quay/endpoints/api/search.py

""" Conduct searches against all registry context. """

import features

from endpoints.api import (ApiResource, parse_args, query_param, truthy_bool, nickname, resource,
                           require_scope, path_param, internal_only, Unauthorized, InvalidRequest,
                           show_if)
from data.database import Repository
from data import model
from data.registry_model import registry_model
from auth.permissions import (OrganizationMemberPermission, ReadRepositoryPermission,
                              UserAdminPermission, AdministerOrganizationPermission,
                              ReadRepositoryPermission)
from auth.auth_context import get_authenticated_user
from auth import scopes
from app import app, avatar, authentication
from flask import abort
from operator import itemgetter
from stringscore import liquidmetal
from util.names import parse_robot_username

import anunidecode # Don't listen to pylint's lies. This import is required.
import math


ENTITY_SEARCH_SCORE = 1
TEAM_SEARCH_SCORE = 2
REPOSITORY_SEARCH_SCORE = 4


@resource('/v1/entities/link/<username>')
@internal_only
class LinkExternalEntity(ApiResource):
  """ Resource for linking external entities to internal users. """
  @nickname('linkExternalUser')
  def post(self, username):
    if not authentication.federated_service:
      abort(404)

    # Only allowed if there is a logged in user.
    if not get_authenticated_user():
      raise Unauthorized()

    # Try to link the user with the given *external* username, to an internal record.
    (user, err_msg) = authentication.link_user(username)
    if user is None:
      raise InvalidRequest(err_msg, payload={'username': username})

    return {
      'entity': {
        'name': user.username,
        'kind': 'user',
        'is_robot': False,
        'avatar': avatar.get_data_for_user(user)
      }
    }


@resource('/v1/entities/<prefix>')
class EntitySearch(ApiResource):
  """ Resource for searching entities. """
  @path_param('prefix', 'The prefix of the entities being looked up')
  @parse_args()
  @query_param('namespace', 'Namespace to use when querying for org entities.', type=str,
               default='')
  @query_param('includeTeams', 'Whether to include team names.', type=truthy_bool, default=False)
  @query_param('includeOrgs', 'Whether to include orgs names.', type=truthy_bool, default=False)
  @nickname('getMatchingEntities')
  def get(self, prefix, parsed_args):
    """ Get a list of entities that match the specified prefix. """

    # Ensure we don't have any unicode characters in the search, as it breaks the search. Nothing
    # being searched can have unicode in it anyway, so this is a safe operation.
    prefix = prefix.encode('unidecode', 'ignore').replace(' ', '').lower()

    teams = []
    org_data = []

    namespace_name = parsed_args['namespace']
    robot_namespace = None
    organization = None

    try:
      organization = model.organization.get_organization(namespace_name)

      # namespace name was an org
      permission = OrganizationMemberPermission(namespace_name)
      if permission.can():
        robot_namespace = namespace_name

        if parsed_args['includeTeams']:
          teams = model.team.get_matching_teams(prefix, organization)

        if (parsed_args['includeOrgs'] and AdministerOrganizationPermission(namespace_name) and
            namespace_name.startswith(prefix)):
          org_data = [{
            'name': namespace_name,
            'kind': 'org',
            'is_org_member': True,
            'avatar': avatar.get_data_for_org(organization),
          }]

    except model.organization.InvalidOrganizationException:
      # namespace name was a user
      user = get_authenticated_user()
      if user and user.username == namespace_name:
        # Check if there is admin user permissions (login only)
        admin_permission = UserAdminPermission(user.username)
        if admin_permission.can():
          robot_namespace = namespace_name

    # Lookup users in the database for the prefix query.
    users = model.user.get_matching_users(prefix, robot_namespace, organization, limit=10,
                                          exact_matches_only=not features.PARTIAL_USER_AUTOCOMPLETE)

    # Lookup users via the user system for the prefix query. We'll filter out any users that
    # already exist in the database.
    external_users, federated_id, _ = authentication.query_users(prefix, limit=10)
    filtered_external_users = []
    if external_users and federated_id is not None:
      users = list(users)
      user_ids = [user.id for user in users]

