test_question_cmd: test fixes and cleanup

chatmastermind/ai_factory.py

@@ -3,18 +3,20 @@ Creates different AI instances, based on the given configuration.
 """
 
 import argparse
-from typing import cast
+from typing import cast, Optional
 from .configuration import Config, AIConfig, OpenAIConfig
 from .ai import AI, AIError
 from .ais.openai import OpenAI
 
 
-def create_ai(args: argparse.Namespace, config: Config) -> AI:  # noqa: 11
+def create_ai(args: argparse.Namespace, config: Config,  # noqa: 11
+              def_ai: Optional[str] = None,
+              def_model: Optional[str] = None) -> AI:
     """
-    Creates an AI subclass instance from the given arguments
-    and configuration file. If AI has not been set in the
-    arguments, it searches for the ID 'default'. If that
-    is not found, it uses the first AI in the list.
+    Creates an AI subclass instance from the given arguments and configuration file.
+    If AI has not been set in the arguments, it searches for the ID 'default'. If
+    that is not found, it uses the first AI in the list. It's also possible to
+    specify a default AI and model using 'def_ai' and 'def_model'.
     """
     ai_conf: AIConfig
     if hasattr(args, 'AI') and args.AI:
@@ -22,6 +24,8 @@ def create_ai(args: argparse.Namespace, config: Config) -> AI:  # noqa: 11
             ai_conf = config.ais[args.AI]
         except KeyError:
             raise AIError(f"AI ID '{args.AI}' does not exist in this configuration")
+    elif def_ai:
+        ai_conf = config.ais[def_ai]
     elif 'default' in config.ais:
         ai_conf = config.ais['default']
     else:
@@ -34,6 +38,8 @@ def create_ai(args: argparse.Namespace, config: Config) -> AI:  # noqa: 11
         ai = OpenAI(cast(OpenAIConfig, ai_conf))
         if hasattr(args, 'model') and args.model:
             ai.config.model = args.model
+        elif def_model:
+            ai.config.model = def_model
         if hasattr(args, 'max_tokens') and args.max_tokens:
             ai.config.max_tokens = args.max_tokens
         if hasattr(args, 'temperature') and args.temperature:

chatmastermind/ais/openai.py

@@ -44,7 +44,7 @@ class OpenAI(AI):
             frequency_penalty=self.config.frequency_penalty,
             presence_penalty=self.config.presence_penalty)
         question.answer = Answer(response['choices'][0]['message']['content'])
-        question.tags = otags
+        question.tags = set(otags) if otags is not None else None
         question.ai = self.ID
         question.model = self.config.model
         answers: list[Message] = [question]

chatmastermind/commands/question.py

@@ -2,6 +2,7 @@ import sys
 import argparse
 from pathlib import Path
 from itertools import zip_longest
+from copy import deepcopy
 from ..configuration import Config
 from ..chat import ChatDB
 from ..message import Message, MessageFilter, MessageError, Question, source_code
@@ -105,13 +106,75 @@ def make_request(ai: AI, chat: ChatDB, message: Message, args: argparse.Namespac
         print(response.tokens)
 
