updates

2025-09-10 21:41:42 +03:00 · 2025-09-10 21:41:42 +03:00 · 1ca7efe4d1
commit 1ca7efe4d1
parent f7ed0cb14e
14 changed files with 3020 additions and 2164 deletions
--- a/.env.example
+++ b/.env.example
@ -24,10 +24,13 @@ OPENAI_MODEL=gpt-4o-mini
 OPENAI_EMBEDDINGS_MODEL=text-embedding-3-small

 # LiveKit Configuration (for interview feature)
-LIVEKIT_URL=ws://localhost:7880
+LIVEKIT_URL=wss://hackaton-eizc9zqk.livekit.cloud
 LIVEKIT_API_KEY=devkey
 LIVEKIT_API_SECRET=devkey_secret_32chars_minimum_length

 # App Configuration
 APP_ENV=development
 DEBUG=true
+
+# Domain for Caddy (use your domain for automatic HTTPS)
+DOMAIN=hr.aiquity.xyz
--- a/.env.local
+++ b/.env.local
@ -0,0 +1,2 @@
+NEXT_PUBLIC_API_BASE_URL=http://localhost:8000/api
+NEXT_PUBLIC_LIVEKIT_URL=ws://localhost:7880
--- a/106
+++ b/106
@ -0,0 +1,106 @@
+# Caddyfile for HR AI Backend with automatic HTTPS
+# Environment variable DOMAIN will be used, defaults to localhost
+
+{$DOMAIN:localhost} {
+    # Backend API routes
+    handle /api/* {
+        reverse_proxy backend:8000
+    }
+
+    # Health check endpoint
+    handle /health {
+        reverse_proxy backend:8000
+    }
+
+    # LiveKit WebSocket and HTTP endpoints
+    handle /livekit/* {
+        reverse_proxy livekit:7880
+    }
+
+    # LiveKit WebSocket upgrade
+    handle /rtc {
+        reverse_proxy livekit:7880
+    }
+
+    # Frontend (SPA) - serve everything else
+    handle {
+        reverse_proxy frontend:3000
+        
+        # SPA fallback - serve index.html for all non-API routes
+        @notapi {
+            not path /api/*
+            not path /health
+            not path /livekit/*
+            not path /rtc
+            file {
+                try_files {path} /index.html
+            }
+        }
+        rewrite @notapi /index.html
+    }
+
+    # Enable gzip compression
+    encode gzip
+
+    # Security headers
+    header {
+        # HSTS
+        Strict-Transport-Security "max-age=31536000; includeSubDomains; preload"
+        
+        # XSS Protection
+        X-Content-Type-Options "nosniff"
+        X-Frame-Options "DENY"
+        X-XSS-Protection "1; mode=block"
+        
+        # CORS for API (adjust origins as needed)
+        Access-Control-Allow-Origin "*"
+        Access-Control-Allow-Methods "GET, POST, PUT, DELETE, OPTIONS"
+        Access-Control-Allow-Headers "Content-Type, Authorization"
+    }
+
+    # Logging
+    log {
+        output file /var/log/caddy/access.log
+        format json
+    }
+}
+
+# Development/localhost configuration (no HTTPS)
+localhost {
+    # Backend API routes
+    handle /api/* {
+        reverse_proxy backend:8000
+    }
+
+    # Health check endpoint  
+    handle /health {
+        reverse_proxy backend:8000
+    }
+
+    # LiveKit endpoints
+    handle /livekit/* {
+        reverse_proxy livekit:7880
+    }
+
+    handle /rtc {
+        reverse_proxy livekit:7880
+    }
+
+    # Frontend
+    handle {
+        reverse_proxy frontend:3000
+        
+        @notapi {
+            not path /api/*
+            not path /health
+            not path /livekit/*
+            not path /rtc
+            file {
+                try_files {path} /index.html
+            }
+        }
+        rewrite @notapi /index.html
+    }
+
+    encode gzip
+}
--- a/75
+++ b/75
@ -0,0 +1,75 @@
+FROM --platform=linux/amd64 python:3.11-slim
+
+# Set working directory
+WORKDIR /app
+
+# Install system dependencies
+RUN apt-get update && apt-get install -y \
+    gcc \
+    g++ \
+    curl \
+    git \
+    && rm -rf /var/lib/apt/lists/*
+
+# Install uv for faster package management
+RUN pip install uv
+
+# Copy dependency files
+COPY pyproject.toml uv.lock ./
+
+# Install Python dependencies
+RUN uv sync --frozen --no-dev
+
+# Install Playwright and Chromium for PDF generation
+RUN uv run playwright install-deps
+RUN uv run playwright install chromium
+
+# Copy application code
+COPY . .
