithea.lib.vcs.backends.base import EmptyChangeset

from kallithea.lib.utils2 import str2bool, safe_str, get_changeset_safe, \

    safe_unicode, remove_prefix, time_to_datetime, aslist, Optional, safe_int, \

    get_clone_url, urlreadable

from kallithea.lib.compat import json

from kallithea.lib.caching_query import FromCache

from kallithea.model.meta import Base, Session

URL_SEP = '/'

log = logging.getLogger(__name__)

#==============================================================================

# BASE CLASSES

#==============================================================================

_hash_key = lambda k: hashlib.md5(safe_str(k)).hexdigest()

class BaseModel(object):

    """

    """

    @classmethod

    def _get_keys(cls):

        """return column names for this model """

        return class_mapper(cls).c.keys()

    def get_dict(self):

        """

        return dict with keys and values corresponding

        to this model data """

        d = {}

        for k in self._get_keys():

            d[k] = getattr(self, k)

        # also use __json__() if present to get additional fields

        _json_attr = getattr(self, '__json__', None)

        if _json_attr:

            # update with attributes from __json__

            if callable(_json_attr):

                _json_attr = _json_attr()

            for k, val in _json_attr.iteritems():

                d[k] = val

        return d

    def get_appstruct(self):

        to this model data """

        l = []

        for k in self._get_keys():

            l.append((k, getattr(self, k),))

        return l

    def populate_obj(self, populate_dict):

        """populate model with data from given populate_dict"""

        for k in self._get_keys():

            if k in populate_dict:

                setattr(self, k, populate_dict[k])

    @classmethod

    def query(cls):

        return Session().query(cls)

    @classmethod

    def get(cls, id_):

        if id_:

            return cls.query().get(id_)

    @classmethod

    def get_or_404(cls, id_):

        try:

            id_ = int(id_)

        except (TypeError, ValueError):

            raise HTTPNotFound

        res = cls.query().get(id_)

        if not res:

            raise HTTPNotFound

        return res

    @classmethod

    def getAll(cls):

        # deprecated and left for backward compatibility

        return cls.get_all()

    @classmethod

    def get_all(cls):

        return cls.query().all()

    @classmethod

    def delete(cls, id_):

        obj = cls.query().get(id_)

        Session().delete(obj)

    @landing_rev.setter

    def landing_rev(self, val):

        if ':' not in val:

            raise ValueError('value must be delimited with `:` and consist '

                             'of <rev_type>:<rev>, got %s instead' % val)

        self._landing_revision = val

    @hybrid_property

    def locked(self):

        # always should return [user_id, timelocked]

        if self._locked:

            _lock_info = self._locked.split(':')

            return int(_lock_info[0]), _lock_info[1]

        return [None, None]

    @locked.setter

    def locked(self, val):

        if val and isinstance(val, (list, tuple)):

            self._locked = ':'.join(map(str, val))

        else:

            self._locked = None

    @hybrid_property

    def changeset_cache(self):

        from kallithea.lib.vcs.backends.base import EmptyChangeset

        dummy = EmptyChangeset().__json__()

        if not self._changeset_cache:

            return dummy

        try:

            return json.loads(self._changeset_cache)

        except TypeError:

            return dummy

    @changeset_cache.setter

    def changeset_cache(self, val):

        try:

            self._changeset_cache = json.dumps(val)

        except Exception:

            log.error(traceback.format_exc())

    @classmethod

    def url_sep(cls):

        return URL_SEP

    @classmethod

    def normalize_repo_name(cls, repo_name):

        """

        Normalizes os specific repo_name to the format internally stored inside

        :param cls:

        :param repo_name:

        """

        return cls.url_sep().join(repo_name.split(os.sep))

    @classmethod

    def get_by_repo_name(cls, repo_name):

        q = Session().query(cls).filter(cls.repo_name == repo_name)

        q = q.options(joinedload(Repository.fork))\

                .options(joinedload(Repository.user))\

                .options(joinedload(Repository.group))

        return q.scalar()

    @classmethod

    def get_by_full_path(cls, repo_full_path):

        repo_name = repo_full_path.split(cls.base_path(), 1)[-1]

        repo_name = cls.normalize_repo_name(repo_name)

        return cls.get_by_repo_name(repo_name.strip(URL_SEP))

