README.md

@@ -0,0 +1,131 @@
+# greek_lang
+
+Проект для изучения греческого языка: перевод слов между языками (греческий, русский, английский) с помощью OpenAI, сохранение в базе данных (PostgreSQL) вместе с транскрипцией, описанием, примерами, категорией и аудио‑произношением (генерация через Google TTS). Также включает телеграм‑бота (aiogram) с хранением состояния в Redis.
+
+## Возможности
+- Перевод слова между языками ru/en/el (ISO‑639‑1) через OpenAI.
+- Автогенерация и сохранение:
+  - леммы, транскрипции (IPA), перевода, описания;
+  - части речи, семантической категории, примера и его перевода;
+  - краткой этимологии;
+  - аудио‑файла произношения (aiogTTS).
+- Сохранение результатов в PostgreSQL (SQLAlchemy + Alembic миграции).
+- Телеграм‑бот на aiogram с Redis FSM‑хранилищем.
+- Конфигурация через .env и pydantic‑settings; инициализация зависимостей через dependency_injector.
+
+## Требования
+- Python 3.13+
+- PostgreSQL 14+
+- Redis 6+
+- OpenAI API ключ
+
+Рекомендуется менеджер зависимостей uv и утилита go-task (для Taskfile.yml).
+
+## Установка
+1. Клонируйте репозиторий:
+   git clone <repo-url>
+   cd greek_lang
+
+2. Установите зависимости (через uv):
+   - Обновить/синхронизировать зависимости: uv sync
+   - Обновить lockfile: uv lock --upgrade && uv sync
+
+3. Создайте файл .env в корне проекта (см. пример ниже).
+
+## Настройка окружения (.env)
+Значения по умолчанию указаны в скобках. Переменные без значения обязательны к заполнению.
+
+# OpenAI
+API_KEY=sk-...
+
+# Telegram Bot
+TG_TOKEN=123456:ABC...
+
+# Логи (опционально: отправка ошибок в Telegram)
+LOG_TELEGRAM_BOT_TOKEN=123456:DEF...   # опционально
+LOG_TELEGRAM_CHAT_ID=123456789         # опционально
+
+# PostgreSQL
+DB_HOST=127.0.0.1
+DB_PORT=5432
+DB_NAME=greek_lang
+DB_USER=greek_lang
+DB_PASSWORD=greek_lang
+DB_POOL_SIZE=20
+DB_POOL_MAX_OVERFLOW=5
+DB_CONNECT_WAIT_TIMEOUT_SECONDS=5
+DB_DEBUG=false
+
+# Redis
+REDIS_HOST=127.0.0.1
+REDIS_PORT=6379
+REDIS_DB=0
+REDIS_USERNAME=
+REDIS_PASSWORD=
+REDIS_POOL_SIZE=100
+
+Примечания:
+- Переменные считываются классами конфигурации из src/greek_lang/configs/*.py.
+- Для Alembic миграций можно переопределить URL через переменную окружения ALEMBIC_DB_URL. Если не задано, Alembic сам соберёт URL из .env.
+
+## Инициализация базы и миграции
+Alembic уже сконфигурирован (см. src/greek_lang/database/migrations/ и alembic.ini).
+
+- Применить миграции к последней версии:
+  alembic -c src/greek_lang/database/alembic.ini upgrade head
+
+- Откатить на один шаг:
+  alembic -c src/greek_lang/database/alembic.ini downgrade -1
+
+- Использовать конкретный URL БД (в обход .env):
+  ALEMBIC_DB_URL="postgresql://user:pass@host:5432/dbname" \
+  alembic -c src/greek_lang/database/alembic.ini upgrade head
+
+## Использование
+### CLI переводчик
+Пример запуска CLI для перевода одного слова:
+
+python -m cli.translate "έμπορος" -s el -t ru
+
+- -s/--source: исходный язык (ru|en|el), по умолчанию el
+- -t/--target: язык перевода (ru|en|el), по умолчанию ru
+
+При запуске создаются контейнеры зависимостей, выполняется запрос к OpenAI, генерируется аудио‑произношение и всё сохраняется в таблицу glossary_word.
+
+### Телеграм‑бот
+Убедитесь, что Redis и .env настроены. Затем запустите бота:
+
+python -m greek_lang.tg_bot
+
+Бот удалит вебхук и начнёт polling. Команда /start ответит тестовым сообщением (заглушка). В коде предусмотрена интеграция Redis FSM‑хранилища и готовность к добавлению диалогов (aiogram-dialog).
+
+## Архитектура (кратко)
+- Контейнеры зависимостей: src/greek_lang/container.py и под‑контейнеры для конфигов, БД, OpenAI, Redis.
+- Конфигурация: pydantic‑settings, .env читается из корня проекта (src/greek_lang/configs/__init__.py).
+- Перевод: src/greek_lang/translator.py — orchestration: OpenAI → TTS → вставка/обновление в БД (UPSERT по term).
+- OpenAI: src/greek_lang/openai_manager — типизированный парсинг ответа через client.beta.chat.completions.parse в модель WordInfo.
+- БД: SQLAlchemy 2.0, Alembic миграции; сущности в glossaries/models.py и openai_manager/models.py.
+- Аудио: aiogTTS, генерация в BytesIO, хранение в столбце LargeBinary.
+- Логи: JSON‑логгер, опциональная отправка ошибок в Telegram.
+
+## Разработка
+- Линт/типы через go-task (см. Taskfile.yml):
+  - Запуск проверок: task check
+  - Только mypy: task mypy
+  - Ruff lint/format: task ruff-check / task ruff-fix / task ruff-format
+  - Обновить зависимости: task deps-update
+  - Очистить кэши: task clean
+
+- Установка dev‑зависимостей (uv):
+  uv sync --group dev
+
+Примечание: В pyproject.toml определён entry point [project.scripts], но пока не реализована функция greek_lang:main. Используйте вызовы модулей через python -m как указано выше.
+
+## Частые проблемы
+- Нет доступа к OpenAI: проверьте переменную API_KEY и сетевые ограничения.
+- Ошибка подключения к Postgres: проверьте DB_HOST/DB_PORT/DB_USER/DB_PASSWORD и что БД создана.
+- Alembic не видит URL: либо задайте ALEMBIC_DB_URL, либо убедитесь, что .env заполнен корректно.
+- Redis недоступен: проверьте REDIS_* переменные и доступность сервиса.
+
+## Лицензия
+Не указана. При необходимости добавьте раздел с выбранной лицензией.

Taskfile.yml

@@ -35,6 +35,12 @@ tasks:
       - task ruff-fix
       - task ruff-format
 
+  deps-update:
+    desc: "Update all dependencies (uv lock --upgrade && uv sync)"
+    cmds:
+      - uv lock --upgrade
+      - uv sync
+
   clean:
     desc: "Clean cache and temporary files"
     cmds:

checkers/check_openai.py

@@ -0,0 +1,19 @@
+import asyncio
+
+from greek_lang.container import init_main_container
+from greek_lang.openai_manager.manager import OpenAiManager
+
+
+async def main():
+    async with init_main_container() as container:
+        open_ai_manager: OpenAiManager = await container.openai_container().ai_manager()
+        source_lang = "el"
+        target_lang = "ru"
+        word_response = await open_ai_manager.get_gpt_response(
+            word="έμπορος", source_lang=source_lang, target_lang=target_lang
+        )
+        print(word_response)
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

