tests.py

"""Unit tests for testing the application functionality."""
import copy
import ntpath
import os
import uuid
from collections import OrderedDict
from threading import Thread
from unittest.mock import patch, MagicMock, mock_open
from zipfile import ZipFile

import flask
import rapidjson as json
import shutil
import sys
import time
import unittest
from multiprocessing import Process
from unittest import TestLoader
from datetime import datetime
from typing import Dict, List, Tuple, Type, Any

from conllu import TokenList
from flask import Flask
from gensim.models import Word2Vec
from graphannis.errors import NoSuchCorpus
from lxml import etree
from networkx import MultiDiGraph, Graph
from requests import HTTPError
from sqlalchemy.exc import OperationalError, InvalidRequestError
from sqlalchemy.orm import session
from werkzeug.wrappers import Response

import mcserver
from mcserver.app import create_app, db, start_updater, full_init, log_exception
from mcserver.app.api.exerciseAPI import map_exercise_data_to_database, get_graph_data
from mcserver.app.api.kwicAPI import handle_exercise_data
from mcserver.app.api.vocabularyAPI import check_vocabulary
from mcserver.app.models import ResourceType, FileType, ExerciseType, ExerciseData, \
    NodeMC, LinkMC, GraphData, Phenomenon, CustomCorpus, AnnisResponse, Solution, DownloadableFile, Language, \
    VocabularyCorpus, TextComplexityMeasure, CitationLevel, FrequencyItem, Choice, XapiStatement, ExerciseMC, \
    CorpusMC, make_solution_element_from_salt_id, Sentence, ReferenceableText, Dependency, PartOfSpeech
from mcserver.app.services import AnnotationService, CorpusService, FileService, CustomCorpusService, \
    DatabaseService, XMLservice, TextService, FrequencyService, ExerciseService, TextComplexityService
from mcserver.config import TestingConfig, Config
from mcserver.models_auto import Corpus, Exercise, UpdateInfo, LearningResult
from mocks import Mocks, MockResponse, MockW2V, MockQuery, TestHelper
from openapi.openapi_server.models import VocabularyForm, VocabularyMC, ExerciseForm, \
    KwicForm, VectorNetworkForm, MatchingExercise, ExerciseTypePath, H5PForm, TextComplexity


class McTestCase(unittest.TestCase):
    """The test suite for the main application."""

    def mocked_requests_get(*args, **kwargs):
        if TestingConfig.SIMULATE_HTTP_ERROR:
            raise HTTPError
        else:
            url: str = args[0]
            if url == Config.CTS_API_BASE_URL:
                if kwargs['params']['request'] == 'GetCapabilities':
                    return MockResponse(Mocks.cts_capabilities_xml)
                return MockResponse(Mocks.cts_reff_xml)

    def setUp(self):
        """Initializes the testing environment."""
        self.start_time = time.time()
        if os.path.exists(Config.GRAPH_DATABASE_DIR):
            shutil.rmtree(Config.GRAPH_DATABASE_DIR)
        self.class_name: str = str(self.__class__)
        TestHelper.update_flask_app(self.class_name, create_app)

    def tearDown(self):
        """Finishes testing by removing the traces."""
        print("{0}: {1} seconds".format(self.id(), "%.2f" % (time.time() - self.start_time)))

    @staticmethod
    def add_corpus(corpus: Corpus):
        """ Adds a corpus to the database. """
        db.session.add(corpus)
        db.session.query(UpdateInfo).delete()
        ui_cts: UpdateInfo = UpdateInfo.from_dict(resource_type=ResourceType.cts_data.name,
                                                  last_modified_time=datetime.utcnow().timestamp(), created_time=1)
        db.session.add(ui_cts)
        DatabaseService.commit()

    @staticmethod
    def clear_folder(folder_path: str):
        """ Deletes every file in a folder. """
        for f in [x for x in os.listdir(folder_path) if x != ".gitignore"]:
            os.remove(os.path.join(folder_path, f))

    def test_api_bad_request(self):
        """Returns validation errors as JSON."""
        response: Response = Mocks.app_dict[self.class_name].client.get(Config.SERVER_URI_CORPORA)
        self.assertEqual(response.status_code, 400)

    def test_api_corpus_delete(self):
        """ Deletes a single corpus. """
        db.session.query(Corpus).delete()
        response: Response = Mocks.app_dict[self.class_name].client.delete(
            f"{Config.SERVER_URI_CORPORA}/1")
        self.assertEqual(response.status_code, 404)
        McTestCase.add_corpus(Mocks.corpora[0])
        response = Mocks.app_dict[self.class_name].client.delete(f"{Config.SERVER_URI_CORPORA}/{Mocks.corpora[0].cid}")
        data_json: dict = json.loads(response.get_data())
        self.assertEqual(data_json, True)
        db.session.query(Corpus).delete()
        db.session.query(UpdateInfo).delete()
        # dirty hack so we can reuse it in other tests
        session.make_transient(Mocks.corpora[0])

    def test_api_corpus_get(self):
        """ Gets information about a single corpus. """
        response: Response = Mocks.app_dict[self.class_name].client.get(
            f"{Config.SERVER_URI_CORPORA}/{Mocks.corpora[0].cid}")
        self.assertEqual(response.status_code, 404)
        McTestCase.add_corpus(Mocks.corpora[0])
        response: Response = Mocks.app_dict[self.class_name].client.get(
            f"{Config.SERVER_URI_CORPORA}/{Mocks.corpora[0].cid}")
        data_json: dict = json.loads(response.get_data())
        old_dict: dict = Mocks.corpora[0].to_dict()
        self.assertEqual(data_json, old_dict)
        db.session.query(Corpus).delete()
        db.session.query(UpdateInfo).delete()
        # dirty hack so we can reuse it in other tests
        session.make_transient(Mocks.corpora[0])

    def test_api_corpus_list_get(self):
        """Adds multiple texts to the database and queries them all."""

        def expect_result(self: McTestCase, mock: MagicMock, lut: str, result: Any,
                          lmt: datetime = datetime.utcnow()):
            ui: UpdateInfo = UpdateInfo.from_dict(resource_type=ResourceType.cts_data.name,
                                                  last_modified_time=lmt.timestamp(), created_time=1)
            mock.session.query.return_value = MockQuery(ui)
            response: Response = Mocks.app_dict[self.class_name].client.get(
                TestingConfig.SERVER_URI_CORPORA, query_string=dict(last_update_time=lut))
            data_json = json.loads(response.get_data())
            if data_json:
                result = [x.to_dict() for x in result]
            self.assertEqual(data_json, result)

        with patch.object(mcserver.app.services.databaseService, "db") as mock_db:
            expect_result(self, mock_db, str(int(datetime.utcnow().timestamp() * 1000)), None,
                          datetime.fromtimestamp(0))
            db.session.add_all(Mocks.corpora)
            DatabaseService.commit()
            expect_result(self, mock_db, "0", Mocks.corpora, datetime.fromtimestamp(time.time()))
        db.session.query(Corpus).delete()
        db.session.query(UpdateInfo).delete()
        # dirty hack so we can reuse it in other tests
        session.make_transient(Mocks.corpora[0])

    def test_api_corpus_patch(self):
        """ Changes information about a single corpus. """
        response: Response = Mocks.app_dict[self.class_name].client.patch(
            f"{Config.SERVER_URI_CORPORA}/{Mocks.corpora[0].cid}")
        self.assertEqual(response.status_code, 404)
        McTestCase.add_corpus(Mocks.corpora[0])
        old_title: str = Mocks.corpora[0].title
        new_title: str = "new_title"
        response: Response = Mocks.app_dict[self.class_name].client.patch(
            f"{Config.SERVER_URI_CORPORA}/{Mocks.corpora[0].cid}", data=dict(title=new_title))
        data_json: dict = json.loads(response.get_data())
        old_dict: dict = Mocks.corpora[0].to_dict()
        self.assertEqual(data_json["title"], old_dict["title"])
        Mocks.corpora[0].title = old_title
        db.session.query(Corpus).delete()
        db.session.query(UpdateInfo).delete()
        # dirty hack so we can reuse it in other tests
        session.make_transient(Mocks.corpora[0])