 
+def create_msg_args(msg: Message, args: argparse.Namespace) -> argparse.Namespace:
+    """
+    Takes an existing message and CLI arguments, and returns modified args based
+    on the members of the given message. Used e.g. when repeating messages, where
+    it's necessary to determine the correct AI, module and output tags to use
+    (either from the existing message or the given args).
+    """
+    msg_args = args
+    # if AI, model or output tags have not been specified,
+    # use those from the original message
+    if (args.AI is None
+       or args.model is None  # noqa: W503
+       or args.output_tags is None):  # noqa: W503
+        msg_args = deepcopy(args)
+        if args.AI is None and msg.ai is not None:
+            msg_args.AI = msg.ai
+        if args.model is None and msg.model is not None:
+            msg_args.model = msg.model
+        if args.output_tags is None and msg.tags is not None:
+            msg_args.output_tags = msg.tags
+    return msg_args
+
+
+def repeat_messages(messages: list[Message], chat: ChatDB, args: argparse.Namespace, config: Config) -> None:
+    """
+    Repeat the given messages using the given arguments.
+    """
+    ai: AI
+    for msg in messages:
+        msg_args = create_msg_args(msg, args)
+        ai = create_ai(msg_args, config)
+        print(f"--------- Repeating message '{msg.msg_id()}': ---------")
+        # overwrite the latest message if requested or empty
+        # -> but not if it's in the DB!
+        if ((msg.answer is None or msg_args.overwrite is True)
+           and (not chat.msg_in_db(msg))):  # noqa: W503
+            msg.clear_answer()
+            make_request(ai, chat, msg, msg_args)
+        # otherwise create a new one
+        else:
+            msg_args.ask = [msg.question]
+            message = create_message(chat, msg_args)
+            make_request(ai, chat, message, msg_args)
+
+
+def invert_input_tag_args(args: argparse.Namespace) -> None:
+    """
+    Changes the semantics of the INPUT tags for this command:
+    * not tags specified on the CLI -> no tags are selected
+    * empty tags specified on the CLI -> all tags are selected
+    """
+    if args.or_tags is None:
+        args.or_tags = set()
+    elif len(args.or_tags) == 0:
+        args.or_tags = None
+    if args.and_tags is None:
+        args.and_tags = set()
+    elif len(args.and_tags) == 0:
+        args.and_tags = None
+
+
 def question_cmd(args: argparse.Namespace, config: Config) -> None:
     """
     Handler for the 'question' command.
     """
-    mfilter = MessageFilter(tags_or=args.or_tags if args.or_tags is not None else set(),
-                            tags_and=args.and_tags if args.and_tags is not None else set(),
-                            tags_not=args.exclude_tags if args.exclude_tags is not None else set())
+    invert_input_tag_args(args)
+    mfilter = MessageFilter(tags_or=args.or_tags,
+                            tags_and=args.and_tags,
+                            tags_not=args.exclude_tags)
     chat = ChatDB.from_dir(cache_path=Path(config.cache),
                            db_path=Path(config.db),
                            mfilter=mfilter)
@@ -121,30 +184,24 @@ def question_cmd(args: argparse.Namespace, config: Config) -> None:
     if args.create:
         return
 
-    # create the correct AI instance
-    ai: AI = create_ai(args, config)
-
     # === ASK ===
     if args.ask:
+        ai: AI = create_ai(args, config)
         make_request(ai, chat, message, args)
     # === REPEAT ===
     elif args.repeat is not None:
+        repeat_msgs: list[Message] = []
+        # repeat latest message
+        if len(args.repeat) == 0:
             lmessage = chat.msg_latest(loc='cache')
             if lmessage is None:
                 print("No message found to repeat!")
                 sys.exit(1)
+            repeat_msgs.append(lmessage)
+        # repeat given message(s)
         else:
-            print(f"Repeating message '{lmessage.msg_id()}':")
-        # overwrite the latest message if requested or empty
-        if lmessage.answer is None or args.overwrite is True:
-            lmessage.clear_answer()
-            make_request(ai, chat, lmessage, args)
-        # otherwise create a new one
-        else:
-            args.ask = [lmessage.question]
-            message = create_message(chat, args)
-            make_request(ai, chat, message, args)
-
+            repeat_msgs = chat.msg_find(args.repeat, loc='disk')
+        repeat_messages(repeat_msgs, chat, args, config)
     # === PROCESS ===
     elif args.process is not None:
         # TODO: process either all questions without an

chatmastermind/main.py

@@ -34,13 +34,13 @@ def create_parser() -> argparse.ArgumentParser:
 