      # Filter the users if any are already found via the database. We do so by looking up all
      # the found users in the federated user system.
      federated_query = model.user.get_federated_logins(user_ids, federated_id)
      found = {result.service_ident for result in federated_query}
      filtered_external_users = [user for user in external_users if not user.username in found]

    def entity_team_view(team):
      result = {
        'name': team.name,
        'kind': 'team',
        'is_org_member': True,
        'avatar': avatar.get_data_for_team(team)
      }
      return result

    def user_view(user):
      user_json = {
        'name': user.username,
        'kind': 'user',
        'is_robot': user.robot,
        'avatar': avatar.get_data_for_user(user)
      }

      if organization is not None:
        user_json['is_org_member'] = user.robot or user.is_org_member

      return user_json

    def external_view(user):
      result = {
        'name': user.username,
        'kind': 'external',
        'title': user.email or '',
        'avatar': avatar.get_data_for_external_user(user)
      }
      return result

    team_data = [entity_team_view(team) for team in teams]
    user_data = [user_view(user) for user in users]
    external_data = [external_view(user) for user in filtered_external_users]

    return {
      'results': team_data + user_data + org_data + external_data
    }


def search_entity_view(username, entity, get_short_name=None):
  kind = 'user'
  title = 'user'
  avatar_data = avatar.get_data_for_user(entity)
  href = '/user/' + entity.username

  if entity.organization:
    kind = 'organization'
    title = 'org'
    avatar_data = avatar.get_data_for_org(entity)
    href = '/organization/' + entity.username
  elif entity.robot:
    parts = parse_robot_username(entity.username)
    if parts[0] == username:
      href = '/user/' + username + '?tab=robots&showRobot=' + entity.username
    else:
      href = '/organization/' + parts[0] + '?tab=robots&showRobot=' + entity.username

    kind = 'robot'
    title = 'robot'
    avatar_data = None

  data = {
    'title': title,
    'kind': kind,
    'avatar': avatar_data,
    'name': entity.username,
    'score': ENTITY_SEARCH_SCORE,
    'href': href
  }

  if get_short_name:
    data['short_name'] = get_short_name(entity.username)

  return data


def conduct_team_search(username, query, encountered_teams, results):
  """ Finds the matching teams where the user is a member. """
  matching_teams = model.team.get_matching_user_teams(query, get_authenticated_user(), limit=5)
  for team in matching_teams:
    if team.id in encountered_teams:
      continue

    encountered_teams.add(team.id)

    results.append({
      'kind': 'team',
      'name': team.name,
      'organization': search_entity_view(username, team.organization),
      'avatar': avatar.get_data_for_team(team),
      'score': TEAM_SEARCH_SCORE,
      'href': '/organization/' + team.organization.username + '/teams/' + team.name
    })


def conduct_admined_team_search(username, query, encountered_teams, results):
  """ Finds matching teams in orgs admined by the user. """
  matching_teams = model.team.get_matching_admined_teams(query, get_authenticated_user(), limit=5)
  for team in matching_teams:
    if team.id in encountered_teams:
      continue

    encountered_teams.add(team.id)

    results.append({
      'kind': 'team',
      'name': team.name,
      'organization': search_entity_view(username, team.organization),
      'avatar': avatar.get_data_for_team(team),
      'score': TEAM_SEARCH_SCORE,
      'href': '/organization/' + team.organization.username + '/teams/' + team.name
    })


def conduct_repo_search(username, query, results, offset=0, limit=5):
  """ Finds matching repositories. """
  matching_repos = model.repository.get_filtered_matching_repositories(query, username, limit=limit,
                                                                       repo_kind=None,
                                                                       offset=offset)

  for repo in matching_repos:
    # TODO: make sure the repo.kind.name doesn't cause extra queries
    results.append(repo_result_view(repo, username))


def conduct_namespace_search(username, query, results):
  """ Finds matching users and organizations. """
  matching_entities = model.user.get_matching_user_namespaces(query, username, limit=5)
  for entity in matching_entities:
    results.append(search_entity_view(username, entity))


def conduct_robot_search(username, query, results):
  """ Finds matching robot accounts. """
  def get_short_name(name):
    return parse_robot_username(name)[1]

