add translator

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

@@ -0,0 +1,32 @@
+"""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 +1,32 @@
+"""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/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
 
@@ -37,6 +37,7 @@ 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 +76,8 @@ 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(

src/greek_lang/openai_manager/manager.py

@@ -5,39 +5,47 @@ 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(
         ...,
-        description="description in {target_language}",
+        description="lemma description in {target_language}",
     )
     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(
         ...,
-        description="semantic category in {target_language}",
+        description=f"semantic category in {{target_language}}",
     )
     etymology: 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/translator.py

@@ -0,0 +1,47 @@
+from dependency_injector.wiring import inject, Provide
+from sqlalchemy.ext.asyncio import async_sessionmaker, AsyncSession
+
+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="έμπορος", source_lang=source_lang)
+
+    async with db_session_maker() as db_session, db_session.begin():
+        glossary_word = GlossaryWord(
+            term=word_response.lemma,
+            language=source_lang.value,
+            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(),
+        )
+        db_session.add(glossary_word)
+    return glossary_word