    @classmethod

    def get_repo_forks(cls, repo_id):

        return cls.query().filter(Repository.fork_id == repo_id)

    @classmethod

    def base_path(cls):

        """

        Returns base path where all repos are stored

        :param cls:

        """

        q = Session().query(Ui)\

            .filter(Ui.ui_key == cls.url_sep())

        q = q.options(FromCache("sql_cache_short", "repository_repo_path"))

        return q.one().ui_value

    @property

    def forks(self):

        """

        Return forks of this repo

        """

        return Repository.get_repo_forks(self.repo_id)

    @property

    def parent(self):

        """

        Returns fork parent

        """

        if not groups:

            groups = cls.query().all()

        repo_groups = []

        if show_empty_group:

            repo_groups = [('-1', u'-- %s --' % _('top level'))]

        repo_groups.extend([cls._generate_choice(x) for x in groups])

        repo_groups = sorted(repo_groups, key=lambda t: t[1].split(cls.SEP)[0])

        return repo_groups

    @classmethod

    def url_sep(cls):

        return URL_SEP

    @classmethod

    def get_by_group_name(cls, group_name, cache=False, case_insensitive=False):

        if case_insensitive:

            gr = cls.query()\

                .filter(cls.group_name.ilike(group_name))

        else:

            gr = cls.query()\

                .filter(cls.group_name == group_name)

        if cache:

            gr = gr.options(FromCache(

                            "sql_cache_short",

                            "get_group_%s" % _hash_key(group_name)

                            )

            )

        return gr.scalar()

    @property

    def parents(self):

        parents_recursion_limit = 10

        groups = []

        if self.parent_group is None:

            return groups

        cur_gr = self.parent_group

        groups.insert(0, cur_gr)

        cnt = 0

        while 1:

            cnt += 1

            gr = getattr(cur_gr, 'parent_group', None)

            cur_gr = cur_gr.parent_group

            if gr is None:

                break

            if cnt == parents_recursion_limit:

                log.error(('more than %s parents found for group %s, stopping '

                           'recursive parent fetching' % (parents_recursion_limit, self)))

                break

            groups.insert(0, gr)

        return groups

    @property

    def children(self):

        return RepoGroup.query().filter(RepoGroup.parent_group == self)

    @property

    def name(self):

        return self.group_name.split(RepoGroup.url_sep())[-1]

    @property

    def full_path(self):

        return self.group_name

    @property

    def full_path_splitted(self):

        return self.group_name.split(RepoGroup.url_sep())

    @property

    def repositories(self):

        return Repository.query()\

                .filter(Repository.group == self)\

                .order_by(Repository.repo_name)

    @property

    def repositories_recursive_count(self):

        cnt = self.repositories.count()

        def children_count(group):

            cnt = 0

            for child in group.children:

                cnt += child.repositories.count()

                cnt += children_count(child)

            return cnt

        return cnt + children_count(self)

    def _recursive_objects(self, include_repos=True):

        all_ = []

        def _get_members(root_gr):

            if include_repos:

                for r in root_gr.repositories:

# -*- coding: utf-8 -*-

# This program is free software: you can redistribute it and/or modify

# it under the terms of the GNU 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 General Public License for more details.

#

# You should have received a copy of the GNU General Public License

# along with this program.  If not, see <http://www.gnu.org/licenses/>.

"""

http://formencode.org/module-formencode.validators.html

we can create our own validators

The table below outlines the options which can be used in a schema in addition to the validators themselves

pre_validators          []     These validators will be applied before the schema

chained_validators      []     These validators will be applied after the schema

allow_extra_fields      False     If True, then it is not an error when keys that aren't associated with a validator are present

filter_extra_fields     False     If True, then keys that aren't associated with a validator are removed

if_key_missing          NoDefault If this is given, then any keys that aren't available but are expected will be replaced with this value (and then validated). This does not override a present .if_missing attribute on validators. NoDefault is a special FormEncode class to mean that no default values has been specified and therefore missing keys shouldn't take a default value.

ignore_key_missing      False     If True, then missing keys will be missing in the result, if the validator doesn't have .if_missing on it already