     # a parent parser for all commands that support tag selection
     tag_parser = argparse.ArgumentParser(add_help=False)
-    tag_arg = tag_parser.add_argument('-t', '--or-tags', nargs='+',
+    tag_arg = tag_parser.add_argument('-t', '--or-tags', nargs='*',
                                       help='List of tags (one must match)', metavar='OTAGS')
     tag_arg.completer = tags_completer  # type: ignore
-    atag_arg = tag_parser.add_argument('-k', '--and-tags', nargs='+',
+    atag_arg = tag_parser.add_argument('-k', '--and-tags', nargs='*',
                                        help='List of tags (all must match)', metavar='ATAGS')
     atag_arg.completer = tags_completer  # type: ignore
-    etag_arg = tag_parser.add_argument('-x', '--exclude-tags', nargs='+',
+    etag_arg = tag_parser.add_argument('-x', '--exclude-tags', nargs='*',
                                        help='List of tags to exclude', metavar='XTAGS')
     etag_arg.completer = tags_completer  # type: ignore
     otag_arg = tag_parser.add_argument('-o', '--output-tags', nargs='+',

tests/test_question_cmd.py

@@ -4,9 +4,9 @@ import argparse
 import tempfile
 from pathlib import Path
 from unittest import mock
-from unittest.mock import MagicMock, call, ANY
-from typing import Optional
-from chatmastermind.configuration import Config
+from unittest.mock import MagicMock, call
+from typing import Optional, Union
+from chatmastermind.configuration import Config, AIConfig
 from chatmastermind.commands.question import create_message, question_cmd
 from chatmastermind.tags import Tag
 from chatmastermind.message import Message, Question, Answer
@@ -14,6 +14,56 @@ from chatmastermind.chat import Chat, ChatDB
 from chatmastermind.ai import AI, AIResponse, Tokens, AIError
 
 
+class FakeAI(AI):
+    """
+    A mocked version of the 'AI' class.
+    """
+    ID: str
+    name: str
+    config: AIConfig
+
+    def models(self) -> list[str]:
+        raise NotImplementedError
+
+    def tokens(self, data: Union[Message, Chat]) -> int:
+        return 123
+
+    def print(self) -> None:
+        pass
+
+    def print_models(self) -> None:
+        pass
+
+    def __init__(self, ID: str, model: str, error: bool = False):
+        self.ID = ID
+        self.model = model
+        self.error = error
+
+    def request(self,
+                question: Message,
+                chat: Chat,
+                num_answers: int = 1,
+                otags: Optional[set[Tag]] = None) -> AIResponse:
+        """
+        Mock the 'ai.request()' function by either returning fake
+        answers or raising an exception.
+        """
+        if self.error:
+            raise AIError
+        question.answer = Answer("Answer 0")
+        question.tags = set(otags) if otags is not None else None
+        question.ai = self.ID
+        question.model = self.model
+        answers: list[Message] = [question]
+        for n in range(1, num_answers):
+            answers.append(Message(question=question.question,
+                                   answer=Answer(f"Answer {n}"),
+                                   tags=otags,
+                                   ai=self.ID,
+                                   model=self.model))
+        return AIResponse(answers, Tokens(10, 10, 20))
+
+
 class TestQuestionCmdBase(unittest.TestCase):
     def assert_messages_equal(self, msg1: list[Message], msg2: list[Message]) -> None:
         """
@@ -24,6 +74,18 @@ class TestQuestionCmdBase(unittest.TestCase):
             # exclude the file_path, compare only Q, A and metadata
             self.assertTrue(m1.equals(m2, file_path=False, verbose=True))
 
+    def mock_create_ai(self, args: argparse.Namespace, config: Config) -> AI:
+        """
+        Mocked 'create_ai' that returns a 'FakeAI' instance.
+        """
+        return FakeAI(args.AI, args.model)
+
+    def mock_create_ai_with_error(self, args: argparse.Namespace, config: Config) -> AI:
+        """
+        Mocked 'create_ai' that returns a 'FakeAI' instance.
+        """
+        return FakeAI(args.AI, args.model, error=True)
+
 
 class TestMessageCreate(TestQuestionCmdBase):
     """
@@ -227,8 +289,8 @@ class TestQuestionCmd(TestQuestionCmdBase):
             ask=['What is the meaning of life?'],
             num_answers=1,
             output_tags=['science'],
-            AI='openai',
-            model='gpt-3.5-turbo',
+            AI='FakeAI',
+            model='FakeModel',
             or_tags=None,
             and_tags=None,
             exclude_tags=None,
@@ -239,57 +301,37 @@ class TestQuestionCmd(TestQuestionCmdBase):
             process=None,
             overwrite=None
         )
-        # create a mock AI instance
-        self.ai = MagicMock(spec=AI)
-        self.ai.request.side_effect = self.mock_request
 