    def test_api_exercise_get(self):
        """ Retrieves an existing exercise by its exercise ID. """
        db.session.query(Exercise).delete()
        response: Response = Mocks.app_dict[self.class_name].client.get(
            Config.SERVER_URI_EXERCISE, query_string=dict(eid=""))
        self.assertEqual(response.status_code, 404)
        old_urn: str = Mocks.exercise.urn
        Mocks.exercise.urn = ""
        db.session.add(Mocks.exercise)
        DatabaseService.commit()
        ar: AnnisResponse = AnnisResponse(solutions=[], graph_data=GraphData(links=[], nodes=[]))
        with patch.object(CorpusService, "get_corpus", side_effect=[ar, Mocks.annis_response]):
            response = Mocks.app_dict[self.class_name].client.get(Config.SERVER_URI_EXERCISE,
                                                                  query_string=dict(eid=Mocks.exercise.eid))
            self.assertEqual(response.status_code, 404)
            Mocks.exercise.urn = old_urn
            DatabaseService.commit()
            response = Mocks.app_dict[self.class_name].client.get(Config.SERVER_URI_EXERCISE,
                                                                  query_string=dict(eid=Mocks.exercise.eid))
            ar: AnnisResponse = AnnisResponse.from_dict(json.loads(response.get_data(as_text=True)))
            self.assertEqual(len(ar.graph_data.nodes), 52)
            db.session.query(Exercise).delete()
            session.make_transient(Mocks.exercise)

    def test_api_exercise_list_get(self):
        """ Retrieves a list of available exercises. """
        ui_exercises: UpdateInfo = UpdateInfo.from_dict(resource_type=ResourceType.exercise_list.name,
                                                        last_modified_time=1, created_time=1)
        db.session.add(ui_exercises)
        DatabaseService.commit()
        args: dict = dict(lang="fr", last_update_time=int(time.time()))
        response: Response = Mocks.app_dict[self.class_name].client.get(TestingConfig.SERVER_URI_EXERCISE_LIST,
                                                                        query_string=args)
        self.assertEqual(json.loads(response.get_data()), [])
        args["last_update_time"] = 0
        db.session.add(Mocks.exercise)
        DatabaseService.commit()
        response = Mocks.app_dict[self.class_name].client.get(TestingConfig.SERVER_URI_EXERCISE_LIST, query_string=args)
        exercises: List[MatchingExercise] = []
        for exercise_dict in json.loads(response.get_data(as_text=True)):
            exercise_dict["search_values"] = json.dumps(exercise_dict["search_values"])
            exercise_dict["solutions"] = json.dumps(exercise_dict["solutions"])
            exercises.append(MatchingExercise.from_dict(exercise_dict))
        self.assertEqual(len(exercises), 1)
        args = dict(lang=Language.English.value, vocabulary=VocabularyCorpus.agldt.name, frequency_upper_bound=500)
        response = Mocks.app_dict[self.class_name].client.get(TestingConfig.SERVER_URI_EXERCISE_LIST, query_string=args)
        exercises: List[dict] = json.loads(response.get_data())
        self.assertTrue(exercises[0]["matching_degree"])
        db.session.query(Exercise).delete()
        db.session.query(UpdateInfo).delete()
        session.make_transient(Mocks.exercise)

    def test_api_exercise_post(self):
        """ Creates a new exercise from scratch. """
        db.session.query(UpdateInfo).delete()
        ui_exercises: UpdateInfo = UpdateInfo.from_dict(resource_type=ResourceType.exercise_list.name,
                                                        last_modified_time=1, created_time=1)
        db.session.add(ui_exercises)
        DatabaseService.commit()
        ef: ExerciseForm = ExerciseForm(urn=Mocks.exercise.urn, type=ExerciseType.matching.value,
                                        search_values=Mocks.exercise.search_values, instructions='abc')
        with patch.object(mcserver.app.api.exerciseAPI, "get_graph_data") as mock_ggd:
            mock_ggd.side_effect = [Mocks.annis_response_dict, ValueError()]
            with patch.object(mcserver.app.services.textComplexityService.TextComplexityService,
                              "text_complexity", return_value=Mocks.text_complexity):
                response: Response = Mocks.app_dict[self.class_name].client.post(
                    Config.SERVER_URI_EXERCISE, headers=Mocks.headers_form_data, data=ef.to_dict())
                ar: AnnisResponse = AnnisResponse.from_dict(json.loads(response.get_data(as_text=True)))
                self.assertEqual(len(ar.solutions), 3)
                response = Mocks.app_dict[self.class_name].client.post(
                    Config.SERVER_URI_EXERCISE, headers=Mocks.headers_form_data, data=ef.to_dict())
                self.assertEqual(response.status_code, 500)
                Mocks.app_dict[self.class_name].app_context.push()
                db.session.query(UpdateInfo).delete()

    def test_api_file_get(self):
        """Gets an existing exercise"""
        ui_file: UpdateInfo = UpdateInfo.from_dict(resource_type=ResourceType.file_api_clean.name,
                                                   last_modified_time=1, created_time=1)
        db.session.add(ui_file)
        DatabaseService.commit()
        # create a fake old file, to be deleted on the next GET request
        FileService.create_tmp_file(FileType.XML, "old")
        args: dict = dict(type=FileType.XML, id=Mocks.exercise.eid, solution_indices=[0])
        response: Response = Mocks.app_dict[self.class_name].client.get(TestingConfig.SERVER_URI_FILE,
                                                                        query_string=args)
        self.assertEqual(response.status_code, 404)
        file_path: str = os.path.join(Config.TMP_DIRECTORY, Mocks.exercise.eid + "." + FileType.XML)
        file_content: str = "<xml></xml>"
        with open(file_path, "w+") as f:
            f.write(file_content)
        ui_file.last_modified_time = datetime.utcnow().timestamp()
        DatabaseService.commit()
        del ui_file
        response = Mocks.app_dict[self.class_name].client.get(TestingConfig.SERVER_URI_FILE, query_string=args)
        os.remove(file_path)
        self.assertEqual(response.get_data(as_text=True), file_content)
        # add the mapped exercise to the database
        db.session.add(Mocks.exercise)
        DatabaseService.commit()
        args["type"] = FileType.PDF
        response = Mocks.app_dict[self.class_name].client.get(TestingConfig.SERVER_URI_FILE, query_string=args)
        # the PDFs are not deterministically reproducible because the creation date etc. is written into them
        self.assertTrue(response.data.startswith(Mocks.exercise_pdf))
        db.session.query(Exercise).delete()
        session.make_transient(Mocks.exercise)

    def test_api_file_post(self):
        """ Posts exercise data to be saved temporarily or permanently on the server, e.g. for downloading. """
        learning_result: str = Mocks.xapi_json_string
        Mocks.app_dict[self.class_name].client.post(TestingConfig.SERVER_URI_FILE, headers=Mocks.headers_form_data,
                                                    data=dict(learning_result=learning_result))
        lrs: List[LearningResult] = DatabaseService.query(LearningResult)
        self.assertEqual(len(lrs), 1)
        data_dict: dict = dict(file_type=FileType.XML, urn=Mocks.urn_custom, html_content="<html></html>")
        response: Response = Mocks.app_dict[self.class_name].client.post(
            TestingConfig.SERVER_URI_FILE, headers=Mocks.headers_form_data, data=data_dict)
        file_name = json.loads(response.data.decode("utf-8"))
        self.assertTrue(file_name.endswith(".xml"))
        os.remove(os.path.join(Config.TMP_DIRECTORY, file_name))
        LearningResult.query.delete()