<name> = formencode.validators.<name of validator>

<name> must equal form name

list=[1,2,3,4,5]

for SELECT use formencode.All(OneOf(list), Int())

"""

import logging

import formencode

from formencode import All

from pylons.i18n.translation import _

from kallithea import BACKENDS

from kallithea.model import validators as v

log = logging.getLogger(__name__)

class LoginForm(formencode.Schema):

    allow_extra_fields = True

    filter_extra_fields = True

    username = v.UnicodeString(

        strip=True,

        min=1,

        not_empty=True,

        messages={

           'empty': _(u'Please enter a login'),

           'tooShort': _(u'Enter a value %(min)i characters long or more')}

    )

    password = v.UnicodeString(

        strip=False,

        min=3,

        not_empty=True,

        messages={

        return self.sa.query(PullRequest)\

                .filter(PullRequest.other_repo == repo)\

                .filter(PullRequest.status != PullRequest.STATUS_CLOSED).count()

    def mark_as_fork(self, repo, fork, user):

        repo = self.__get_repo(repo)

        fork = self.__get_repo(fork)

        if fork and repo.repo_id == fork.repo_id:

            raise Exception("Cannot set repository as fork of itself")

        if fork and repo.repo_type != fork.repo_type:

            raise RepositoryError("Cannot set repository as fork of repository with other type")

        repo.fork = fork

        self.sa.add(repo)

        return repo

    def _handle_rc_scm_extras(self, username, repo_name, repo_alias,

                              action=None):

        from kallithea import CONFIG

        from kallithea.lib.base import _get_ip_addr

        try:

            from pylons import request

            environ = request.environ

        except TypeError:

            # we might use this outside of request context, let's fake the

            # environ data

            from webob import Request

            environ = Request.blank('').environ

        extras = {

            'ip': _get_ip_addr(environ),

            'username': username,

            'action': action or 'push_local',

            'repository': repo_name,

            'scm': repo_alias,

            'config': CONFIG['__file__'],

            'server_url': get_server_url(environ),

            'make_lock': None,

            'locked_by': [None, None]

        }

        _set_extras(extras)

    def _handle_push(self, repo, username, action, repo_name, revisions):

        """

        Triggers push action hooks

        :param repo: SCM repo

        :param username: username who pushes

        :param repo_name: name of repo

        :param revisions: list of revisions that we pushed

        """

        self._handle_rc_scm_extras(username, repo_name, repo_alias=repo.alias)

        _scm_repo = repo._repo

        # trigger push hook

        if repo.alias == 'hg':

            log_push_action(_scm_repo.ui, _scm_repo, node=revisions[0])

        elif repo.alias == 'git':

            log_push_action(None, _scm_repo, _git_revs=revisions)

    def _get_IMC_module(self, scm_type):

        """

        Returns InMemoryCommit class based on scm_type

        :param scm_type:

        """

        if scm_type == 'hg':

            from kallithea.lib.vcs.backends.hg import MercurialInMemoryChangeset

            return MercurialInMemoryChangeset

        if scm_type == 'git':

            from kallithea.lib.vcs.backends.git import GitInMemoryChangeset

            return GitInMemoryChangeset

        raise Exception('Invalid scm_type, must be one of hg,git got %s'

                        % (scm_type,))

    def pull_changes(self, repo, username):

        dbrepo = self.__get_repo(repo)

        clone_uri = dbrepo.clone_uri

        if not clone_uri:

            raise Exception("This repository doesn't have a clone uri")

        repo = dbrepo.scm_instance

        repo_name = dbrepo.repo_name

        try:

            if repo.alias == 'git':

                repo.fetch(clone_uri)

                # git doesn't really have something like post-fetch action

                # we fake that now. #TODO: extract fetched revisions somehow

                # here

                self._handle_push(repo,

                                  username=username,

                                  action='push_remote',

                                  repo_name=repo_name,

                                  revisions=[])

            else:

        self._handle_push(repo,

                          username=user.username,

                          action='push_local',

                          repo_name=repo_name,

                          revisions=[tip.raw_id])

        return tip

    def _sanitize_path(self, f_path):

        if f_path.startswith('/') or f_path.startswith('.') or '../' in f_path:

            raise NonRelativePathError('%s is not an relative path' % f_path)

        if f_path:

            f_path = os.path.normpath(f_path)

        return f_path

    def get_nodes(self, repo_name, revision, root_path='/', flat=True):

        """

        Recursively walk root dir and return a set of all paths found.