+
+# Create directories for agent communication
+RUN mkdir -p /tmp/agent_commands
+
+# Expose the port
+EXPOSE 8000
+
+# Health check
+HEALTHCHECK --interval=30s --timeout=10s --start-period=60s --retries=3 \
+  CMD curl -f http://localhost:8000/health || exit 1
+
+# Create startup script
+COPY <<EOF /app/start.sh
+#!/bin/bash
+set -e
+
+# Load environment variables from .env file if it exists
+# Only set variables that are not already set
+if [ -f .env ]; then
+    echo "Loading environment variables from .env file..."
+    set -a  # automatically export all variables
+    source .env
+    set +a  # stop auto-exporting
+fi
+
+# Run database migrations
+echo "Running database migrations..."
+uv run alembic upgrade head
+
+# Start Celery worker in background
+echo "Starting Celery worker in background..."
+uv run celery -A celery_worker.celery_app worker --loglevel=info --pool=solo &
+
+# Start FastAPI server
+if [ "\$APP_ENV" = "development" ]; then
+    echo "Starting FastAPI development server..."
+    exec uv run fastapi dev main.py --host 0.0.0.0 --port 8000
+else
+    echo "Starting FastAPI production server..."
+    exec uv run uvicorn app.main:app --host 0.0.0.0 --port 8000
+fi
+EOF
+
+RUN chmod +x /app/start.sh
+
+# Default command
+CMD ["/app/start.sh"]
--- a/INTERVIEW_INTEGRATION.md
+++ b/INTERVIEW_INTEGRATION.md
@ -42,7 +42,7 @@ POST /api/interview/{resumeId}/token
 {
  "token": "livekit_jwt_token_here",
  "roomName": "interview_room_123",
-  "serverUrl": "wss://your-livekit-server.com"
+  "serverUrl": "ws://your-livekit-server.com"
 }
 ```

@ -50,7 +50,7 @@ POST /api/interview/{resumeId}/token

 #### Environment Variables
 ```env
-NEXT_PUBLIC_LIVEKIT_URL=wss://your-livekit-server.com
+NEXT_PUBLIC_LIVEKIT_URL=ws://your-livekit-server.com
 LIVEKIT_API_KEY=your_api_key
 LIVEKIT_API_SECRET=your_api_secret
 ```
--- a/ai_interviewer_agent.py
+++ b/ai_interviewer_agent.py
@ -82,6 +82,28 @@ class InterviewAgent:
            return self.sections[self.current_section]
        return {}

+    def _format_questions(self, questions):
+        """
+        Форматирует список вопросов в строку, независимо от их структуры
+        """
+        if not questions:
+            return "Нет вопросов"
+            
+        formatted = []
+        for question in questions:
+            if isinstance(question, str):
+                # Простая строка
+                formatted.append(question)
+            elif isinstance(question, dict):
+                # Объект с полями (например, из LLM генерации)
+                question_text = question.get("question", question.get("text", str(question)))
+                formatted.append(question_text)
+            else:
+                # Любой другой тип - приводим к строке
+                formatted.append(str(question))
+        
+        return ", ".join(formatted)
+
    def get_next_question(self) -> str:
        """Получить следующий вопрос"""
        section = self.get_current_section()
@ -129,7 +151,7 @@ class InterviewAgent:
        # Формируем план интервью для агента
        sections_info = "\n".join(
            [
-                f"- {section.get('name', 'Секция')}: {', '.join(section.get('questions', []))}"
+                f"- {section.get('name', 'Секция')}: {self._format_questions(section.get('questions', []))}"
                for section in self.sections
            ]
        )
@ -748,10 +770,23 @@ async def entrypoint(ctx: JobContext):


 def main():
-    logging.basicConfig(level=logging.INFO)
-    asyncio.set_event_loop_policy(
-        asyncio.WindowsSelectorEventLoopPolicy()
-    )  # фикс для Windows
+    # Настройка логирования для продакшена
+    if os.getenv("APP_ENV") == "production":
+        logging.basicConfig(
+            level=logging.INFO,
+            format='{"timestamp": "%(asctime)s", "level": "%(levelname)s", "logger": "%(name)s", "message": "%(message)s", "module": "%(module)s", "line": %(lineno)d}',
+            datefmt='%Y-%m-%dT%H:%M:%S'
+        )
+    else:
+        logging.basicConfig(
+            level=logging.DEBUG if os.getenv("DEBUG") == "true" else logging.INFO,
+            format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
+        )
+    
+    # Устанавливаем политику цикла событий только для Windows
+    if os.name == "nt":  # Windows
+        asyncio.set_event_loop_policy(asyncio.WindowsProactorEventLoopPolicy())