    def test_api_frequency_get(self):
        """ Requests a frequency analysis for a given URN. """
        expected_fa: List[FrequencyItem] = [
            FrequencyItem(values=[Dependency.object.name], phenomena=[Phenomenon.DEPENDENCY], count=1),
            FrequencyItem(values=[PartOfSpeech.adjective.name], phenomena=[Phenomenon.UPOSTAG], count=1)]
        with patch.object(
                mcserver.app.services.corpusService.CorpusService, "get_frequency_analysis",
                side_effect=[[FrequencyItem(values=[], phenomena=[], count=0)], expected_fa]):
            response: Response = Mocks.app_dict[self.class_name].client.get(
                TestingConfig.SERVER_URI_FREQUENCY, query_string=dict(urn=Mocks.urn_custom))
            result_list: List[dict] = json.loads(response.get_data(as_text=True))
            fa: List[FrequencyItem] = [FrequencyItem.from_dict(x) for x in result_list]
            self.assertEqual(len(fa), 1)
            response = Mocks.app_dict[self.class_name].client.get(
                TestingConfig.SERVER_URI_FREQUENCY, query_string=dict(urn=Mocks.urn_custom))
            result_list: List[dict] = json.loads(response.get_data(as_text=True))
            fa: List[FrequencyItem] = [FrequencyItem.from_dict(x) for x in result_list]
            self.assertEqual(fa[0].values, expected_fa[0].values)
            self.assertEqual(fa[1].values[0], None)

    def test_api_h5p_get(self):
        """ Requests a H5P JSON file for a given exercise. """
        response: Response = Mocks.app_dict[self.class_name].client.get(
            TestingConfig.SERVER_URI_H5P, query_string=dict(eid=Config.EXERCISE_ID_TEST, lang=Language.English.value))
        self.assertIn(Mocks.h5p_json_cloze[1:-1], response.get_data(as_text=True))
        args: dict = dict(eid=Mocks.exercise.eid, lang=Language.English.value, solution_indices=[0])
        response = Mocks.app_dict[self.class_name].client.get(TestingConfig.SERVER_URI_H5P, query_string=args)
        self.assertEqual(response.status_code, 404)
        db.session.add(Mocks.exercise)
        DatabaseService.commit()
        response = Mocks.app_dict[self.class_name].client.get(TestingConfig.SERVER_URI_H5P, query_string=args)
        self.assertIn(Mocks.h5p_json_cloze[1:-1], response.get_data(as_text=True))
        Mocks.exercise.exercise_type = ExerciseType.kwic.value
        DatabaseService.commit()
        response = Mocks.app_dict[self.class_name].client.get(TestingConfig.SERVER_URI_H5P, query_string=args)
        self.assertEqual(response.status_code, 422)
        Mocks.exercise.exercise_type = ExerciseType.matching.value
        DatabaseService.commit()
        response = Mocks.app_dict[self.class_name].client.get(TestingConfig.SERVER_URI_H5P, query_string=args)
        self.assertIn(Mocks.h5p_json_matching[1:-1], response.get_data(as_text=True))
        Mocks.exercise.exercise_type = ExerciseType.cloze.value
        args["lang"] = "fr"
        response = Mocks.app_dict[self.class_name].client.get(TestingConfig.SERVER_URI_H5P, query_string=args)
        self.assertIn(Mocks.h5p_json_cloze[1:-1], response.get_data(as_text=True))
        db.session.query(Exercise).delete()
        session.make_transient(Mocks.exercise)

    def test_api_h5p_post(self):
        """The POST method for the H5P REST API. It offers client-side H5P exercises for download as ZIP archives."""
        exercise: Exercise = copy.deepcopy(Mocks.exercise)
        hf: H5PForm = H5PForm(eid=Mocks.exercise.eid, exercise_type_path=ExerciseTypePath.DRAG_TEXT,
                              lang=Language.English.value, solution_indices=[])
        response: Response = Mocks.app_dict[self.class_name].client.post(
            TestingConfig.SERVER_URI_H5P, headers=Mocks.headers_form_data, data=hf.to_dict())
        self.assertEqual(response.status_code, 404)
        db.session.add(exercise)
        DatabaseService.commit()
        response = Mocks.app_dict[self.class_name].client.post(
            TestingConfig.SERVER_URI_H5P, headers=Mocks.headers_form_data, data=hf.to_dict())
        self.assertEqual(len(response.get_data()), 1940145)
        with patch.object(mcserver.app.api.h5pAPI, "get_text_field_content", return_value=""):
            response = Mocks.app_dict[self.class_name].client.post(
                TestingConfig.SERVER_URI_H5P, headers=Mocks.headers_form_data, data=hf.to_dict())
            self.assertEqual(response.status_code, 422)
        db.session.query(Exercise).delete()

    def test_api_kwic_post(self):
        """ Posts an AQL query to create a KWIC visualization in SVG format. """
        ed1: ExerciseData = ExerciseService.map_graph_data_to_exercise(
            Mocks.annis_response_dict["graph_data_raw"],
            "", [Solution(target=make_solution_element_from_salt_id(
                'salt:/urn:custom:latinLit:proiel.pal-agr.lat:1.1.1/doc1#sent159692tok1'))])
        ed2: ExerciseData = ExerciseService.map_graph_data_to_exercise(
            Mocks.annis_response_dict["graph_data_raw"],
            "", [Solution(target=make_solution_element_from_salt_id(
                'salt:/urn:custom:latinLit:proiel.pal-agr.lat:1.1.1/doc1#sent159695tok10'))])
        ed2.graph.nodes = ed2.graph.nodes[42:]
        mr: MockResponse = MockResponse(json.dumps([ed1.serialize(), ed2.serialize()]))
        kf: KwicForm = KwicForm(ctx_left=5, ctx_right=5, search_values=Mocks.exercise.search_values,
                                urn=Mocks.urn_custom)
        with patch.object(mcserver.app.services.corpusService.requests, "post", return_value=mr):
            response: Response = Mocks.app_dict[self.class_name].client.post(
                TestingConfig.SERVER_URI_KWIC, headers=Mocks.headers_form_data, data=kf.to_dict())
            self.assertTrue(response.data.startswith(Mocks.kwic_svg))

    def test_api_not_found(self):
        """Checks the 404 response in case of an invalid API query URL."""
        response: Response = Mocks.app_dict[self.class_name].client.get("/")
        self.assertEqual(response.status_code, 404)

    def test_api_raw_text_get(self):
        """ Retrieves the raw text for a given URN. """
        with patch.object(mcserver.app.services.corpusService.CorpusService, "get_corpus") as mock_get_corpus:
            with patch.object(mcserver.app.services.textComplexityService.TextComplexityService,
                              "text_complexity", return_value=Mocks.text_complexity):
                mock_get_corpus.return_value = AnnisResponse(
                    graph_data=GraphData(links=[], nodes=[]), solutions=[])
                response: Response = Mocks.app_dict[self.class_name].client.get(
                    TestingConfig.SERVER_URI_RAW_TEXT, query_string=dict(urn=Mocks.urn_custom))
                self.assertEqual(response.status_code, 404)
                mock_get_corpus.return_value = Mocks.annis_response
                response = Mocks.app_dict[self.class_name].client.get(TestingConfig.SERVER_URI_RAW_TEXT,
                                                                      query_string=dict(urn=Mocks.urn_custom))
                ar: AnnisResponse = AnnisResponse.from_dict(json.loads(response.get_data(as_text=True)))
                self.assertEqual(len(ar.graph_data.nodes), 52)
                ar_copy: AnnisResponse = AnnisResponse.from_dict(Mocks.annis_response.to_dict())
                ar_copy.graph_data.nodes = []
                mock_get_corpus.return_value = ar_copy
                response = Mocks.app_dict[self.class_name].client.get(TestingConfig.SERVER_URI_RAW_TEXT,
                                                                      query_string=dict(urn=Mocks.urn_custom))
                self.assertEqual(response.status_code, 404)