        :param repo_name: name of repository

        :param revision: revision for which to list nodes

        :param root_path: root path to list

        :param flat: return as a list, if False returns a dict with description

        """

        _files = list()

        _dirs = list()

        try:

            _repo = self.__get_repo(repo_name)

            changeset = _repo.scm_instance.get_changeset(revision)

            root_path = root_path.lstrip('/')

            for topnode, dirs, files in changeset.walk(root_path):

                for f in files:

                    _files.append(f.path if flat else {"name": f.path,

                                                       "type": "file"})

                for d in dirs:

                    _dirs.append(d.path if flat else {"name": d.path,

                                                      "type": "dir"})

        except RepositoryError:

            log.debug(traceback.format_exc())

            raise

        return _dirs, _files

    def create_nodes(self, user, repo, message, nodes, parent_cs=None,

                     author=None, trigger_push_hook=True):

        """

        Commits specified nodes to repo.

        :param repo: Kallithea Repository object

        :param message: commit message

        :param nodes: mapping {filename:{'content':content},...}

        :param parent_cs: parent changeset, can be empty than it's initial commit

        :param trigger_push_hook: trigger push hooks

        """

        user = self._get_user(user)

        scm_instance = repo.scm_instance_no_cache()

        processed_nodes = []

        for f_path in nodes:

            f_path = self._sanitize_path(f_path)

            content = nodes[f_path]['content']

            f_path = safe_str(f_path)

            # decoding here will force that we have proper encoded values

            # in any other case this will throw exceptions and deny commit

            if isinstance(content, (basestring,)):

                content = safe_str(content)

            elif isinstance(content, (file, cStringIO.OutputType,)):

                content = content.read()

            else:

                raise Exception('Content is of unrecognized type %s' % (

                    type(content)

                ))

            processed_nodes.append((f_path, content))

        message = safe_unicode(message)

        IMC = self._get_IMC_module(scm_instance.alias)

        imc = IMC(scm_instance)

        if not parent_cs:

            parent_cs = EmptyChangeset(alias=scm_instance.alias)

        if isinstance(parent_cs, EmptyChangeset):

            # EmptyChangeset means we we're editing empty repository

            parents = None

        else:

            parents = [parent_cs]

        # add multiple nodes

        for path, content in processed_nodes:

            imc.add(FileNode(path, content=content))

        tip = imc.commit(message=message,

                         author=author,

                         parents=parents,

                         branch=parent_cs.branch)

        self.mark_for_invalidation(repo.repo_name)

        if trigger_push_hook:

            self._handle_push(scm_instance,

                              username=user.username,

                              action='push_local',

                              repo_name=repo.repo_name,

                              revisions=[tip.raw_id])

        return tip

    def update_nodes(self, user, repo, message, nodes, parent_cs=None,

                     author=None, trigger_push_hook=True):

        user = self._get_user(user)

        scm_instance = repo.scm_instance_no_cache()

        message = safe_unicode(message)

        imc_class = self._get_IMC_module(scm_instance.alias)

        imc = imc_class(scm_instance)

        if not parent_cs:

            parent_cs = EmptyChangeset(alias=scm_instance.alias)

        if isinstance(parent_cs, EmptyChangeset):

            # EmptyChangeset means we we're editing empty repository

            parents = None

        else:

            parents = [parent_cs]

        # add multiple nodes

        for _filename, data in nodes.items():

            # new filename, can be renamed from the old one

            filename = self._sanitize_path(data['filename'])

            old_filename = self._sanitize_path(_filename)

            content = data['content']

            filenode = FileNode(old_filename, content=content)

            op = data['op']

            if op == 'add':

                imc.add(filenode)

            elif op == 'del':

                imc.remove(filenode)

            elif op == 'mod':

                if filename != old_filename:

                    #TODO: handle renames, needs vcs lib changes

                    imc.remove(filenode)

                    imc.add(FileNode(filename, content=content))

                else:

                    imc.change(filenode)

        # commit changes

        tip = imc.commit(message=message,

                         author=author,

                         parents=parents,

                         branch=parent_cs.branch)

        self.mark_for_invalidation(repo.repo_name)

        if trigger_push_hook:

            self._handle_push(scm_instance,

                              username=user.username,

                              action='push_local',

                              repo_name=repo.repo_name,

                              revisions=[tip.raw_id])

    def delete_nodes(self, user, repo, message, nodes, parent_cs=None,

                     author=None, trigger_push_hook=True):

        """

        Deletes specified nodes from repo.