cli/translate.py

@@ -0,0 +1,48 @@
+import asyncio
+import click
+
+from greek_lang.container import init_main_container
+from greek_lang.languages import LanguageEnum
+from greek_lang.translator import translate
+
+
+async def _translate(word: str, source_code: str, target_code: str) -> None:
+    try:
+        source_lang = LanguageEnum(source_code.lower())
+        target_lang = LanguageEnum(target_code.lower())
+    except KeyError as exc:
+        raise click.BadParameter(f"Unsupported language code: {exc.args[0]}") from exc
+
+    async with init_main_container():
+        result = await translate(word, source_lang, target_lang=target_lang)
+
+    click.echo(result)
+
+
+@click.command(help="Перевод слова между языками библиотеки greek_lang.")
+@click.argument("word")
+@click.option(
+    "-s",
+    "--source",
+    "source_code",
+    default="el",
+    show_default=True,
+    type=click.Choice([lang.name for lang in LanguageEnum], case_sensitive=False),
+    help="Код исходного языка (ISO-639-1, как в LanguageEnum).",
+)
+@click.option(
+    "-t",
+    "--target",
+    "target_code",
+    default="ru",
+    show_default=True,
+    type=click.Choice([lang.name for lang in LanguageEnum], case_sensitive=False),
+    help="Код языка перевода (ISO-639-1, как в LanguageEnum).",
+)
+def cli(word: str, source_code: str, target_code: str) -> None:
+    """Обёртка, которая запускает асинхронный перевод через asyncio."""
+    asyncio.run(_translate(word, source_code, target_code))
+
+
+if __name__ == "__main__":
+    cli()

pyproject.toml

@@ -8,17 +8,27 @@ authors = [
 ]
 requires-python = ">=3.13"
 dependencies = [
+    "aiogram>=3.21.0",
+    "aiogram-dialog>=2.4.0",
     "aiogtts>=1.1.1",
     "alembic>=1.16.1",
     "asyncpg>=0.30.0",
+    "click>=8.2.1",
     "dependency-injector>=4.47.1",
     "greenlet>=3.2.3",
+    "legacy-cgi>=2.6.3",
     "openai>=1.84.0",
+    "orjson>=3.11.1",
     "pendulum>=3.1.0",
     "psycopg2-binary>=2.9.10",
     "pydantic>=2.11.5",
     "pydantic-settings>=2.9.1",
+    "python-json-logger>=3.3.0",
+    "redis>=6.4.0",
+    "requests>=2.32.4",
+    "sentry-sdk>=2.34.1",
     "sqlalchemy>=2.0.41",
+    "telebot>=0.0.5",
 ]
 
 [project.scripts]
@@ -36,4 +46,5 @@ dev = [
     "pre-commit>=4.2.0",
     "pyupgrade>=3.20.0",
     "ruff>=0.11.13",
+    "types-requests>=2.32.4.20250809",
 ]

src/greek_lang/audio/manager.py

@@ -0,0 +1,19 @@
+from __future__ import annotations
+
+import io
+
+from aiogtts import aiogTTS  # type: ignore[import-untyped]
+
+from ..languages import LanguageEnum
+
+
+async def get_pronunciation(text: str, source_lang: LanguageEnum) -> io.BytesIO:
+    aiogtts = aiogTTS()
+    buffer = io.BytesIO()
+    await aiogtts.write_to_fp(
+        text=text,
+        fp=buffer,
+        slow=True,
+        lang=source_lang.value,
+    )
+    return buffer

src/greek_lang/configs/container.py

@@ -3,6 +3,7 @@ from dependency_injector import containers, providers
 from .db_config import PostgresConfig
 from .log_config import LoggerConfig
 from .openai_config import OpenAiConfig
+from .redis_conn import RedisConfig
 from .tg_bot_config import TgBotConfig
 
 
@@ -11,5 +12,6 @@ class ConfigContainer(containers.DeclarativeContainer):
     postgres_config: providers.Provider[PostgresConfig] = providers.Singleton(
         PostgresConfig
     )
+    redis_config: providers.Provider[RedisConfig] = providers.Singleton(RedisConfig)
     tg_bot_config: providers.Provider[TgBotConfig] = providers.Singleton(TgBotConfig)
     openai_config: providers.Provider[OpenAiConfig] = providers.Singleton(OpenAiConfig)

src/greek_lang/configs/log_config.py

@@ -1,4 +1,5 @@
 import pydantic
+from pydantic_settings import SettingsConfigDict
 
 from . import EnvConfig
 
@@ -6,3 +7,7 @@ from . import EnvConfig
 class LoggerConfig(EnvConfig):
     telegram_bot_token: pydantic.SecretStr | None = None
     telegram_chat_id: int | None = None
+
+    model_config = SettingsConfigDict(
+        env_prefix="LOG_",
+    )

src/greek_lang/configs/redis_conn.py

@@ -0,0 +1,17 @@
+import pydantic
+from pydantic_settings import SettingsConfigDict
+
+from . import EnvConfig
+
+
+class RedisConfig(EnvConfig):
+    host: str = pydantic.Field(default="127.0.0.1")
+    port: int = pydantic.Field(default=6379)
+    db: int = pydantic.Field(default=0)
+    username: str | None = pydantic.Field(default=None)
+    password: pydantic.SecretStr | None = pydantic.Field(default=None)
+    pool_size: int = pydantic.Field(default=100)
+
+    model_config = SettingsConfigDict(
+        env_prefix="REDIS_",
+    )

src/greek_lang/configs/tg_bot_config.py

@@ -1,7 +1,12 @@
 import pydantic
+from pydantic_settings import SettingsConfigDict
 
 from . import EnvConfig
 
 
 class TgBotConfig(EnvConfig):
+    model_config = SettingsConfigDict(
+        env_prefix="TG_",
+    )
+
     token: pydantic.SecretStr

src/greek_lang/container.py

@@ -6,6 +6,7 @@ from dependency_injector import containers, providers
 from .configs.container import ConfigContainer
 from .database.container import DatabaseContainer
 from .openai_manager.container import OpenAiContainer
+from .redis_db.container import RedisContainer
 
 
 class MainContainer(containers.DeclarativeContainer):
@@ -18,6 +19,9 @@ class MainContainer(containers.DeclarativeContainer):
     openai_container = providers.Container(
         OpenAiContainer, config_container=config_container
     )
+    redis_container = providers.Container(
+        RedisContainer, config_container=config_container
+    )
 
 
 @contextlib.asynccontextmanager

src/greek_lang/database/app_models.py

@@ -4,8 +4,12 @@ import types
 def get_app_models_modules() -> list[types.ModuleType]:
     from greek_lang.glossaries import models as glossaries_models
     from greek_lang.openai_manager import models as openai_manager_models
+    from greek_lang.users import models as users_models
+    from greek_lang.srs import models as srs_models
 
     return [
         glossaries_models,
         openai_manager_models,
+        users_models,
+        srs_models,
     ]

src/greek_lang/database/migrations/versions/20250621_2051_55f95da68641_.py

@@ -0,0 +1,36 @@
+"""empty message
+
+Revision ID: 55f95da68641
+Revises: 19fc4bee7a9f
+Create Date: 2025-06-21 20:51:15.097769
+
+"""
+
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+
+
+# revision identifiers, used by Alembic.
+revision: str = "55f95da68641"
+down_revision: Union[str, None] = "19fc4bee7a9f"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    """Upgrade schema."""
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.drop_column("glossary_word", "audio_file")
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    """Downgrade schema."""
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.add_column(
+        "glossary_word",
+        sa.Column("audio_file", sa.TEXT(), autoincrement=False, nullable=True),
+    )
+    # ### end Alembic commands ###

src/greek_lang/database/migrations/versions/20250621_2051_78357f437f61_.py

@@ -0,0 +1,35 @@
+"""empty message
+
+Revision ID: 78357f437f61
+Revises: 55f95da68641
+Create Date: 2025-06-21 20:51:29.437692
+
+"""
+
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+
+
+# revision identifiers, used by Alembic.
+revision: str = "78357f437f61"
+down_revision: Union[str, None] = "55f95da68641"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    """Upgrade schema."""
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.add_column(
+        "glossary_word", sa.Column("audio_file", sa.LargeBinary(), nullable=True)
+    )
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    """Downgrade schema."""
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.drop_column("glossary_word", "audio_file")
+    # ### end Alembic commands ###