    def test_api_static_exercises_get(self):
        """ Retrieves static exercises from the frontend and publishes deep URLs for each one of them. """
        exercises: List[Tuple[str, str, str]] = [
            (ExerciseTypePath.FILL_BLANKS,) + Mocks.h5p_json_fill_blanks_1,
            (ExerciseTypePath.FILL_BLANKS,) + Mocks.h5p_json_fill_blanks_3,
            (ExerciseTypePath.FILL_BLANKS,) + Mocks.h5p_json_fill_blanks_4,
            (ExerciseTypePath.FILL_BLANKS,) + Mocks.h5p_json_fill_blanks_13,
            (ExerciseTypePath.DRAG_TEXT, "1_en", Mocks.h5p_json_cloze),
            (ExerciseTypePath.MULTI_CHOICE, "1_en", Mocks.h5p_json_multi_choice),
            (ExerciseTypePath.MULTI_CHOICE, "2_en", Mocks.h5p_json_multi_choice_2),
            (ExerciseTypePath.MULTI_CHOICE,) + Mocks.h5p_json_multi_choice_9,
            (ExerciseTypePath.VOC_LIST, "1_en", Mocks.h5p_json_voc_list)]
        paths: List[str] = []
        for exercise in exercises:
            file_name: str = exercise[1] + ".json"
            file_path: str = os.path.join(Config.TMP_DIRECTORY, exercise[0], "content", file_name)
            os.makedirs(os.path.split(file_path)[0], exist_ok=True)
            json.dump(json.loads(exercise[2]), open(file_path, "w+"))
            paths.append(file_path)
        with ZipFile(TestingConfig.STATIC_EXERCISES_ZIP_FILE_PATH, "w") as z:
            for path in paths:
                z.write(path)
        for exercise in exercises:
            shutil.rmtree(os.path.join(Config.TMP_DIRECTORY, exercise[0]), ignore_errors=True)
        zip_content: bytes = open(TestingConfig.STATIC_EXERCISES_ZIP_FILE_PATH, "rb").read()
        with patch.object(
                mcserver.app.api.staticExercisesAPI.requests, "get", side_effect=[
                    MockResponse("{}", ok=False), MockResponse("{}", content=zip_content)]):
            with patch.object(AnnotationService, "get_udpipe",
                              return_value=Mocks.static_exercises_udpipe_string) as mock_udpipe:
                response: Response = Mocks.app_dict[self.class_name].client.get(
                    TestingConfig.SERVER_URI_STATIC_EXERCISES)
                self.assertEqual(response.status_code, 503)
                response = Mocks.app_dict[self.class_name].client.get(TestingConfig.SERVER_URI_STATIC_EXERCISES)
                os.remove(TestingConfig.STATIC_EXERCISES_ZIP_FILE_PATH)
                self.assertGreater(len(response.get_data(as_text=True)), 1900)
                response = Mocks.app_dict[self.class_name].client.get(TestingConfig.SERVER_URI_STATIC_EXERCISES)
                self.assertEqual(mock_udpipe.call_count, 1)

    def test_api_text_complexity_get(self):
        """ Calculates text complexity measures for a given URN. """
        tc: TextComplexity = TextComplexityService.text_complexity("n_w", Mocks.urn_custom, Mocks.graph_data)
        self.assertEqual(tc.n_w, 52)
        with patch.object(mcserver.app.services.corpusService.CorpusService, "get_corpus",
                          return_value=Mocks.annis_response):
            with patch.object(mcserver.app.services.textComplexityService.TextComplexityService,
                              "text_complexity", return_value=Mocks.text_complexity):
                args: dict = dict(urn=Mocks.urn_custom, measure=TextComplexityMeasure.all.name)
                response: Response = Mocks.app_dict[self.class_name].client.get(
                    TestingConfig.SERVER_URI_TEXT_COMPLEXITY, query_string=args)
                self.assertEqual(response.get_data(as_text=True), json.dumps(Mocks.text_complexity.to_dict()))

    @patch('MyCapytain.retrievers.cts5.requests.get', side_effect=mocked_requests_get)
    def test_api_valid_reff_get(self, mock_get: MagicMock):  #
        """ Retrieves possible citations for a given URN. """
        args: dict = dict(urn=Mocks.urn_custom[:-14])
        response: Response = Mocks.app_dict[self.class_name].client.get(TestingConfig.SERVER_URI_VALID_REFF,
                                                                        query_string=args)
        self.assertEqual(len(json.loads(response.data.decode("utf-8"))), 3)
        McTestCase.clear_folder(Config.REFF_CACHE_DIRECTORY)
        args["urn"] = f"{Mocks.urn_custom[:-13]}4"
        response = Mocks.app_dict[self.class_name].client.get(TestingConfig.SERVER_URI_VALID_REFF, query_string=args)
        self.assertEqual(response.status_code, 404)
        McTestCase.clear_folder(Config.REFF_CACHE_DIRECTORY)
        args["urn"] = f"{Mocks.urn_custom[:-13]}abc"
        response = Mocks.app_dict[self.class_name].client.get(TestingConfig.SERVER_URI_VALID_REFF, query_string=args)
        self.assertEqual(response.status_code, 400)
        McTestCase.clear_folder(Config.REFF_CACHE_DIRECTORY)
        TestingConfig.SIMULATE_HTTP_ERROR = True
        self.assertEqual(len(CorpusService.get_standard_corpus_reff(Mocks.urn[:-8])), 0)
        TestingConfig.SIMULATE_HTTP_ERROR = False
        reff: List[str] = CorpusService.get_standard_corpus_reff(Mocks.urn[:-8])
        self.assertEqual(len(reff), 7)

    def test_api_vector_network_get(self):
        """ Builds a network of semantically similar vectors for a given list of words. """
        with patch.object(mcserver.app.api.vectorNetworkAPI, "add_edges", side_effect=Mocks.mock_add_eges):
            with patch.object(mcserver.app.api.vectorNetworkAPI.Word2Vec, "load", return_value=MockW2V()):
                args: dict = dict(search_regex='ueritas', nearest_neighbor_count=150, min_count=6)
                response: Response = Mocks.app_dict[self.class_name].client.get(
                    TestingConfig.SERVER_URI_VECTOR_NETWORK, query_string=args)
                svg_string: str = json.loads(response.get_data(as_text=True))
                self.assertGreater(len(svg_string), 6500)

    def test_api_vector_network_post(self):
        """ Returns contexts that are semantically similar to a given query. """
        mock_data: str = "This is a sentence.\nAnd here is yet another one.\n"
        with patch("mcserver.app.api.vectorNetworkAPI.open", mock_open(read_data=mock_data)):
            with patch.object(mcserver.app.api.vectorNetworkAPI.Word2Vec, "load", return_value=MockW2V()):
                vnf: VectorNetworkForm = VectorNetworkForm(search_regex='uera', nearest_neighbor_count=10)
                response: Response = Mocks.app_dict[self.class_name].client.post(
                    TestingConfig.SERVER_URI_VECTOR_NETWORK, headers=Mocks.headers_form_data, data=vnf.to_dict())
                self.assertEqual(len(json.loads(response.get_data(as_text=True))), 2)