        :param repo: Kallithea Repository object

        :param message: commit message

        :param nodes: mapping {filename:{'content':content},...}

        :param parent_cs: parent changeset, can be empty than it's initial commit

        :param trigger_push_hook: trigger push hooks

        """

        user = self._get_user(user)

        scm_instance = repo.scm_instance_no_cache()

        processed_nodes = []

        for f_path in nodes:

            f_path = self._sanitize_path(f_path)

            # content can be empty but for compatibility it allows same dicts

            # structure as add_nodes

            content = nodes[f_path].get('content')

            processed_nodes.append((f_path, content))

        message = safe_unicode(message)

        IMC = self._get_IMC_module(scm_instance.alias)

        imc = IMC(scm_instance)

        if not parent_cs:

            parent_cs = EmptyChangeset(alias=scm_instance.alias)

        if isinstance(parent_cs, EmptyChangeset):

            # EmptyChangeset means we we're editing empty repository

            parents = None

        else:

            parents = [parent_cs]

        # add multiple nodes

        for path, content in processed_nodes:

            imc.remove(FileNode(path, content=content))

        tip = imc.commit(message=message,

                         author=author,

                         parents=parents,

                         branch=parent_cs.branch)

        self.mark_for_invalidation(repo.repo_name)

        if trigger_push_hook:

            self._handle_push(scm_instance,

                              username=user.username,

                              action='push_local',

                              repo_name=repo.repo_name,

                              revisions=[tip.raw_id])

        return tip

    def get_unread_journal(self):

        return self.sa.query(UserLog).count()

    def get_repo_landing_revs(self, repo=None):

        """

        Generates select option with tags branches and bookmarks (for hg only)

        grouped by type

        :param repo:

        """

        hist_l = []

        choices = []

        repo = self.__get_repo(repo)

        hist_l.append(['rev:tip', _('latest tip')])

        choices.append('rev:tip')

        if not repo:

            return choices, hist_l

        user = User.get_by_email(user_email)

        if user:

            log.debug('password reset user found %s' % user)

            link = h.canonical_url('reset_password_confirmation', key=user.api_key)

            reg_type = EmailNotificationModel.TYPE_PASSWORD_RESET

            body = EmailNotificationModel().get_email_tmpl(reg_type,

                                                           'txt',

                                                           user=user.short_contact,

                                                           reset_url=link)

            html_body = EmailNotificationModel().get_email_tmpl(reg_type,

                                                           'html',

                                                           user=user.short_contact,

                                                           reset_url=link)

            log.debug('sending email')

            run_task(tasks.send_email, [user_email],

                     _("Password reset link"), body, html_body)

            log.info('send new password mail to %s' % user_email)

        else:

            log.debug("password reset email %s not found" % user_email)

        return True

    def reset_password(self, data):

        from kallithea.lib.celerylib import tasks, run_task

        from kallithea.lib import auth

        user_email = data['email']

        user = User.get_by_email(user_email)

        new_passwd = auth.PasswordGenerator().gen_password(8,

                        auth.PasswordGenerator.ALPHABETS_BIG_SMALL)

        if user:

            user.password = auth.get_crypt_password(new_passwd)

            Session().add(user)

            Session().commit()

            log.info('change password for %s' % user_email)

        if new_passwd is None:

            raise Exception('unable to generate new password')

        run_task(tasks.send_email, [user_email],

                 _('Your new password'),

                 _('Your new Kallithea password:%s') % (new_passwd,))

        log.info('send new password mail to %s' % user_email)

        return True

    def fill_data(self, auth_user, user_id=None, api_key=None, username=None):

        """

        Fetches auth_user by user_id,or api_key if present.

        Fills auth_user attributes with those taken from database.

        present in database