-    def input_message(self, args: argparse.Namespace) -> Message:
-        """
-        Create the expected input message for a question using the
-        given arguments.
-        """
-        # NOTE: we only use the first question from the "ask" list
-        # -> message creation using "question.create_message()" is
-        #    tested above
-        # the answer is always empty for the input message
-        return Message(Question(args.ask[0]),
-                       tags=args.output_tags,
+    def create_single_message(self, args: argparse.Namespace, with_answer: bool = True) -> Message:
+        message = Message(Question(args.ask[0]),
+                          tags=set(args.output_tags) if args.output_tags is not None else None,
                           ai=args.AI,
-                       model=args.model)
-
-    def mock_request(self,
-                     question: Message,
-                     chat: Chat,
-                     num_answers: int = 1,
-                     otags: Optional[set[Tag]] = None) -> AIResponse:
-        """
-        Mock the 'ai.request()' function
-        """
-        question.answer = Answer("Answer 0")
-        question.tags = set(otags) if otags else None
-        question.ai = 'FakeAI'
-        question.model = 'FakeModel'
-        answers: list[Message] = [question]
-        for n in range(1, num_answers):
-            answers.append(Message(question=question.question,
-                                   answer=Answer(f"Answer {n}"),
-                                   tags=otags,
-                                   ai='FakeAI',
-                                   model='FakeModel'))
-        return AIResponse(answers, Tokens(10, 10, 20))
+                          model=args.model,
+                          file_path=Path(self.cache_dir.name) / '0001.txt')
+        if with_answer:
+            message.answer = Answer('Answer 0')
+        message.to_file()
+        return message
 
     def message_list(self, tmp_dir: tempfile.TemporaryDirectory) -> list[Path]:
         # exclude '.next'
         return sorted([f for f in Path(tmp_dir.name).glob('*.[ty]*')])
 
+
+class TestQuestionCmdAsk(TestQuestionCmd):
+
     @mock.patch('chatmastermind.commands.question.create_ai')
     def test_ask_single_answer(self, mock_create_ai: MagicMock) -> None:
         """
         Test single answer with no errors.
         """
-        mock_create_ai.return_value = self.ai
-        expected_question = self.input_message(self.args)
-        expected_responses = self.mock_request(expected_question,
+        mock_create_ai.side_effect = self.mock_create_ai
+        expected_question = Message(Question(self.args.ask[0]),
+                                    tags=self.args.output_tags,
+                                    ai=self.args.AI,
+                                    model=self.args.model)
+        fake_ai = self.mock_create_ai(self.args, self.config)
+        expected_responses = fake_ai.request(expected_question,
                                              Chat([]),
                                              self.args.num_answers,
                                              self.args.output_tags).messages
@@ -297,11 +339,6 @@ class TestQuestionCmd(TestQuestionCmdBase):
         # execute the command
         question_cmd(self.args, self.config)
 
-        # check for correct request call
-        self.ai.request.assert_called_once_with(expected_question,
-                                                ANY,
-                                                self.args.num_answers,
-                                                self.args.output_tags)
         # check for the expected message files
         chat = ChatDB.from_dir(Path(self.cache_dir.name),
                                Path(self.db_dir.name))
@@ -318,9 +355,13 @@ class TestQuestionCmd(TestQuestionCmdBase):
         chat = MagicMock(spec=ChatDB)
         mock_from_dir.return_value = chat
 
-        mock_create_ai.return_value = self.ai
-        expected_question = self.input_message(self.args)
-        expected_responses = self.mock_request(expected_question,
+        mock_create_ai.side_effect = self.mock_create_ai
+        expected_question = Message(Question(self.args.ask[0]),
+                                    tags=self.args.output_tags,
+                                    ai=self.args.AI,
+                                    model=self.args.model)
+        fake_ai = self.mock_create_ai(self.args, self.config)
+        expected_responses = fake_ai.request(expected_question,
                                              Chat([]),
                                              self.args.num_answers,
                                              self.args.output_tags).messages
@@ -328,12 +369,6 @@ class TestQuestionCmd(TestQuestionCmdBase):
         # execute the command
         question_cmd(self.args, self.config)
 