    def test_api_vocabulary_get(self):
        """ Retrieves sentence ID and matching degree for each sentence in the query text. """
        with patch.object(mcserver.app.services.corpusService.requests, "get",
                          return_value=MockResponse(json.dumps(Mocks.annis_response.to_dict()))):
            args: dict = dict(query_urn=Mocks.urn_custom, show_oov=True, vocabulary=VocabularyCorpus.agldt.name,
                              frequency_upper_bound=6000)
            response = Mocks.app_dict[self.class_name].client.get(TestingConfig.SERVER_URI_VOCABULARY,
                                                                  query_string=args)
            sentences: List[Sentence] = [Sentence.from_dict(x) for x in json.loads(response.get_data(as_text=True))]
            self.assertEqual(sentences[0].matching_degree, 92.85714285714286)

    def test_api_vocabulary_post(self):
        """ Indicates for each token of a corpus whether it is covered by a reference vocabulary. """
        with patch.object(mcserver.app.services.corpusService.CorpusService, "get_corpus",
                          return_value=Mocks.annis_response):
            vf: VocabularyForm = VocabularyForm(frequency_upper_bound=500, query_urn=Mocks.urn_custom,
                                                vocabulary=VocabularyMC.AGLDT)
            response: Response = Mocks.app_dict[self.class_name].client.post(
                TestingConfig.SERVER_URI_VOCABULARY, data=vf.to_dict())
            ar: AnnisResponse = AnnisResponse.from_dict(json.loads(response.get_data(as_text=True)))
            self.assertTrue(NodeMC.from_dict(ar.graph_data.nodes[3].to_dict()).is_oov)

    def test_app_init(self):
        """Creates an MCserver app in testing mode."""
        CorpusService.init_graphannis_logging()
        self.assertTrue(os.path.exists(Config.GRAPHANNIS_LOG_PATH))
        os.remove(Config.GRAPHANNIS_LOG_PATH)
        with patch.object(sys, 'argv', Mocks.test_args):
            app: Flask = mcserver.get_app()
            self.assertIsInstance(app, Flask)
            self.assertTrue(app.config["TESTING"])
            db.session.query(UpdateInfo).delete()
        Mocks.app_dict[self.class_name].app_context.push()
        db.session.query(Corpus).delete()

    def test_commit(self):
        """Commits the last action to the database and, if it fails, rolls back the current session."""

        def commit():
            raise OperationalError("", [], "")

        with patch.object(mcserver.app.services.databaseService, "db") as mock_db:
            mock_db.session.commit.side_effect = commit
            with self.assertRaises(OperationalError):
                McTestCase.add_corpus(Mocks.corpora[0])
        db.session.query(Corpus).delete()
        db.session.query(UpdateInfo).delete()
        session.make_transient(Mocks.corpora[0])

    def test_create_app(self):
        """Creates a new Flask application and configures it. Initializes the application and the database."""
        with patch.object(sys, "argv", [None, None, Config.FLASK_MIGRATE]):
            with patch.object(mcserver.app, "init_app_common", return_value=Flask(__name__)):
                cfg: Type[Config] = TestingConfig
                old_uri: str = cfg.SQLALCHEMY_DATABASE_URI
                create_app(cfg)
                self.assertEqual(cfg.SQLALCHEMY_DATABASE_URI, Config.DATABASE_URL_LOCAL)
                cfg.SQLALCHEMY_DATABASE_URI = old_uri
        Mocks.app_dict[self.class_name].app_context.push()

    def test_get_favicon(self):
        """Sends the favicon to browsers, which is used, e.g., in the tabs as a symbol for our application."""
        response: Response = Mocks.app_dict[self.class_name].client.get(Config.SERVER_URI_FAVICON)
        with open(os.path.join(Config.ASSETS_DIRECTORY, Config.FAVICON_FILE_NAME), "rb") as f:
            content: bytes = f.read()
            data_received: bytes = response.get_data()
            self.assertEqual(content, data_received)

    def test_map_exercise_data_to_database(self):
        """Maps exercise data to the database and saves it for later access."""
        ui_exercises: UpdateInfo = UpdateInfo.from_dict(resource_type=ResourceType.exercise_list.name,
                                                        last_modified_time=1, created_time=1)
        db.session.add(ui_exercises)
        DatabaseService.commit()
        exercise_expected: Exercise = Mocks.exercise
        exercise: Exercise = map_exercise_data_to_database(
            solutions=[Solution.from_dict(x) for x in json.loads(exercise_expected.solutions)],
            exercise_data=ExerciseData(json_dict=Mocks.exercise_data.serialize()),
            instructions=exercise_expected.instructions, exercise_type=exercise_expected.exercise_type,
            exercise_type_translation=exercise_expected.exercise_type_translation, xml_guid=exercise_expected.eid,
            conll=exercise_expected.conll, correct_feedback=exercise_expected.correct_feedback,
            partially_correct_feedback=exercise_expected.partially_correct_feedback, urn=Mocks.urn_custom,
            incorrect_feedback=exercise_expected.incorrect_feedback, search_values=exercise_expected.search_values,
            general_feedback=exercise_expected.general_feedback, work_author=exercise_expected.work_author,
            work_title=exercise_expected.work_title, language=exercise_expected.language)
        expected_values: List[str] = [
            exercise_expected.conll, exercise_expected.general_feedback, exercise_expected.incorrect_feedback,
            exercise_expected.search_values, exercise_expected.partially_correct_feedback,
            exercise_expected.correct_feedback, exercise_expected.instructions,
            exercise_expected.exercise_type_translation, exercise_expected.exercise_type, exercise_expected.solutions,
            exercise_expected.eid]
        actual_values: List[str] = [
            exercise.conll, exercise.general_feedback, exercise.incorrect_feedback, exercise.search_values,
            exercise.partially_correct_feedback, exercise.correct_feedback, exercise.instructions,
            exercise.exercise_type_translation, exercise.exercise_type, exercise.solutions, exercise.eid]
        self.assertEqual(expected_values, actual_values)
        exercise_from_db: Exercise = DatabaseService.query(Exercise, first=True)
        self.assertEqual(exercise, exercise_from_db)
        db.session.query(Exercise).delete()
        db.session.query(UpdateInfo).delete()
        session.make_transient(Mocks.exercise)

    @patch('MyCapytain.retrievers.cts5.requests.get', side_effect=mocked_requests_get)
    def test_update_corpora(self, mock_get: MagicMock):
        """Checks the remote repositories for new corpora to be included in our database."""
        CorpusService.update_corpora()
        self.assertEqual(len(CorpusService.existing_corpora), 1)
        ec: Corpus = CorpusService.existing_corpora[0]
        ec.title = ""
        DatabaseService.commit()
        McTestCase.add_corpus(CorpusMC.from_dict(source_urn="123"))
        cls: List[CitationLevel] = [ec.citation_level_1, ec.citation_level_2, ec.citation_level_3]
        CorpusService.update_corpus(ec.title, ec.source_urn, ec.author, cls, ec)
        self.assertFalse(ec.title)
        CorpusService.update_corpora()
        self.assertTrue(ec.title)
        db.session.query(Corpus).delete()
        db.session.query(UpdateInfo).delete()


class CorpusTestCase(unittest.TestCase):
    def setUp(self):
        """Initializes the testing environment."""
        self.start_time = time.time()
        self.class_name: str = str(self.__class__)
        TestHelper.update_flask_app(self.class_name, create_app)

    def tearDown(self):
        """Finishes testing by removing the traces."""
        print("{0}: {1} seconds".format(self.id(), "%.2f" % (time.time() - self.start_time)))