src/greek_lang/database/migrations/versions/20250716_1013_6b43c7ed8c78_.py

@@ -0,0 +1,43 @@
+"""empty message
+
+Revision ID: 6b43c7ed8c78
+Revises: 78357f437f61
+Create Date: 2025-07-16 10:13:26.574794
+
+"""
+
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+
+
+# revision identifiers, used by Alembic.
+revision: str = "6b43c7ed8c78"
+down_revision: Union[str, None] = "78357f437f61"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    """Upgrade schema."""
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.alter_column(
+        "openai_token_usage",
+        "response_fingerprint",
+        existing_type=sa.TEXT(),
+        nullable=True,
+    )
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    """Downgrade schema."""
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.alter_column(
+        "openai_token_usage",
+        "response_fingerprint",
+        existing_type=sa.TEXT(),
+        nullable=False,
+    )
+    # ### end Alembic commands ###

src/greek_lang/database/migrations/versions/20250810_1240_d30d80dee5a3_.py

@@ -0,0 +1,34 @@
+"""empty message
+
+Revision ID: d30d80dee5a3
+Revises: 6b43c7ed8c78
+Create Date: 2025-08-10 12:40:24.118166
+
+"""
+
+from typing import Sequence, Union
+
+from alembic import op
+
+
+# revision identifiers, used by Alembic.
+revision: str = "d30d80dee5a3"
+down_revision: Union[str, None] = "6b43c7ed8c78"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    """Upgrade schema."""
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.create_unique_constraint(
+        op.f("uq_glossary_word_term"), "glossary_word", ["term"]
+    )
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    """Downgrade schema."""
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.drop_constraint(op.f("uq_glossary_word_term"), "glossary_word", type_="unique")
+    # ### end Alembic commands ###

src/greek_lang/database/migrations/versions/20250816_1753_747797032526_.py

@@ -0,0 +1,54 @@
+"""empty message
+
+Revision ID: 747797032526
+Revises: d30d80dee5a3
+Create Date: 2025-08-16 17:53:23.785592
+
+"""
+
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+
+
+# revision identifiers, used by Alembic.
+revision: str = "747797032526"
+down_revision: Union[str, None] = "d30d80dee5a3"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    """Upgrade schema."""
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.create_table(
+        "users",
+        sa.Column("id", sa.BigInteger(), nullable=False),
+        sa.Column("is_bot", sa.Boolean(), nullable=False),
+        sa.Column("first_name", sa.String(), nullable=True),
+        sa.Column("last_name", sa.String(), nullable=True),
+        sa.Column("username", sa.String(), nullable=True),
+        sa.Column("language_code", sa.String(length=8), nullable=True),
+        sa.Column("is_premium", sa.Boolean(), nullable=True),
+        sa.Column("added_to_attachment_menu", sa.Boolean(), nullable=True),
+        sa.Column(
+            "registered_at",
+            sa.DateTime(timezone=True),
+            server_default=sa.text("now()"),
+            nullable=False,
+        ),
+        sa.PrimaryKeyConstraint("id", name=op.f("pk_users")),
+    )
+    op.create_index(op.f("ix_users_id"), "users", ["id"], unique=False)
+    op.create_index(op.f("ix_users_username"), "users", ["username"], unique=False)
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    """Downgrade schema."""
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.drop_index(op.f("ix_users_username"), table_name="users")
+    op.drop_index(op.f("ix_users_id"), table_name="users")
+    op.drop_table("users")
+    # ### end Alembic commands ###

src/greek_lang/database/migrations/versions/20250816_1940_9a2898513cf2_.py

@@ -0,0 +1,112 @@
+"""empty message
+
+Revision ID: 9a2898513cf2
+Revises: 747797032526
+Create Date: 2025-08-16 19:40:06.376743
+
+"""
+
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+
+
+# revision identifiers, used by Alembic.
+revision: str = "9a2898513cf2"
+down_revision: Union[str, None] = "747797032526"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    """Upgrade schema."""
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.create_table(
+        "srs_progress",
+        sa.Column("id", sa.BigInteger(), nullable=False),
+        sa.Column("user_id", sa.BigInteger(), nullable=False),
+        sa.Column("word_id", sa.BigInteger(), nullable=False),
+        sa.Column(
+            "due_at",
+            sa.DateTime(timezone=True),
+            server_default=sa.text("now()"),
+            nullable=False,
+        ),
+        sa.Column("interval_days", sa.Integer(), nullable=False),
+        sa.Column("ease", sa.Float(), nullable=False),
+        sa.Column("reps", sa.Integer(), nullable=False),
+        sa.Column("lrn_step", sa.Integer(), nullable=False),
+        sa.Column(
+            "state",
+            sa.Enum(
+                "learning", "review", "lapsed", name="reviewstate", native_enum=False
+            ),
+            nullable=False,
+        ),
+        sa.ForeignKeyConstraint(
+            ["user_id"], ["users.id"], name=op.f("fk_srs_progress_user_id_users")
+        ),
+        sa.ForeignKeyConstraint(
+            ["word_id"],
+            ["glossary_word.id"],
+            name=op.f("fk_srs_progress_word_id_glossary_word"),
+        ),
+        sa.PrimaryKeyConstraint("id", name=op.f("pk_srs_progress")),
+        sa.UniqueConstraint("user_id", "word_id", name="uq_srs_user_word"),
+    )
+    op.create_index(
+        op.f("ix_srs_progress_due_at"), "srs_progress", ["due_at"], unique=False
+    )
+    op.create_index(
+        op.f("ix_srs_progress_user_id"), "srs_progress", ["user_id"], unique=False
+    )
+    op.create_index(
+        op.f("ix_srs_progress_word_id"), "srs_progress", ["word_id"], unique=False
+    )
+    op.create_table(
+        "srs_review_log",
+        sa.Column("id", sa.BigInteger(), nullable=False),
+        sa.Column("user_id", sa.BigInteger(), nullable=False),
+        sa.Column("word_id", sa.BigInteger(), nullable=False),
+        sa.Column(
+            "ts",
+            sa.DateTime(timezone=True),
+            server_default=sa.text("now()"),
+            nullable=False,
+        ),
+        sa.Column("grade", sa.Integer(), nullable=False),
+        sa.Column("prev_interval", sa.Integer(), nullable=False),
+        sa.Column("new_interval", sa.Integer(), nullable=False),
+        sa.Column("prev_ease", sa.Float(), nullable=False),
+        sa.Column("new_ease", sa.Float(), nullable=False),
+        sa.ForeignKeyConstraint(
+            ["user_id"], ["users.id"], name=op.f("fk_srs_review_log_user_id_users")
+        ),
+        sa.ForeignKeyConstraint(
+            ["word_id"],
+            ["glossary_word.id"],
+            name=op.f("fk_srs_review_log_word_id_glossary_word"),
+        ),
+        sa.PrimaryKeyConstraint("id", name=op.f("pk_srs_review_log")),
+    )
+    op.create_index(
+        op.f("ix_srs_review_log_user_id"), "srs_review_log", ["user_id"], unique=False
+    )
+    op.create_index(
+        op.f("ix_srs_review_log_word_id"), "srs_review_log", ["word_id"], unique=False
+    )
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    """Downgrade schema."""
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.drop_index(op.f("ix_srs_review_log_word_id"), table_name="srs_review_log")
+    op.drop_index(op.f("ix_srs_review_log_user_id"), table_name="srs_review_log")
+    op.drop_table("srs_review_log")
+    op.drop_index(op.f("ix_srs_progress_word_id"), table_name="srs_progress")
+    op.drop_index(op.f("ix_srs_progress_user_id"), table_name="srs_progress")
+    op.drop_index(op.f("ix_srs_progress_due_at"), table_name="srs_progress")
+    op.drop_table("srs_progress")
+    # ### end Alembic commands ###

src/greek_lang/glossaries/models.py

@@ -3,7 +3,7 @@ from __future__ import annotations
 import datetime
 import enum
 
-from sqlalchemy import BigInteger, Text, DateTime, Enum, func
+from sqlalchemy import BigInteger, Text, DateTime, Enum, func, LargeBinary
 from sqlalchemy.dialects.postgresql import ARRAY
 from sqlalchemy.orm import Mapped, mapped_column
 
@@ -11,7 +11,8 @@ from ..database.base import Base
 from ..languages import LanguageEnum
 