-        # check for correct request call
-        self.ai.request.assert_called_once_with(expected_question,
-                                                chat,
-                                                self.args.num_answers,
-                                                self.args.output_tags)
-
         # check for the correct ChatDB calls:
         # - initial question has been written (prior to the actual request)
         # - responses have been written (after the request)
@@ -350,19 +385,16 @@ class TestQuestionCmd(TestQuestionCmdBase):
         Provoke an error during the AI request and verify that the question
         has been correctly stored in the cache.
         """
-        mock_create_ai.return_value = self.ai
-        expected_question = self.input_message(self.args)
-        self.ai.request.side_effect = AIError
+        mock_create_ai.side_effect = self.mock_create_ai_with_error
+        expected_question = Message(Question(self.args.ask[0]),
+                                    tags=self.args.output_tags,
+                                    ai=self.args.AI,
+                                    model=self.args.model)
 
         # execute the command
         with self.assertRaises(AIError):
             question_cmd(self.args, self.config)
 
-        # check for correct request call
-        self.ai.request.assert_called_once_with(expected_question,
-                                                ANY,
-                                                self.args.num_answers,
-                                                self.args.output_tags)
         # check for the expected message files
         chat = ChatDB.from_dir(Path(self.cache_dir.name),
                                Path(self.db_dir.name))
@@ -370,33 +402,34 @@ class TestQuestionCmd(TestQuestionCmdBase):
         self.assertEqual(len(self.message_list(self.cache_dir)), 1)
         self.assert_messages_equal(cached_msg, [expected_question])
 
+
+class TestQuestionCmdRepeat(TestQuestionCmd):
+
     @mock.patch('chatmastermind.commands.question.create_ai')
     def test_repeat_single_question(self, mock_create_ai: MagicMock) -> None:
         """
         Repeat a single question.
         """
-        # 1. ask a question
-        mock_create_ai.return_value = self.ai
-        expected_question = self.input_message(self.args)
-        expected_responses = self.mock_request(expected_question,
-                                               Chat([]),
-                                               self.args.num_answers,
-                                               self.args.output_tags).messages
-        question_cmd(self.args, self.config)
-        chat = ChatDB.from_dir(Path(self.cache_dir.name),
-                               Path(self.db_dir.name))
-        cached_msg = chat.msg_gather(loc='cache')
-        self.assertEqual(len(self.message_list(self.cache_dir)), 1)
-        self.assert_messages_equal(cached_msg, expected_responses)
+        mock_create_ai.side_effect = self.mock_create_ai
+        # create a message
+        message = self.create_single_message(self.args)
 
-        # 2. repeat the last question (without overwriting)
+        # repeat the last question (without overwriting)
         # -> expect two identical messages (except for the file_path)
         self.args.ask = None
         self.args.repeat = []
         self.args.overwrite = False
-        expected_responses += expected_responses
+        fake_ai = self.mock_create_ai(self.args, self.config)
+        expected_response = fake_ai.request(message,
+                                            Chat([]),
+                                            self.args.num_answers,
+                                            set(self.args.output_tags)).messages
+        expected_responses = expected_response + expected_response
         question_cmd(self.args, self.config)
+        chat = ChatDB.from_dir(Path(self.cache_dir.name),
+                               Path(self.db_dir.name))
         cached_msg = chat.msg_gather(loc='cache')
+        print(self.message_list(self.cache_dir))
         self.assertEqual(len(self.message_list(self.cache_dir)), 2)
         self.assert_messages_equal(cached_msg, expected_responses)
 