    @patch('mcserver.app.services.corpusService.CorpusService.update_corpora')
    def test_check_corpus_list_age(self, mock_update: MagicMock):
        """Checks whether the list of available corpora needs to be updated."""
        db.session.query(UpdateInfo).delete()
        ui_cts: UpdateInfo = UpdateInfo.from_dict(resource_type=ResourceType.cts_data.name,
                                                  last_modified_time=1, created_time=1)
        db.session.add(ui_cts)
        DatabaseService.commit()
        utc_now: datetime = datetime.utcnow()
        CorpusService.check_corpus_list_age(Mocks.app_dict[self.class_name].app)
        ui_cts: UpdateInfo = DatabaseService.query(
            UpdateInfo, filter_by=dict(resource_type=ResourceType.cts_data.name), first=True)
        self.assertGreater(ui_cts.last_modified_time, utc_now.timestamp())
        db.session.query(UpdateInfo).delete()

    def test_extract_custom_corpus_text(self):
        """ Extracts text from the relevant parts of a (custom) corpus. """
        new_text_parts: List[ReferenceableText] = CustomCorpusService.extract_custom_corpus_text(
            Mocks.text_parts, ["", ""], ["", "0"], "", 1, [False, True])
        self.assertEqual(len(new_text_parts), 0)
        new_text_parts = CustomCorpusService.extract_custom_corpus_text(Mocks.text_parts, ["", ""], ["", "0"], "", 1)
        self.assertEqual(new_text_parts[0].text, Mocks.text_parts[0].text_value)
        new_text_parts = CustomCorpusService.extract_custom_corpus_text(Mocks.text_parts, ["1"], ["3"], "")
        self.assertEqual(new_text_parts[0].text, Mocks.text_parts[0].text_value)

    def test_get_corpus(self):
        """ Loads the text for a standard corpus from the CTS API or cache. """
        ar: AnnisResponse = CorpusService.get_corpus("")
        self.assertEqual(len(ar.graph_data.nodes), 0)

    def test_get_custom_corpus_annotations(self):
        """ Retrieves the annotated text for a custom non-PROIEL corpus, e.g. a textbook. """
        mock_conll: List[TokenList] = Mocks.annotations + [TokenList([], metadata=OrderedDict([("sent_id", "3")]))]
        with patch.object(CustomCorpusService, "get_custom_corpus_text", return_value=Mocks.text_list):
            with patch.object(AnnotationService, "get_udpipe", return_value=Mocks.udpipe_string):
                with patch.object(AnnotationService, "parse_conll_string", return_value=mock_conll):
                    conll: List[TokenList] = CustomCorpusService.get_custom_corpus_annotations(Mocks.urn + "@1-2")
                    self.assertEqual(len(conll), 1)

    def test_get_custom_corpus_reff(self):
        """ Retrieves possible citations for given URN. """
        CustomCorpusService.custom_corpora[4].text_parts = Mocks.text_parts
        reff: List[str] = CustomCorpusService.get_custom_corpus_reff(Mocks.urn_custom[:-15])
        self.assertEqual(len(reff), 0)
        McTestCase.clear_folder(Config.REFF_CACHE_DIRECTORY)
        reff = CustomCorpusService.get_custom_corpus_reff(Mocks.urn_custom[:-14])
        self.assertEqual(len(reff), 1)
        McTestCase.clear_folder(Config.REFF_CACHE_DIRECTORY)
        reff = CustomCorpusService.get_custom_corpus_reff(Mocks.urn_custom[:-9])
        self.assertEqual(len(reff), 2)
        reff = CustomCorpusService.get_custom_corpus_reff(Mocks.urn_custom[:-9])
        self.assertEqual(len(reff), 2)
        McTestCase.clear_folder(Config.REFF_CACHE_DIRECTORY)
        CustomCorpusService.custom_corpora[4].text_parts = []
        with patch.object(CustomCorpusService, "init_custom_corpus",
                          return_value=CustomCorpusService.custom_corpora[4]):
            source_urn: str = CustomCorpusService.custom_corpora[4].corpus.source_urn
            reff = CustomCorpusService.get_custom_corpus_reff(source_urn)
            self.assertEqual(len(reff), 0)
        McTestCase.clear_folder(Config.REFF_CACHE_DIRECTORY)

    def test_get_custom_corpus_text(self):
        """ Retrieves the text for a custom corpus, e.g. a textbook. """
        text_list: List[ReferenceableText] = CustomCorpusService.get_custom_corpus_text(Mocks.urn)
        self.assertEqual(len(text_list), 0)

    def test_init_custom_corpus(self):
        """Adds custom corpora to the corpus list, e.g. the PROIEL corpora."""
        with patch.object(CustomCorpusService, "get_treebank_annotations", return_value=Mocks.annotations):
            cc: CustomCorpus = CustomCorpusService.init_custom_corpus(CustomCorpusService.custom_corpora[0])
            self.assertEqual(len(cc.text_parts), 1)

    def test_process_corpus_data(self):
        """Builds a graph from annotated text data."""
        disk_urn: str = AnnotationService.get_disk_urn(Mocks.urn_custom)
        AnnotationService.map_conll_to_graph(corpus_name=Mocks.urn_custom, conll=Mocks.annotations,
                                             file_name=disk_urn)
        result: dict = CorpusService.process_corpus_data(
            urn=Mocks.urn_custom, annotations=Mocks.annotations, aqls=[Phenomenon.UPOSTAG],
            exercise_type=ExerciseType.cloze, search_phenomena=[Phenomenon.UPOSTAG])
        gd: GraphData = AnnotationService.map_graph_data(result["graph_data_raw"])
        self.assertEqual(len(gd.nodes), len(Mocks.nodes))
        urn_parts: List[str] = Mocks.urn_custom.split(":")
        base_urn: str = Mocks.urn_custom.replace(":" + urn_parts[-1], "")
        target_corpus: CustomCorpus = next(
            (x for x in CustomCorpusService.custom_corpora if x.corpus.source_urn == base_urn), None)
        CustomCorpusService.init_custom_corpus(target_corpus)
        text_parts_list: List[ReferenceableText] = CorpusService.load_text_list(Mocks.urn_custom)
        self.assertEqual(len(text_parts_list), 1)


class CommonTestCase(unittest.TestCase):
    def setUp(self):
        """Initializes the testing environment."""
        self.start_time = time.time()
        self.class_name: str = str(self.__class__)
        TestHelper.update_flask_app(self.class_name, create_app)

    def tearDown(self):
        """Finishes testing by removing the traces."""
        print("{0}: {1} seconds".format(self.id(), "%.2f" % (time.time() - self.start_time)))

    def test_add_dependency_frequencies(self):
        """ Performs a frequency analysis for dependency annotations in a corpus. """
        gd_copy: GraphData = GraphData.from_dict(Mocks.graph_data.to_dict())
        gd_copy.links[0].udep_deprel = "safebpfw"
        gd_copy.links[48].udep_deprel = "fkonürwür"
        fis: List[FrequencyItem] = []
        FrequencyService.add_dependency_frequencies(gd_copy, fis)
        self.assertEqual(len(fis), 134)

    def test_add_edges(self):
        """Adds edges to an existing graph based on a list of keys and constraints to their similarity and frequency."""
        from mcserver.app.api.vectorNetworkAPI import add_edges
        w2v: Word2Vec = Word2Vec([x.split() for x in Mocks.raw_text.split(". ")], min_count=1, sample=0)
        graph: Graph = Graph()
        add_edges(["fortis"], w2v, 4, 1, graph)
        self.assertGreater(len(graph.edges), 1)