 
-class LexicalCategoryEnum(str, enum.Enum):
+@enum.unique
+class LexicalCategoryEnum(enum.StrEnum):
     noun = "noun"
     verb = "verb"
     adjective = "adjective"
@@ -37,6 +38,7 @@ class GlossaryWord(Base):
     term: Mapped[str] = mapped_column(
         Text(),
         nullable=False,
+        unique=True,
     )
     language: Mapped[LanguageEnum] = mapped_column(
         Enum(LanguageEnum, native_enum=False),
@@ -75,8 +77,8 @@ class GlossaryWord(Base):
         Text(),
         nullable=True,
     )
-    audio_file: Mapped[str | None] = mapped_column(
-        Text(),
+    audio_file: Mapped[bytes | None] = mapped_column(
+        LargeBinary(),
         nullable=True,
     )
     created_at: Mapped[datetime.datetime] = mapped_column(
@@ -97,7 +99,7 @@ class GlossaryWord(Base):
 
     def __repr__(self) -> str:
         return (
-            f"<GlossaryWord(id={self.id}, term='{self.term}', language='{self.language.value}', "
+            f"<GlossaryWord(id={self.id}, term='{self.term}', language='{self.language}', "
             f"translation='{self.translation}', transcription='{self.transcription}', "
             f"lexical_category='{self.lexical_category}', meaning_category='{self.meaning_category}')>"
         )

src/greek_lang/languages.py

@@ -2,7 +2,7 @@ import enum
 
 
 @enum.unique
-class LanguageEnum(str, enum.Enum):
+class LanguageEnum(enum.StrEnum):
     ru = "ru"
     en = "en"
     el = "el"

src/greek_lang/logger.py

@@ -0,0 +1,255 @@
+import base64
+import contextlib
+import contextvars
+import datetime
+import logging
+import socket
+import sys
+import typing
+from collections.abc import Iterator
+from logging.config import dictConfig
+from types import TracebackType
+
+import orjson
+import sentry_sdk
+
+from .configs.log_config import LoggerConfig
+
+
+extra_log_context: contextvars.ContextVar[dict[str, str]] = contextvars.ContextVar(
+    "extra_log_context"
+)
+
+
+@contextlib.contextmanager
+def extra_log_context_manager(new_context: dict[str, str]) -> Iterator[None]:
+    extra_log_context_data = {}
+    with contextlib.suppress(LookupError):
+        extra_log_context_data = extra_log_context.get()
+    extra_log_context_data.update(new_context)
+    token = extra_log_context.set(extra_log_context_data)
+    try:
+        yield
+    finally:
+        extra_log_context.reset(token)
+
+
+class NonLoggableExceptionsFilter(logging.Filter):
+    exclude_exception_types: typing.Sequence[type[Exception]]
+
+    def __init__(
+        self,
+        *,
+        exclude_exception_types: typing.Sequence[type[Exception]] = (),
+        name: str = "",
+    ):
+        self.exclude_exception_types = exclude_exception_types
+        super().__init__(name=name)
+
+    def filter(self, record: logging.LogRecord) -> bool:
+        if record.exc_info is None:
+            return True
+        try:
+            exception_type = record.exc_info[0]
+        except TypeError:
+            return True
+        return exception_type not in self.exclude_exception_types
+
+
+def default_json_serializer(obj: object) -> str:
+    match obj:
+        case bytes() as b:
+            try:
+                return b.decode("utf-8")
+            except UnicodeDecodeError:
+                return base64.b64encode(b).decode("ascii")
+        case datetime.timedelta() as td:
+            return str(td.total_seconds())
+        case datetime.datetime() as dt:
+            return dt.isoformat()
+        case datetime.date() as d:
+            return d.isoformat()
+        case _:
+            raise TypeError(f"Type {type(obj)} not serializable")
+
+
+def json_serializer(data: dict[str, typing.Any], **_: typing.Any) -> str:
+    extra_log_context_data = {}
+    with contextlib.suppress(LookupError):
+        extra_log_context_data = extra_log_context.get()
+    data.update({"extra_log_context": extra_log_context_data})
+    return orjson.dumps(
+        data,
+        default=default_json_serializer,
+    ).decode()
+
+
+def get_dict_config(
+    *,
+    sentry_dsn: str | None = None,
+    tg_token: str | None = None,
+    tg_chat: int | None = None,
+    exclude_exception_types: typing.Sequence[type[Exception]] = (),
+    formatters_extension_dict: dict[str, typing.Any] | None = None,
+    filters_extension_dict: dict[str, typing.Any] | None = None,
+    handlers_extension_dict: dict[str, typing.Any] | None = None,
+    loggers_extension_dict: dict[str, typing.Any] | None = None,
+) -> dict[str, typing.Any]:
+    hostname: str = socket.gethostname()
+    null_handler: dict[str, str] = {
+        "class": "logging.NullHandler",
+    }
+    formatters = {
+        "verbose": {
+            "format": f"%(asctime)s [%(levelname)s] [{hostname} %(name)s:%(lineno)s] %(message)s"
+        },
+        "json": {
+            "()": "pythonjsonlogger.jsonlogger.JsonFormatter",
+            "json_serializer": json_serializer,
+            "format": "%(asctime)s %(levelname)s %(name)s %(filename)s %(lineno)s %(message)s",
+        },
+    } | (formatters_extension_dict or {})
+    filters = {
+        "non_loggable_exceptions": {
+            "()": NonLoggableExceptionsFilter,
+            "exclude_exception_types": exclude_exception_types,
+        },
+    } | (filters_extension_dict or {})
+    handlers = {
+        "console_handler": {
+            "class": "logging.StreamHandler",
+            "formatter": "verbose",
+            "filters": [],
+        },
+        "telegram_handler": {
+            "class": "greek_lang.utils.telegram_log.handler.TelegramHandler",
+            "token": tg_token,
+            "chat_id": tg_chat,
+            "logger_name": "console_handler",
+            "level": "ERROR",
+            "formatter": "verbose",
+            "filters": ["non_loggable_exceptions"],
+        }
+        if tg_token and tg_token
+        else null_handler,
+        "sentry_handler": {
+            "class": "sentry_sdk.integrations.logging.EventHandler",
+            "level": "ERROR",
+            "formatter": "verbose",
+            "filters": ["non_loggable_exceptions"],
+        }
+        if sentry_dsn is not None
+        else null_handler,
+    } | (handlers_extension_dict or {})
+    loggers = {
+        "root": {
+            "level": "DEBUG",
+            "handlers": ["console_handler", "telegram_handler", "sentry_handler"],
+        },
+        "console": {
+            "level": "DEBUG",
+            "handlers": ["console_handler"],
+            "propagate": False,
+        },
+        "telegram.bot": {
+            "propagate": False,
+        },
+        "httpx": {
+            "level": "DEBUG",
+            "propagate": True,
+        },
+    } | (loggers_extension_dict or {})
+    return {
+        "version": 1,
+        "disable_existing_loggers": False,
+        "formatters": formatters,
+        "filters": filters,
+        "handlers": handlers,
+        "loggers": loggers,
+    }
+
+
+def create_tg_info_logger(
+    *,
+    tg_token: str,
+    tg_chat: str,
+) -> logging.Logger:
+    logger_name = "tg_info"
+    dict_config = {
+        "version": 1,
+        "disable_existing_loggers": False,
+        "handlers": {
+            "telegram_handler": {
+                "class": "petuh_bot.utils.telegram_log.handler.TelegramHandler",
+                "logger_name": "console",
+                "token": tg_token,
+                "chat_id": tg_chat,
+                "level": "INFO",
+            }
+        },
+        "loggers": {
+            logger_name: {
+                "handlers": ["telegram_handler"],
+                "propagate": False,
+            },
+        },
+    }
+    logging.config.dictConfig(dict_config)
+    return logging.getLogger(logger_name)
+
+
+def init_root_logger(
+    sentry_dsn: str | None = None,
+    tg_token: str | None = None,
+    tg_chat: int | None = None,
+    exclude_exception_types: typing.Sequence[type[Exception]] = (),
+    formatters_extension_dict: dict[str, typing.Any] | None = None,
+    filters_extension_dict: dict[str, typing.Any] | None = None,
+    handlers_extension_dict: dict[str, typing.Any] | None = None,
+    loggers_extension_dict: dict[str, typing.Any] | None = None,
+) -> logging.Logger:
+    if sentry_dsn is not None:
+        sentry_sdk.init(
+            dsn=sentry_dsn,
+            traces_sample_rate=1.0,
+            default_integrations=True,
+        )
+    dict_config = get_dict_config(
+        sentry_dsn=sentry_dsn,
+        tg_token=tg_token,
+        tg_chat=tg_chat,
+        exclude_exception_types=exclude_exception_types,
+        formatters_extension_dict=formatters_extension_dict,
+        filters_extension_dict=filters_extension_dict,
+        handlers_extension_dict=handlers_extension_dict,
+        loggers_extension_dict=loggers_extension_dict,
+    )
+    dictConfig(dict_config)
+    return logging.getLogger()
+
+
+loggers_ext: dict[str, typing.Any] = {}
+
+
+def _exc_hook_patched(
+    exc_type: type[BaseException],
+    exc_val: BaseException,
+    exc_tb: TracebackType,
+) -> None:
+    if isinstance(exc_val, KeyboardInterrupt):
+        return
+    logging.critical(
+        f"Uncaught exception: {exc_type}", exc_info=(exc_type, exc_val, exc_tb)
+    )
+
+
+def setup() -> None:
+    config = LoggerConfig()
+    init_root_logger(
+        tg_token=config.telegram_bot_token.get_secret_value()
+        if config.telegram_bot_token
+        else None,
+        tg_chat=config.telegram_chat_id,
+        loggers_extension_dict=loggers_ext,
+    )
+    sys.excepthook = _exc_hook_patched  # type: ignore[assignment]