@@ -405,31 +438,29 @@ class TestQuestionCmd(TestQuestionCmdBase):
         """
         Repeat a single question and overwrite the old one.
         """
-        # 1. ask a question
-        mock_create_ai.return_value = self.ai
-        expected_question = self.input_message(self.args)
-        expected_responses = self.mock_request(expected_question,
-                                               Chat([]),
-                                               self.args.num_answers,
-                                               self.args.output_tags).messages
-        question_cmd(self.args, self.config)
+        mock_create_ai.side_effect = self.mock_create_ai
+        # create a message
+        message = self.create_single_message(self.args)
         chat = ChatDB.from_dir(Path(self.cache_dir.name),
                                Path(self.db_dir.name))
         cached_msg = chat.msg_gather(loc='cache')
         assert cached_msg[0].file_path
         cached_msg_file_id = cached_msg[0].file_path.stem
-        self.assertEqual(len(self.message_list(self.cache_dir)), 1)
-        self.assert_messages_equal(cached_msg, expected_responses)
 
-        # 2. repeat the last question (WITH overwriting)
+        # repeat the last question (WITH overwriting)
         # -> expect a single message afterwards
         self.args.ask = None
         self.args.repeat = []
         self.args.overwrite = True
+        fake_ai = self.mock_create_ai(self.args, self.config)
+        expected_response = fake_ai.request(message,
+                                            Chat([]),
+                                            self.args.num_answers,
+                                            set(self.args.output_tags)).messages
         question_cmd(self.args, self.config)
         cached_msg = chat.msg_gather(loc='cache')
         self.assertEqual(len(self.message_list(self.cache_dir)), 1)
-        self.assert_messages_equal(cached_msg, expected_responses)
+        self.assert_messages_equal(cached_msg, expected_response)
         # also check that the file ID has not been changed
         assert cached_msg[0].file_path
         self.assertEqual(cached_msg_file_id, cached_msg[0].file_path.stem)
@@ -439,35 +470,31 @@ class TestQuestionCmd(TestQuestionCmdBase):
         """
         Repeat a single question after an error.
         """
-        # 1. ask a question and provoke an error
-        mock_create_ai.return_value = self.ai
-        expected_question = self.input_message(self.args)
-        self.ai.request.side_effect = AIError
-        with self.assertRaises(AIError):
-            question_cmd(self.args, self.config)
+        mock_create_ai.side_effect = self.mock_create_ai
+        # create a question WITHOUT an answer
+        # -> just like after an error, which is tested above
+        question = self.create_single_message(self.args, with_answer=False)
         chat = ChatDB.from_dir(Path(self.cache_dir.name),
                                Path(self.db_dir.name))
         cached_msg = chat.msg_gather(loc='cache')
         assert cached_msg[0].file_path
         cached_msg_file_id = cached_msg[0].file_path.stem
-        self.assertEqual(len(self.message_list(self.cache_dir)), 1)
-        self.assert_messages_equal(cached_msg, [expected_question])
 
-        # 2. repeat the last question (without overwriting)
+        # repeat the last question (without overwriting)
         # -> expect a single message because if the original has
         #    no answer, it should be overwritten by default
         self.args.ask = None
         self.args.repeat = []
         self.args.overwrite = False
-        self.ai.request.side_effect = self.mock_request
-        expected_responses = self.mock_request(expected_question,
+        fake_ai = self.mock_create_ai(self.args, self.config)
+        expected_response = fake_ai.request(question,
                                             Chat([]),
                                             self.args.num_answers,
                                             self.args.output_tags).messages
         question_cmd(self.args, self.config)
         cached_msg = chat.msg_gather(loc='cache')
         self.assertEqual(len(self.message_list(self.cache_dir)), 1)
-        self.assert_messages_equal(cached_msg, expected_responses)
+        self.assert_messages_equal(cached_msg, expected_response)
         # also check that the file ID has not been changed
         assert cached_msg[0].file_path
         self.assertEqual(cached_msg_file_id, cached_msg[0].file_path.stem)
@@ -477,21 +504,15 @@ class TestQuestionCmd(TestQuestionCmdBase):
         """
         Repeat a single question with new arguments.
         """
-        # 1. ask a question
-        mock_create_ai.return_value = self.ai
-        expected_question = self.input_message(self.args)
-        expected_responses = self.mock_request(expected_question,
-                                               Chat([]),
-                                               self.args.num_answers,
-                                               self.args.output_tags).messages
-        question_cmd(self.args, self.config)
+        mock_create_ai.side_effect = self.mock_create_ai
+        # create a message
+        message = self.create_single_message(self.args)
         chat = ChatDB.from_dir(Path(self.cache_dir.name),
                                Path(self.db_dir.name))
         cached_msg = chat.msg_gather(loc='cache')
-        self.assertEqual(len(self.message_list(self.cache_dir)), 1)
-        self.assert_messages_equal(cached_msg, expected_responses)
+        assert cached_msg[0].file_path
 