    def test_add_urn_to_sentences(self):
        """ Adds the relevant URN for every annotated sentence. """
        conll: List[TokenList] = copy.deepcopy(Mocks.annotations)
        text_list: List[ReferenceableText] = [
            ReferenceableText(conll[0].tokens[0]["form"], Mocks.urn),
            ReferenceableText("", Mocks.urn_custom), ReferenceableText(conll[0].tokens[0]["form"], Mocks.urn_custom)]
        conll[0].tokens[0]["form"] += "."
        conll.append(TokenList(tokens=[
            {"id": 1, "form": "Caesar.", "lemma": "Caeso", "upostag": "VERB", "xpostag": "L3|modJ|tem3|gen4|stAV",
             "feats": {"Mood": "Ind", "Number": "Sing", "Person": "1", "Tense": "Fut", "VerbForm": "Fin",
                       "Voice": "Pass"}, "head": 0, "deprel": "root", "deps": None, "misc": {"ref": "1.1"}}],
            metadata=OrderedDict([("sent_id", "2"), ("urn", "")])))
        conll += copy.deepcopy(Mocks.annotations)
        AnnotationService.add_urn_to_sentences(text_list, conll)
        self.assertEqual(conll[0].metadata["urn"], Mocks.urn)
        self.assertEqual(conll[1].metadata["urn"], "")
        self.assertEqual(conll[2].metadata["urn"], Mocks.urn_custom)

    def test_check_vocabulary(self):
        """ Checks whether the lemmata of a given graph/text match a reference vocabulary. """
        gd_copy: GraphData = GraphData.from_dict(Mocks.graph_data.to_dict())
        new_node: NodeMC = NodeMC.from_dict(gd_copy.nodes[0].to_dict())
        new_node.id = "salt:/urn:custom:latinLit:proiel.pal-agr.lat:1.1.1/doc1#sent159693tok1"
        gd_copy.nodes.append(NodeMC)
        sentences: List[Sentence] = check_vocabulary(Mocks.graph_data, {""})
        self.assertEqual(sentences[1].matching_degree, 3.225806451612903)

    def test_create_xml_string(self):
        """Exports the exercise data to the Moodle XML format. See https://docs.moodle.org/35/en/Moodle_XML_format ."""
        xml_string: str = XMLservice.create_xml_string(
            ExerciseMC.from_dict(exercise_type=ExerciseType.matching.value, last_access_time=0, eid=str(uuid.uuid4())),
            [], FileType.PDF, [])
        self.assertEqual(xml_string, Mocks.exercise_xml)

    def test_dependency_imports(self):
        """Verifies that all necessary dependencies are installed by trying to import them."""
        are_dependencies_missing = False
        try:
            import blinker  # for signalling
            import gunicorn  # for production server
            import psycopg2  # for database access via SQLAlchemy
            import coverage  # for code coverage in unit tests
        except ModuleNotFoundError:
            are_dependencies_missing = True
        self.assertFalse(are_dependencies_missing)

    def test_find_matches(self):
        """ Finds matches for a given URN and AQL and returns the corresponding node IDs. """
        expected_matches: List[str] = ["a", "b"]
        with patch.object(Config.CORPUS_STORAGE_MANAGER, "find",
                          side_effect=[[expected_matches], NoSuchCorpus(""), [expected_matches]]):
            matches: List[str] = CorpusService.find_matches(Mocks.urn_custom[:-6] + "3.1.1", "tok")
            self.assertEqual(matches, expected_matches)
            with patch.object(mcserver.app.services.corpusService.CorpusService, "get_corpus"):
                matches = CorpusService.find_matches(Mocks.urn, "")
                self.assertEqual(matches, expected_matches)

    def test_full_init(self):
        """ Fully initializes the application, including logging."""
        Mocks.app_dict[self.class_name].app.config["TESTING"] = False
        with patch.object(CorpusService, "init_graphannis_logging"):
            with patch.object(mcserver.app, "start_updater") as updater_mock:
                full_init(Mocks.app_dict[self.class_name].app)
                self.assertEqual(updater_mock.call_count, 1)
        Mocks.app_dict[self.class_name].app.config["TESTING"] = True
        db.session.query(UpdateInfo).delete()

    def test_get_annotations_from_string(self):
        """ Gets annotation data from a given string, be it a CoNLL string or a corpus URN. """
        conll: List[TokenList]
        with patch.object(AnnotationService, "get_udpipe", return_value=Mocks.udpipe_string):
            with patch.object(CorpusService, "load_text_list", return_value=Mocks.text_list):
                with patch.object(CorpusService, "get_raw_text", return_value=Mocks.raw_text):
                    conll = CorpusService.get_annotations_from_string(Mocks.urn)
                    self.assertEqual(len(conll[0]), 5)
                mdg: MultiDiGraph = CorpusService.get_graph(Mocks.urn)
                self.assertEqual(len(mdg.nodes), 5)
                mdg = CorpusService.get_graph(f"{Mocks.urn}@1-1")
                self.assertEqual(len(mdg.nodes), 5)
        with patch.object(CustomCorpusService, "get_treebank_annotations", return_value=Mocks.annotations):
            conll = CorpusService.get_annotations_from_string(Mocks.urn_custom)
            self.assertEqual(len(conll[0]), 6)
        with patch.object(CustomCorpusService, "get_custom_corpus_annotations",
                          return_value=Mocks.annotations * 2):
            urn: str = f"{Config.CUSTOM_CORPUS_VIVA_URN}:1.1-1.1"
            conll = CorpusService.get_annotations_from_string(urn)
            self.assertEqual(len(conll), 2)

    def test_get_concept_network(self):
        """Extracts a network of words from vector data in an AI model."""
        from mcserver.app.api.vectorNetworkAPI import get_concept_network
        with patch.object(mcserver.app.api.vectorNetworkAPI, "add_edges", side_effect=Mocks.mock_add_eges):
            with patch.object(mcserver.app.api.vectorNetworkAPI.Word2Vec, "load", return_value=MockW2V()):
                svg_string: str = get_concept_network("ueritas", highlight_regex_string="uera")
                self.assertGreater(len(svg_string), 6500)

    def test_get_frequency_analysis(self):
        """ Collects frequency statistics for various combinations of linguistic annotations in a corpus. """
        with patch.object(
                mcserver.app.services.corpusService.CorpusService, "get_frequency_analysis",
                return_value=[FrequencyItem(values=[], phenomena=[], count=0)]):
            fa: List[FrequencyItem] = CorpusService.get_frequency_analysis(urn=Mocks.urn_custom)
            self.assertEqual(len(fa), 1)
        CorpusService.get_corpus(Mocks.urn_custom)
        with patch.object(CorpusService, "get_corpus", return_value=Mocks.annis_response):
            fa = CorpusService.get_frequency_analysis(Mocks.urn_custom)
            self.assertEqual(len(fa), 214)

    def test_get_graph(self):
        """ Retrieves a graph from the cache or, if not there, builds it from scratch. """
        expected_mdg: MultiDiGraph = MultiDiGraph([(1, 2), (2, 3), (3, 4)])
        with patch.object(Config.CORPUS_STORAGE_MANAGER, "subcorpus_graph", return_value=expected_mdg):
            mdg: MultiDiGraph = CorpusService.get_graph(Mocks.urn)
            self.assertEqual(mdg, expected_mdg)

    def test_get_graph_data(self):
        """Sends annotated text data or a URN to the Corpus Storage Manager in order to get a graph."""
        with patch.object(mcserver.app.services.corpusService.CorpusService, "get_annotations_from_string",
                          return_value=Mocks.annotations):
            with patch.object(mcserver.app.services.corpusService.CorpusService, "process_corpus_data",
                              return_value=Mocks.annis_response_dict):
                result: dict = get_graph_data("", "", [], ExerciseType.matching, [])
                self.assertEqual(result, Mocks.annis_response_dict)