src/greek_lang/openai_manager/manager.py

@@ -5,15 +5,23 @@ import dataclasses
 import pydantic
 from openai import AsyncOpenAI
 
+from greek_lang.languages import LanguageEnum
+from greek_lang.glossaries.models import LexicalCategoryEnum
+
 
 class WordInfo(pydantic.BaseModel):
+    lemma: str = pydantic.Field(
+        ...,
+        description="lemma (base form) - for verbs, use the 1st person singular in present indicative, "
+        "for nouns and adjectives, use the nominative singular masculine (for adjectives)",
+    )
     transcription: str = pydantic.Field(
         ...,
-        description="phonetic transcription in IPA",
+        description="lemma phonetic transcription in IPA",
     )
     translation: str = pydantic.Field(
         ...,
-        description="translation in {target_language}",
+        description="lemma translation in {target_language}",
     )
     description: str = pydantic.Field(
         ...,
@@ -21,19 +29,19 @@ class WordInfo(pydantic.BaseModel):
     )
     part_of_speech: str = pydantic.Field(
         ...,
-        description="part of speech in {target_language}",
+        description=f"part of speech, one of {[cat.value for cat in LexicalCategoryEnum]}",
     )
     example: str = pydantic.Field(
         ...,
-        description="example",
+        description="lemma example",
     )
     example_transcription: str = pydantic.Field(
         ...,
-        description="phonetic transcription in IPA of an example",
+        description="lemma phonetic transcription in IPA of an example",
     )
     example_translation: str = pydantic.Field(
         ...,
-        description="translation of the example in {target_language}",
+        description="lemma translation of the example in {target_language}",
     )
     category: str = pydantic.Field(
         ...,
@@ -53,8 +61,8 @@ class OpenAiManager:
         self,
         *,
         word: str,
-        source_lang: str,
-        target_lang: str,
+        source_lang: LanguageEnum,
+        target_lang: LanguageEnum,
         model: str = "gpt-4o",
     ) -> WordInfo:
         system_message = {
@@ -63,7 +71,7 @@ class OpenAiManager:
         }
         user_message = {
             "role": "user",
-            "content": f'Provide detailed information about the word "{word}" in language {source_lang}, set {{target_language}} = {target_lang}.',
+            "content": f'Provide detailed information about the word "{word}" in language {source_lang!s}, set {{target_language}} = {target_lang!s}.',
         }
         response = await self.client.beta.chat.completions.parse(
             model=model,

src/greek_lang/openai_manager/models.py

@@ -40,7 +40,7 @@ class OpenAiTokenUsage(Base):
     )
     response_fingerprint: Mapped[str] = mapped_column(
         Text(),
-        nullable=False,
+        nullable=True,
         index=True,
     )
     completion_tokens: Mapped[int] = mapped_column(

src/greek_lang/redis_db/container.py

@@ -0,0 +1,24 @@
+from collections.abc import AsyncIterator
+
+from dependency_injector import containers, providers
+
+from ..configs.redis_conn import RedisConfig
+from .redis_conn import create_redis_pool, RedisPool
+
+
+async def create_redis_pool_resource(
+    redis_config: RedisConfig,
+) -> AsyncIterator[RedisPool]:
+    redis_pool = await create_redis_pool(redis_config)
+    try:
+        yield redis_pool
+    finally:
+        await redis_pool.aclose()
+
+
+class RedisContainer(containers.DeclarativeContainer):
+    config_container = providers.DependenciesContainer()
+    redis_pool: providers.Resource[RedisPool] = providers.Resource(
+        create_redis_pool_resource,
+        redis_config=config_container.redis_config,
+    )

src/greek_lang/redis_db/redis_conn.py

@@ -0,0 +1,51 @@
+import dataclasses
+from typing import TypeAlias
+
+import redis
+
+from ..configs.redis_conn import RedisConfig
+
+
+RedisPool: TypeAlias = redis.asyncio.Redis
+
+
+@dataclasses.dataclass(frozen=True)
+class RedisConnectionParams:
+    host: str = "127.0.0.1"
+    port: int = 6379
+    db: int = 0
+    username: str | None = None
+    password: str | None = None
+    max_connections: int = 2**31
+    socket_timeout: float = 5.0
+
+
+def create_redis_single_pool(
+    redis_conn_params: RedisConnectionParams,
+) -> redis.asyncio.Redis:
+    redis_url = f"redis://{redis_conn_params.host}:{redis_conn_params.port}/{redis_conn_params.db}"
+    connection: redis.asyncio.Redis = redis.asyncio.from_url(  # type: ignore[no-untyped-call]
+        redis_url,
+        username=redis_conn_params.username,
+        password=redis_conn_params.password,
+        decode_responses=False,
+        socket_connect_timeout=redis_conn_params.socket_timeout,
+        max_connections=redis_conn_params.max_connections,
+    )
+    return connection
+
+
+def create_redis_pool(
+    redis_config: RedisConfig,
+) -> RedisPool:
+    redis_conn_params = RedisConnectionParams(
+        host=redis_config.host,
+        port=redis_config.port,
+        db=redis_config.db,
+        username=redis_config.username,
+        password=(
+            redis_config.password.get_secret_value() if redis_config.password else None
+        ),
+        max_connections=redis_config.pool_size,
+    )
+    return create_redis_single_pool(redis_conn_params)

src/greek_lang/srs/engine.py

@@ -0,0 +1,24 @@
+from datetime import datetime, timedelta
+
+from .models import UserWordProgress, ReviewState
+
+
+def sm2_update(p: UserWordProgress, grade: int, now: datetime) -> None:
+    # grade: 0..5
+    e = p.ease + (0.1 - (5 - grade) * (0.08 + (5 - grade) * 0.02))
+    p.ease = max(1.3, e)
+    if grade < 3:
+        p.interval_days = 1
+        p.reps = 0
+        p.lrn_step = 0
+        p.state = ReviewState.lapsed
+    else:
+        if p.reps == 0:
+            p.interval_days = 1
+        elif p.reps == 1:
+            p.interval_days = 6
+        else:
+            p.interval_days = round(p.interval_days * p.ease)
+        p.reps += 1
+        p.state = ReviewState.review
+    p.due_at = now + timedelta(days=p.interval_days)