  matching_robots = model.user.get_matching_robots(query, username, limit=5)
  for robot in matching_robots:
    results.append(search_entity_view(username, robot, get_short_name))


def repo_result_view(repo, username, last_modified=None, stars=None, popularity=None):
  kind = 'application' if Repository.kind.get_name(repo.kind_id) == 'application' else 'repository'
  view = {
    'kind': kind,
    'title': 'app' if kind == 'application' else 'repo',
    'namespace': search_entity_view(username, repo.namespace_user),
    'name': repo.name,
    'description': repo.description,
    'is_public': model.repository.is_repository_public(repo),
    'score': REPOSITORY_SEARCH_SCORE,
    'href': '/' + kind + '/' + repo.namespace_user.username + '/' + repo.name
  }

  if last_modified is not None:
    view['last_modified'] = last_modified

  if stars is not None:
    view['stars'] = stars

  if popularity is not None:
    view['popularity'] = popularity

  return view

@resource('/v1/find/all')
class ConductSearch(ApiResource):
  """ Resource for finding users, repositories, teams, etc. """
  @parse_args()
  @query_param('query', 'The search query.', type=str, default='')
  @require_scope(scopes.READ_REPO)
  @nickname('conductSearch')
  def get(self, parsed_args):
    """ Get a list of entities and resources that match the specified query. """
    query = parsed_args['query']
    if not query:
      return {'results': []}

    username = None
    results = []

    if get_authenticated_user():
      username = get_authenticated_user().username

      # Search for teams.
      encountered_teams = set()
      conduct_team_search(username, query, encountered_teams, results)
      conduct_admined_team_search(username, query, encountered_teams, results)

      # Search for robot accounts.
      conduct_robot_search(username, query, results)

    # Search for repos.
    conduct_repo_search(username, query, results)

    # Search for users and orgs.
    conduct_namespace_search(username, query, results)

    # Modify the results' scores via how close the query term is to each result's name.
    for result in results:
      name = result.get('short_name', result['name'])
      lm_score = liquidmetal.score(name, query) or 0.5
      result['score'] = result['score'] * lm_score

    return {'results': sorted(results, key=itemgetter('score'), reverse=True)}


MAX_PER_PAGE = app.config.get('SEARCH_RESULTS_PER_PAGE', 10)
MAX_RESULT_PAGE_COUNT = app.config.get('SEARCH_MAX_RESULT_PAGE_COUNT', 10)

@resource('/v1/find/repositories')
class ConductRepositorySearch(ApiResource):
  """ Resource for finding repositories. """
  @parse_args()
  @query_param('query', 'The search query.', type=str, default='')
  @query_param('page', 'The page.', type=int, default=1)
  @nickname('conductRepoSearch')
  def get(self, parsed_args):
    """ Get a list of apps and repositories that match the specified query. """
    query = parsed_args['query']
    page = min(max(1, parsed_args['page']), MAX_RESULT_PAGE_COUNT)
    offset = (page - 1) * MAX_PER_PAGE
    limit = offset + MAX_PER_PAGE + 1

    username = get_authenticated_user().username if get_authenticated_user() else None

    # Lookup matching repositories.
    matching_repos = list(model.repository.get_filtered_matching_repositories(query, username,
                                                                              repo_kind=None,
                                                                              limit=limit,
                                                                              offset=offset))

    # Load secondary information such as last modified time, star count and action count.
    repository_ids = [repo.id for repo in matching_repos]
    last_modified_map = registry_model.get_most_recent_tag_lifetime_start(matching_repos)
    star_map = model.repository.get_stars(repository_ids)
    action_sum_map = model.log.get_repositories_action_sums(repository_ids)

    # Build the results list.
    results = [repo_result_view(repo, username, last_modified_map.get(repo.id),
                                star_map.get(repo.id, 0),
                                float(action_sum_map.get(repo.id, 0)))
               for repo in matching_repos]

    return {
      'results': results[0:MAX_PER_PAGE],
      'has_additional': len(results) > MAX_PER_PAGE,
      'page': page,
      'page_size': MAX_PER_PAGE,
      'start_index': offset,
    }