    def test_get_matches(self):
        """Retrieves search results from ANNIS for a given corpus and AQL query."""
        disk_urn: str = AnnotationService.get_disk_urn(Mocks.urn_custom)
        AnnotationService.map_conll_to_graph(corpus_name=Mocks.urn_custom, conll=Mocks.annotations,
                                             file_name=disk_urn)
        solutions: List[Solution] = CorpusService.get_matches(Mocks.urn_custom, ['tok ->dep tok'],
                                                              [Phenomenon.DEPENDENCY])
        self.assertEqual(len(solutions), 5)
        solutions = CorpusService.get_matches(Mocks.urn_custom, ['upostag="VERB" ->dep tok'],
                                              [Phenomenon.UPOSTAG, Phenomenon.DEPENDENCY])
        self.assertEqual(len(solutions), 5)
        solutions = CorpusService.get_matches(Mocks.urn_custom, ['tok ->dep tok ->dep tok'],
                                              [Phenomenon.DEPENDENCY, Phenomenon.UPOSTAG])
        self.assertEqual(len(solutions), 3)

    def test_get_pdf_html_string(self):
        """ Builds an HTML string from an exercise, e.g. to construct a PDF from it. """
        Mocks.exercise.exercise_type = ExerciseType.matching.value
        solutions: List[Solution] = [Solution.from_dict(x) for x in json.loads(Mocks.exercise.solutions)]
        result: str = FileService.get_pdf_html_string(Mocks.exercise, Mocks.annotations, FileType.PDF, solutions)
        expected_result: str = '<br><p>Cloze: Assign the words from the pool to the correct gaps!</p><p><table><tr><td>praecepturus</td><td>Caesar</td></tr></table></p>'
        self.assertEqual(result, expected_result)
        Mocks.exercise.exercise_type = ExerciseType.markWords.value
        result = FileService.get_pdf_html_string(Mocks.exercise, Mocks.annotations, FileType.PDF, solutions)
        expected_result = '<p>Cloze: Assign the words from the pool to the correct gaps!</p><p>Caesar et Galli fortes sunt.</p><br><br>'
        self.assertEqual(result, expected_result)
        Mocks.exercise.exercise_type = ExerciseType.cloze.value

    def test_get_raw_text(self):
        """ Retrieves the raw text for a corpus. """
        with patch.object(CorpusService, "get_corpus", return_value=Mocks.annis_response):
            text: str = CorpusService.get_raw_text(Mocks.urn)
            self.assertEqual(len(text), 349)

    def test_get_solutions_by_index(self):
        """ If available, makes use of the solution indices to return only the wanted solutions. """
        solutions: List[Solution] = TextService.get_solutions_by_index(Mocks.exercise, [])
        self.assertEqual(len(solutions), 1)

    def test_get_subgraph(self):
        """ Retrieves subgraph data for a given URN. """
        with patch.object(mcserver.app.services.annotationService.AnnotationService, "get_single_subgraph",
                          return_value=Mocks.graph_data):
            ar: AnnisResponse = CorpusService.get_subgraph(Mocks.urn_custom, 'tok="quarum"', 0, 0)
            self.assertEqual(len(ar.graph_data.nodes), len(Mocks.graph_data.nodes))

    def test_get_treebank_annotations(self):
        """ Retrieves annotations from a treebank. """
        cache_path: str = os.path.join(Config.TREEBANKS_CACHE_DIRECTORY,
                                       ntpath.basename(CustomCorpusService.custom_corpora[4].file_path) + ".json")
        if os.path.exists(cache_path):
            os.remove(cache_path)
        with patch.object(mcserver.app.services.customCorpusService.conllu, "parse",
                          return_value=Mocks.annotations) as parse_mock:
            with patch.object(CustomCorpusService, "get_treebank_sub_annotations", return_value=Mocks.annotations):
                conll: List[TokenList] = CustomCorpusService.get_treebank_annotations(Mocks.urn_custom)
                self.assertIs(conll, Mocks.annotations)
                with patch.object(mcserver.app.services.customCorpusService.json, "loads", side_effect=ValueError):
                    CustomCorpusService.get_treebank_annotations(Mocks.urn_custom)
                    os.remove(cache_path)
                    self.assertEqual(parse_mock.call_count, 2)

    def test_get_treebank_sub_annotations(self):
        """ Retrieves annotations for nested parts of a treebank. """
        annotations: List[TokenList] = Mocks.annotations + \
                                       [TokenList([], metadata=OrderedDict([("sent_id", "2")])),
                                        TokenList([], metadata=OrderedDict([("sent_id", "3")]))]
        conll: List[TokenList] = CustomCorpusService.get_treebank_sub_annotations(
            Mocks.urn + "@1-3", annotations, CustomCorpusService.custom_corpora[4])
        self.assertEqual(len(conll), 3)
        cc: CustomCorpus = CustomCorpusService.custom_corpora[-1]
        urn: str = cc.corpus.source_urn + ":1.1-1.2"
        CustomCorpusService.get_treebank_sub_annotations(urn, [], cc)
        self.assertEqual(len(cc.text_parts), 2)
        conll = CustomCorpusService.get_treebank_sub_annotations(cc.corpus.source_urn + "I.1-I.2", [], cc)
        self.assertEqual(conll, [])

    def test_get_udpipe(self):
        """Annotates a single text with UdPipe. The beginning of the CONLL has to be left out because it contains the
        randomly generated temp file path and thus cannot be predicted exactly."""
        text: str = "Caesar fortis est. Galli moriuntur."
        conll_string: str = AnnotationService.get_udpipe(text)
        self.assertIn(Mocks.udpipe_string, conll_string)

    def test_handle_exercise_data(self):
        """ Constructs an SVG image (for POS and syntactic dependencies) from given annotations. """
        result: str = handle_exercise_data(ExerciseData(json_dict=Mocks.exercise_data.serialize()), 5, 5)
        self.assertTrue(result.startswith('<svg height="160" id="svg1" width="224">'))

    def test_init_db_alembic(self):
        """In Docker, the alembic version is not initially written to the database, so we need to set it manually."""
        if db.engine.dialect.has_table(db.engine, Config.DATABASE_TABLE_ALEMBIC):
            db.engine.execute(f"DROP TABLE {Config.DATABASE_TABLE_ALEMBIC}")
        self.assertEqual(db.engine.dialect.has_table(db.engine, Config.DATABASE_TABLE_ALEMBIC), False)
        DatabaseService.init_db_alembic()
        self.assertEqual(db.engine.dialect.has_table(db.engine, Config.DATABASE_TABLE_ALEMBIC), True)

    def test_init_db_corpus(self):
        """Initializes the corpus table."""
        db.session.query(Corpus).delete()
        cc: CustomCorpus = CustomCorpusService.custom_corpora[0]
        old_corpus: Corpus = Mocks.corpora[0]
        old_corpus.source_urn = cc.corpus.source_urn
        McTestCase.add_corpus(old_corpus)
        del old_corpus
        CorpusService.init_corpora()
        corpus: Corpus = DatabaseService.query(
            Corpus, filter_by=dict(source_urn=cc.corpus.source_urn), first=True)
        self.assertEqual(corpus.title, cc.corpus.title)
        db.session.query(Corpus).delete()
        db.session.query(UpdateInfo).delete()

    def test_init_stop_words_latin(self):
        """Initializes the stop words list for Latin texts and caches it if necessary."""

        def clear_cache():
            if os.path.exists(Config.STOP_WORDS_LATIN_PATH):
                os.remove(Config.STOP_WORDS_LATIN_PATH)

        clear_cache()
        stop_word_list: Dict[str, List[str]] = {"a": ["b"]}
        mr: MockResponse = MockResponse(json.dumps(stop_word_list))
        with patch.object(mcserver.app.services.textService.requests, "get", return_value=mr) as \
                mock_get_request:
            TextService.init_stop_words_latin()
            self.assertEqual(len(TextService.stop_words_latin), 1)