src/greek_lang/srs/models.py

@@ -0,0 +1,67 @@
+import enum
+import datetime as dt
+
+from sqlalchemy import (
+    BigInteger,
+    Integer,
+    Float,
+    DateTime,
+    Enum,
+    ForeignKey,
+    UniqueConstraint,
+    func,
+)
+from sqlalchemy.orm import Mapped, mapped_column
+
+from greek_lang.database.base import Base
+
+
+@enum.unique
+class ReviewState(enum.StrEnum):
+    learning = "learning"
+    review = "review"
+    lapsed = "lapsed"
+
+
+class UserWordProgress(Base):
+    __tablename__ = "srs_progress"
+    id: Mapped[int] = mapped_column(BigInteger, primary_key=True)
+    user_id: Mapped[int] = mapped_column(
+        ForeignKey("users.id"), index=True, nullable=False
+    )
+    word_id: Mapped[int] = mapped_column(
+        ForeignKey("glossary_word.id"), index=True, nullable=False
+    )
+    due_at: Mapped[dt.datetime] = mapped_column(
+        DateTime(timezone=True), index=True, server_default=func.now(), nullable=False
+    )
+    interval_days: Mapped[int] = mapped_column(Integer, nullable=False, default=1)
+    ease: Mapped[float] = mapped_column(Float, nullable=False, default=2.5)
+    reps: Mapped[int] = mapped_column(Integer, nullable=False, default=0)
+    lrn_step: Mapped[int] = mapped_column(Integer, nullable=False, default=0)
+    state: Mapped[ReviewState] = mapped_column(
+        Enum(ReviewState, native_enum=False),
+        nullable=False,
+        default=ReviewState.learning,
+    )
+
+    __table_args__ = (UniqueConstraint("user_id", "word_id", name="uq_srs_user_word"),)
+
+
+class ReviewLog(Base):
+    __tablename__ = "srs_review_log"
+    id: Mapped[int] = mapped_column(BigInteger, primary_key=True)
+    user_id: Mapped[int] = mapped_column(
+        ForeignKey("users.id"), index=True, nullable=False
+    )
+    word_id: Mapped[int] = mapped_column(
+        ForeignKey("glossary_word.id"), index=True, nullable=False
+    )
+    ts: Mapped[dt.datetime] = mapped_column(
+        DateTime(timezone=True), server_default=func.now(), nullable=False
+    )
+    grade: Mapped[int] = mapped_column(Integer, nullable=False)  # 0..5
+    prev_interval: Mapped[int] = mapped_column(Integer, nullable=False)
+    new_interval: Mapped[int] = mapped_column(Integer, nullable=False)
+    prev_ease: Mapped[float] = mapped_column(Float, nullable=False)
+    new_ease: Mapped[float] = mapped_column(Float, nullable=False)

src/greek_lang/srs/selector.py

@@ -0,0 +1,28 @@
+from collections.abc import Sequence
+
+from sqlalchemy import select, func
+from sqlalchemy.ext.asyncio import AsyncSession
+
+from greek_lang.glossaries.models import GlossaryWord
+
+from .models import UserWordProgress
+
+
+async def pick_due_words(
+    db: AsyncSession,
+    user_id: int,
+    limit: int = 10,
+) -> Sequence[tuple[GlossaryWord, UserWordProgress]]:
+    stmt = (
+        select(GlossaryWord, UserWordProgress)
+        .join(UserWordProgress, UserWordProgress.word_id == GlossaryWord.id)
+        .where(
+            UserWordProgress.user_id == user_id,
+            UserWordProgress.due_at <= func.now(),
+        )
+        .order_by(UserWordProgress.due_at)
+        .limit(limit)
+    )
+
+    result = await db.execute(stmt)
+    return result.tuples().all()

src/greek_lang/tg_bot/__main__.py

@@ -0,0 +1,17 @@
+import asyncio
+
+from greek_lang import logger
+from greek_lang.container import init_main_container
+from greek_lang.tg_bot import app
+
+
+async def main() -> None:
+    logger.setup()
+    async with init_main_container():
+        bot = app.create_bot()
+        dispatcher = await app.create_dispatcher()
+        await app.run_bot(bot=bot, dispatcher=dispatcher)
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

src/greek_lang/tg_bot/app.py

@@ -0,0 +1,69 @@
+import pydantic
+from aiogram import Bot, Dispatcher, BaseMiddleware
+from aiogram.fsm.storage.base import BaseStorage, DefaultKeyBuilder
+from aiogram.fsm.storage.redis import RedisStorage
+from aiogram.types import BotCommandScopeAllPrivateChats
+from aiogram_dialog import setup_dialogs
+from dependency_injector.wiring import Provide, inject
+
+from ..configs.container import ConfigContainer
+from ..redis_db.container import RedisContainer
+from ..redis_db.redis_conn import RedisPool
+
+
+@inject
+def create_bot(
+    bot_token: pydantic.SecretStr = Provide[
+        ConfigContainer.tg_bot_config.provided.token
+    ],
+) -> Bot:
+    bot = Bot(
+        token=bot_token.get_secret_value(),
+    )
+    return bot
+
+
+async def create_dispatcher() -> Dispatcher:
+    from .router import router as root_router
+    from .dialogs import dialog
+
+    fsm_storage = await create_fsm_storage()
+
+    dp = Dispatcher(
+        storage=fsm_storage,
+    )
+
+    middlewares: list[BaseMiddleware] = []
+    for middleware in middlewares:
+        dp.update.middleware(middleware)
+
+    dp.include_routers(dialog, root_router)
+
+    setup_dialogs(dp)
+
+    return dp
+
+
+@inject
+async def create_fsm_storage(
+    redis_pool: RedisPool = Provide[RedisContainer.redis_pool],
+) -> BaseStorage:
+    storage = RedisStorage(
+        redis=redis_pool,
+        key_builder=DefaultKeyBuilder(
+            prefix="fsm",
+            with_destiny=True,
+        ),
+    )
+    return storage
+
+
+async def run_bot(
+    bot: Bot,
+    dispatcher: Dispatcher,
+) -> None:
+    await bot.delete_webhook(drop_pending_updates=True)
+    await bot.delete_my_commands(scope=BotCommandScopeAllPrivateChats())
+    await dispatcher.start_polling(
+        bot, allowed_updates=dispatcher.resolve_used_update_types()
+    )

src/greek_lang/tg_bot/commands.py

@@ -0,0 +1,19 @@
+from aiogram import Router
+from aiogram.filters import CommandStart
+from aiogram.types import Message
+from aiogram_dialog import DialogManager, StartMode
+
+from greek_lang.tg_bot.dialogs.states import States
+from greek_lang.users.manager import get_or_create_telegram_user
+
+
+router = Router()
+
+
+@router.message(CommandStart())
+async def start(message: Message, dialog_manager: DialogManager) -> None:
+    user = message.from_user
+    if user is None:
+        return
+    await get_or_create_telegram_user(user)
+    await dialog_manager.start(States.main_menu, mode=StartMode.RESET_STACK)

src/greek_lang/tg_bot/dialogs/__init__.py

@@ -0,0 +1,11 @@
+from aiogram_dialog import Dialog
+
+from .main_menu import windows as main_windows
+from .add_word import windows as add_word_windows
+
+
+dialog = Dialog(
+    main_windows.main_window,
+    add_word_windows.add_word_window,
+    add_word_windows.add_word_result_window,
+)