-        # 2. repeat the last question with new arguments (without overwriting)
+        # repeat the last question with new arguments (without overwriting)
         # -> expect two messages with identical question and answer, but different metadata
         self.args.ask = None
         self.args.repeat = []
@@ -499,15 +520,69 @@ class TestQuestionCmd(TestQuestionCmdBase):
         self.args.output_tags = ['newtag']
         self.args.AI = 'newai'
         self.args.model = 'newmodel'
-        new_expected_question = Message(question=Question(expected_question.question),
+        new_expected_question = Message(question=Question(message.question),
                                         tags=set(self.args.output_tags),
                                         ai=self.args.AI,
                                         model=self.args.model)
-        expected_responses += self.mock_request(new_expected_question,
+        fake_ai = self.mock_create_ai(self.args, self.config)
+        new_expected_response = fake_ai.request(new_expected_question,
                                                 Chat([]),
                                                 self.args.num_answers,
                                                 set(self.args.output_tags)).messages
         question_cmd(self.args, self.config)
         cached_msg = chat.msg_gather(loc='cache')
         self.assertEqual(len(self.message_list(self.cache_dir)), 2)
-        self.assert_messages_equal(cached_msg, expected_responses)
+        self.assert_messages_equal(cached_msg, [message] + new_expected_response)
+
+    @mock.patch('chatmastermind.commands.question.create_ai')
+    def test_repeat_multiple_questions(self, mock_create_ai: MagicMock) -> None:
+        """
+        Repeat multiple  questions.
+        """
+        # 1. === create three questions ===
+        # cached message without an answer
+        message1 = Message(Question('Question 1'),
+                           ai='foo',
+                           model='bla',
+                           file_path=Path(self.cache_dir.name) / '0001.txt')
+        # cached message with an answer
+        message2 = Message(Question('Question 2'),
+                           Answer('Answer 0'),
+                           ai='openai',
+                           model='gpt-3.5-turbo',
+                           file_path=Path(self.cache_dir.name) / '0002.txt')
+        # DB message without an answer
+        message3 = Message(Question('Question 3'),
+                           ai='openai',
+                           model='gpt-3.5-turbo',
+                           file_path=Path(self.db_dir.name) / '0003.txt')
+        message1.to_file()
+        message2.to_file()
+        message3.to_file()
+        questions = [message1, message2, message3]
+        expected_responses: list[Message] = []
+        fake_ai = self.mock_create_ai(self.args, self.config)
+        for question in questions:
+            expected_responses += fake_ai.request(question,
+                                                  Chat([]),
+                                                  self.args.num_answers,
+                                                  set(self.args.output_tags)).messages
+
+        # 2. === repeat all three questions (without overwriting) ===
+        self.args.ask = None
+        self.args.repeat = ['0001', '0002', '0003']
+        self.args.overwrite = False
+        question_cmd(self.args, self.config)
+        # two new files should be in the cache directory
+        # * the repeated cached message with answer
+        # * the repeated DB message
+        # -> the cached message without answer should be overwritten
+        self.assertEqual(len(self.message_list(self.cache_dir)), 4)
+        self.assertEqual(len(self.message_list(self.db_dir)), 1)
+        expected_cache_messages = [expected_responses[0], message2, expected_responses[1], expected_responses[2]]
+        chat = ChatDB.from_dir(Path(self.cache_dir.name),
+                               Path(self.db_dir.name))
+        cached_msg = chat.msg_gather(loc='cache')
+        print(f"Cached: {cached_msg}")
+        print(f"Expected: {expected_cache_messages}")
+        self.assert_messages_equal(cached_msg, expected_cache_messages)