src/greek_lang/tg_bot/dialogs/add_word/handlers.py

@@ -0,0 +1,65 @@
+from aiogram.types import Message, CallbackQuery, BufferedInputFile
+from aiogram_dialog import DialogManager
+from aiogram_dialog.widgets.input import MessageInput
+from aiogram_dialog.widgets.kbd import Button
+from greek_lang.languages import LanguageEnum
+from greek_lang.translator import translate
+from ..states import States
+
+
+async def add_word(
+    message: Message,
+    source: MessageInput | Button,
+    manager: DialogManager,
+) -> None:
+    if not message.text:
+        return
+    word = message.text.strip()
+    if not word:
+        return
+    source_lang = LanguageEnum.ru
+    target_lang = LanguageEnum.el
+    glossary_word = await translate(word, source_lang, target_lang=target_lang)
+
+    # Try to send audio pronunciation back to the user
+    try:
+        audio_bytes = getattr(glossary_word, "audio_file", None)
+        if audio_bytes:
+            # aiogTTS produces MP3 data; send as audio
+            caption = (
+                f"<b>{glossary_word.term}</b> → <b>{glossary_word.translation}</b>"
+            )
+            input_file = BufferedInputFile(
+                audio_bytes, filename=f"{glossary_word.term}.mp3"
+            )
+            await message.answer_audio(
+                audio=input_file, caption=caption, parse_mode="HTML"
+            )
+    except Exception:
+        # Silently ignore audio sending issues to not break the flow
+        pass
+
+    # Store data for the result window
+    manager.dialog_data.update(
+        {
+            "term": glossary_word.term,
+            "translation": glossary_word.translation,
+            "transcription": glossary_word.transcription or "",
+            "lexical_category": getattr(glossary_word, "lexical_category", ""),
+            "description": glossary_word.description or "",
+            "example": glossary_word.example or "",
+            "note": glossary_word.note or "",
+        }
+    )
+
+    # Switch to the result window state
+    await manager.switch_to(States.add_word_result)
+
+
+async def on_add_another(
+    callback: CallbackQuery,
+    button: Button,
+    manager: DialogManager,
+) -> None:
+    await callback.answer()
+    await manager.switch_to(States.add_word)

src/greek_lang/tg_bot/dialogs/add_word/windows.py

@@ -0,0 +1,73 @@
+from typing import Any
+
+from aiogram.enums import ParseMode
+from aiogram_dialog import DialogManager, Window
+from aiogram_dialog.widgets.input import MessageInput
+from aiogram_dialog.widgets.kbd import Button, Row
+from aiogram_dialog.widgets.markup.reply_keyboard import ReplyKeyboardFactory
+from aiogram_dialog.widgets.text import Format, Const
+
+from ..states import States
+from ..base_handlers import cancel_handler
+from . import handlers
+
+
+async def add_word_window_getter(
+    dialog_manager: DialogManager, **kwargs: Any
+) -> dict[str, Any]:
+    return {}
+
+
+add_word_window = Window(
+    Const("<b>Введите слово</b>:"),
+    MessageInput(func=handlers.add_word),
+    Row(
+        Button(
+            Format("❌ Отмена"),
+            on_click=cancel_handler,  # type: ignore[arg-type]
+            id="cancel_add_word",
+        ),
+    ),
+    markup_factory=ReplyKeyboardFactory(
+        input_field_placeholder=Format("Слово..."),
+        resize_keyboard=True,
+        one_time_keyboard=True,
+    ),
+    state=States.add_word,
+    parse_mode=ParseMode.HTML,
+    getter=add_word_window_getter,
+)
+
+
+async def add_word_result_getter(
+    dialog_manager: DialogManager, **kwargs: Any
+) -> dict[str, Any]:
+    # Expose dialog_data fields directly for Format widgets
+    return dialog_manager.dialog_data
+
+
+add_word_result_window = Window(
+    Format(
+        "✅ <b>Слово добавлено</b>!\n\n"
+        "<b>{term}</b> → <b>{translation}</b>\n"
+        "{transcription}\n"
+        "<b>Часть речи</b>: {lexical_category!s}\n"
+        "<b>Описание</b>: {description}\n"
+        "<b>Пример</b>: {example}"
+    ),
+    Row(
+        Button(
+            Const("➕ Добавить ещё"),
+            id="add_another",
+            on_click=handlers.on_add_another,
+        ),
+        Button(
+            Const("🏠 В меню"),
+            on_click=cancel_handler,  # type: ignore[arg-type]
+            id="to_main_menu",
+        ),
+    ),
+    state=States.add_word_result,
+    parse_mode=ParseMode.HTML,
+    getter=add_word_result_getter,
+)

src/greek_lang/tg_bot/dialogs/base_handlers.py

@@ -0,0 +1,16 @@
+from aiogram_dialog import DialogManager, ShowMode
+from aiogram_dialog.api.internal import ReplyCallbackQuery
+from aiogram_dialog.widgets.kbd import Cancel
+
+from .states import States
+
+
+async def cancel_handler(
+    callback_query: ReplyCallbackQuery,
+    button: Cancel,
+    manager: DialogManager,
+) -> None:
+    await manager.switch_to(
+        States.main_menu,
+        show_mode=ShowMode.DELETE_AND_SEND,
+    )

src/greek_lang/tg_bot/dialogs/main_menu/handlers.py

@@ -0,0 +1,26 @@
+from aiogram.types import CallbackQuery
+from aiogram_dialog import DialogManager, ShowMode
+from aiogram_dialog.api.internal import ReplyCallbackQuery
+from aiogram_dialog.widgets.kbd import Cancel, Button
+
+from ..states import States
+
+
+async def on_add_word(
+    callback: CallbackQuery,
+    button: Button,
+    manager: DialogManager,
+) -> None:
+    await callback.answer()
+    await manager.switch_to(States.add_word)
+
+
+async def cancel_handler(
+    callback_query: ReplyCallbackQuery,
+    button: Cancel,
+    manager: DialogManager,
+) -> None:
+    await manager.switch_to(
+        States.add_word,
+        show_mode=ShowMode.DELETE_AND_SEND,
+    )

src/greek_lang/tg_bot/dialogs/main_menu/windows.py

@@ -0,0 +1,32 @@
+from typing import Any
+
+from aiogram.enums import ParseMode
+from aiogram_dialog import DialogManager, Window
+from aiogram_dialog.widgets.kbd import Row, Button
+from aiogram_dialog.widgets.text import Format, Const
+
+from ..states import States
+from . import handlers
+
+
+async def main_getter(dialog_manager: DialogManager, **kwargs: Any) -> dict[str, Any]:
+    return {}
+
+
+main_window = Window(
+    Format(
+        "<b>Выбери действие</b>:",
+        when=lambda data, widget, dialog_manager: data["dialog_data"].get("action")
+        is None,
+    ),
+    Row(
+        Button(
+            Const("Добавить слово"),
+            id="add_word",
+            on_click=handlers.on_add_word,
+        ),
+    ),
+    state=States.main_menu,
+    getter=main_getter,
+    parse_mode=ParseMode.HTML,
+)

src/greek_lang/tg_bot/dialogs/states.py

@@ -0,0 +1,7 @@
+from aiogram.fsm.state import State, StatesGroup
+
+
+class States(StatesGroup):
+    main_menu = State()
+    add_word = State()
+    add_word_result = State()

src/greek_lang/tg_bot/router.py

@@ -0,0 +1,8 @@
+from aiogram import Router
+
+from .commands import router as commands_router
+
+router = Router()
+router.include_routers(
+    commands_router,
+)

src/greek_lang/translator.py

@@ -0,0 +1,73 @@
+from dependency_injector.wiring import inject, Provide
+from sqlalchemy.ext.asyncio import async_sessionmaker, AsyncSession
+from sqlalchemy import func
+from sqlalchemy.dialects.postgresql import insert
+
+from greek_lang.audio.manager import get_pronunciation
+from greek_lang.database.container import DatabaseContainer
+from greek_lang.languages import LanguageEnum
+from greek_lang.openai_manager.container import OpenAiContainer
+from greek_lang.openai_manager.manager import OpenAiManager
+from greek_lang.glossaries.models import GlossaryWord, LexicalCategoryEnum
+
+
+@inject
+async def translate(
+    word: str,
+    source_lang: LanguageEnum,
+    target_lang: LanguageEnum = LanguageEnum.ru,
+    note: str | None = None,
+    tags: tuple[str, ...] = tuple(),
+    open_ai_manager: OpenAiManager = Provide[OpenAiContainer.ai_manager],
+    db_session_maker: async_sessionmaker[AsyncSession] = Provide[
+        DatabaseContainer.async_session_maker,
+    ],
+) -> GlossaryWord:
+    word_response = await open_ai_manager.get_gpt_response(
+        word=word,
+        source_lang=source_lang,
+        target_lang=target_lang,
+    )
+    pronon = await get_pronunciation(text=word_response.lemma, source_lang=source_lang)
+
+    async with db_session_maker() as db_session, db_session.begin():
+        values = {
+            "term": word_response.lemma,
+            "language": source_lang,
+            "transcription": word_response.transcription,
+            "translation": word_response.translation,
+            "description": word_response.description,
+            "lexical_category": LexicalCategoryEnum(word_response.part_of_speech),
+            "meaning_category": word_response.category,
+            "example": f"{word_response.example}({word_response.example_translation})",
+            "etymology": word_response.etymology,
+            "note": note,
+            "tags": list(tags),
+            "audio_file": pronon.getvalue(),
+        }
+        stmt = (
+            insert(GlossaryWord)
+            .values(**values)
+            .on_conflict_do_update(
+                index_elements=[GlossaryWord.term],
+                set_={
+                    "term": values["term"],
+                    "language": values["language"],
+                    "transcription": values["transcription"],
+                    "translation": values["translation"],
+                    "description": values["description"],
+                    "lexical_category": values["lexical_category"],
+                    "meaning_category": values["meaning_category"],
+                    "example": values["example"],
+                    "etymology": values["etymology"],
+                    "note": values["note"],
+                    "tags": values["tags"],
+                    "audio_file": values["audio_file"],
+                    "updated_at": func.now(),
+                },
+            )
+            .returning(GlossaryWord)
+        )
+        result = await db_session.execute(stmt)
+        glossary_word = result.scalar_one()
+    return glossary_word

src/greek_lang/users/manager.py

@@ -0,0 +1,39 @@
+from aiogram.types import User
+from dependency_injector.wiring import Provide, inject
+from sqlalchemy.exc import IntegrityError
+from sqlalchemy.ext.asyncio import AsyncSession, async_sessionmaker
+
+from greek_lang.database.container import DatabaseContainer
+from greek_lang.users.models import User as TgUser
+
+
+@inject
+async def get_or_create_telegram_user(
+    user: User,
+    db_session_maker: async_sessionmaker[AsyncSession] = Provide[
+        DatabaseContainer.async_session_maker
+    ],
+) -> TgUser:
+    async with db_session_maker() as db_session, db_session.begin():
+        telegram_user: TgUser | None = await db_session.get(TgUser, user.id)
+        if telegram_user:
+            return telegram_user
+    try:
+        async with db_session_maker() as db_session, db_session.begin():
+            telegram_user = TgUser(
+                id=user.id,
+                username=user.username,
+                first_name=user.first_name,
+                last_name=user.last_name,
+                language_code=user.language_code,
+                is_bot=user.is_bot,
+                is_premium=user.is_premium,
+                added_to_attachment_menu=user.added_to_attachment_menu,
+            )
+            db_session.add(telegram_user)
+        return telegram_user
+    except IntegrityError:
+        telegram_user = await db_session.get(TgUser, user.id)
+        if telegram_user is None:
+            raise Exception(f"Can't find telegram_user = {user.id}") from None
+        return telegram_user

src/greek_lang/users/models.py

@@ -0,0 +1,26 @@
+from datetime import datetime
+
+from sqlalchemy import BigInteger, Boolean, DateTime, String, func
+from sqlalchemy.orm import Mapped, mapped_column
+
+from greek_lang.database.base import Base
+
+
+class User(Base):
+    __tablename__ = "users"
+
+    id: Mapped[int] = mapped_column(BigInteger, primary_key=True, index=True)
+    is_bot: Mapped[bool] = mapped_column(Boolean, nullable=False)
+    first_name: Mapped[str | None] = mapped_column(String, nullable=True)
+    last_name: Mapped[str | None] = mapped_column(String, nullable=True)
+    username: Mapped[str | None] = mapped_column(String, nullable=True, index=True)
+    language_code: Mapped[str | None] = mapped_column(String(length=8), nullable=True)
+    is_premium: Mapped[bool | None] = mapped_column(Boolean, nullable=True)
+    added_to_attachment_menu: Mapped[bool | None] = mapped_column(
+        Boolean, nullable=True
+    )
+    registered_at: Mapped[datetime] = mapped_column(
+        DateTime(timezone=True),
+        server_default=func.now(),
+        nullable=False,
+    )

src/greek_lang/utils/telegram_log/handler.py

@@ -0,0 +1,103 @@
+import contextlib
+import copy
+import datetime
+import io
+import logging
+import logging.config
+import types
+
+import cgitb  # type: ignore[import-untyped]
+import telebot
+from requests import ReadTimeout
+
+
+ExcInfoType = tuple[type[BaseException], BaseException, types.TracebackType]
+
+
+class TelegramHandler(logging.Handler):
+    bot: telebot.TeleBot
+    chat_id: int
+    logger_name: logging.Logger | None
+
+    def __init__(self, *, token: str, chat_id: int, logger_name: str | None = None):
+        logging.Handler.__init__(self)
+        self.bot = telebot.TeleBot(token)
+        self.chat_id = chat_id
+        self.logger = logging.getLogger(logger_name) if logger_name else None
+
+    @staticmethod
+    def get_tb_data(exc_info: ExcInfoType, output_format: str = "html") -> io.BytesIO:
+        string_io_buffer = io.StringIO()
+        context_width = 11
+        cgitb.Hook(
+            context=context_width,
+            file=string_io_buffer,
+            format=output_format,
+        ).handle(info=exc_info)
+        string_io_buffer.seek(0)
+        encoding = "utf-8"
+        bytes_io_buffer = io.BytesIO(string_io_buffer.read().encode(encoding))
+        bytes_io_buffer.seek(0)
+        return bytes_io_buffer
+
+    @staticmethod
+    def prepare(log_data: str, length: int) -> str:
+        message = log_data[:length]
+        return message
+
+    def emit(self, record: logging.LogRecord) -> None:
+        try:
+            if record.exc_info is None:
+                self.send_plain_text(record)
+            else:
+                self.send_traceback(record)
+        except ReadTimeout:
+            if self.logger:
+                self.logger.error("Telegram request timed out")
+        except BaseException as exc:
+            if self.logger:
+                self.logger.exception(
+                    f"Telegram Log Handler Unexpected Exception Occurred: {exc}"
+                )
+
+    def send_traceback(self, record: logging.LogRecord) -> None:
+        tb_data_html = self.get_tb_data(record.exc_info, output_format="html")  # type: ignore
+        tb_data_plain = self.get_tb_data(record.exc_info, output_format="plain")  # type: ignore
+        with contextlib.closing(tb_data_html), contextlib.closing(tb_data_plain):
+            filename = datetime.datetime.now().strftime("python_tb_%Y-%m-%d_%H_%M_%S")
+            caption = self.get_exc_caption_text(record)
+            self.bot.send_media_group(
+                chat_id=self.chat_id,
+                media=[
+                    telebot.types.InputMediaDocument(
+                        telebot.types.InputFile(
+                            tb_data_html, file_name=filename + ".html"
+                        ),
+                        caption=caption,
+                    ),
+                    telebot.types.InputMediaDocument(
+                        telebot.types.InputFile(
+                            tb_data_plain, file_name=filename + ".txt"
+                        )
+                    ),
+                ],
+                timeout=5,
+            )
+
+    def get_exc_caption_text(self, record: logging.LogRecord) -> str:
+        caption_length = 200
+        no_exc_record = self.get_no_exc_record_copy(record)
+        caption = self.prepare(self.format(no_exc_record), caption_length)
+        return caption
+
+    @staticmethod
+    def get_no_exc_record_copy(record: logging.LogRecord) -> logging.LogRecord:
+        no_exc_record = copy.copy(record)
+        no_exc_record.exc_info = None
+        no_exc_record.exc_text = None
+        return no_exc_record
+
+    def send_plain_text(self, record: logging.LogRecord) -> None:
+        message_length = 4096
+        text = self.prepare(self.format(record), message_length)
+        self.bot.send_message(self.chat_id, text, timeout=5)