+    
    cli.run_app(WorkerOptions(entrypoint_fnc=entrypoint))


--- a/celery_worker/celery_app.py
+++ b/celery_worker/celery_app.py
@ -1,4 +1,6 @@
+import logging
 from celery import Celery
+from celery.signals import setup_logging

 from rag.settings import settings

@ -15,4 +17,24 @@ celery_app.conf.update(
    result_serializer="json",
    timezone="UTC",
    enable_utc=True,
+    worker_log_format="[%(asctime)s: %(levelname)s/%(processName)s] %(message)s",
+    worker_task_log_format="[%(asctime)s: %(levelname)s/%(processName)s][%(task_name)s(%(task_id)s)] %(message)s",
+    task_acks_late=True,
+    worker_prefetch_multiplier=1,
+    task_reject_on_worker_lost=True,
+    result_extended=True,
 )
+
+@setup_logging.connect
+def config_loggers(*args, **kwargs):
+    """Configure logging for Celery worker"""
+    logging.basicConfig(
+        level=logging.INFO,
+        format="[%(asctime)s: %(levelname)s/%(name)s] %(message)s",
+        datefmt="%Y-%m-%d %H:%M:%S",
+    )
+    
+    # Set specific log levels
+    logging.getLogger("celery").setLevel(logging.INFO)
+    logging.getLogger("celery.worker").setLevel(logging.INFO)
+    logging.getLogger("celery.task").setLevel(logging.INFO)
--- a/celery_worker/interview_analysis_task.py
+++ b/celery_worker/interview_analysis_task.py
@ -1,4 +1,3 @@
-import asyncio
 import json
 import logging
 from datetime import datetime
@ -100,16 +99,45 @@ def generate_interview_report(resume_id: int):
            # Сохраняем отчет в БД
            report_instance = _save_report_to_db(db, resume_id, report)

-            # Генерируем и загружаем PDF отчет
+            # Запускаем отдельную задачу для генерации PDF
            if report_instance:
-                asyncio.run(
-                    _generate_and_upload_pdf_report(
-                        db,
-                        report_instance,
-                        resume.applicant_name,
-                        vacancy.get("title", "Unknown Position"),
-                        resume.resume_file_url,
-                    )
+                from celery_worker.tasks import generate_pdf_report_task
+
+                report_data = {
+                    "id": report_instance.id,
+                    "interview_session_id": report_instance.interview_session_id,
+                    "overall_score": report_instance.overall_score,
+                    "technical_skills_score": report_instance.technical_skills_score,
+                    "technical_skills_justification": report_instance.technical_skills_justification,
+                    "technical_skills_concerns": report_instance.technical_skills_concerns,
+                    "communication_score": report_instance.communication_score,
+                    "communication_justification": report_instance.communication_justification,
+                    "communication_concerns": report_instance.communication_concerns,
+                    "problem_solving_score": report_instance.problem_solving_score,
+                    "problem_solving_justification": report_instance.problem_solving_justification,
+                    "problem_solving_concerns": report_instance.problem_solving_concerns,
+                    "experience_relevance_score": report_instance.experience_relevance_score,
+                    "experience_relevance_justification": report_instance.experience_relevance_justification,
+                    "experience_relevance_concerns": report_instance.experience_relevance_concerns,
+                    "cultural_fit_score": report_instance.cultural_fit_score,
+                    "cultural_fit_justification": report_instance.cultural_fit_justification,
+                    "cultural_fit_concerns": report_instance.cultural_fit_concerns,
+                    "recommendation": report_instance.recommendation,
+                    "strengths": report_instance.strengths,
+                    "weaknesses": report_instance.weaknesses,
+                    "red_flags": report_instance.red_flags,
+                    "questions_analysis": report_instance.questions_analysis,
+                    "next_steps": report_instance.next_steps,
+                    "analysis_method": report_instance.analysis_method,
+                    "created_at": report_instance.created_at.isoformat() if report_instance.created_at else None,
+                    "updated_at": report_instance.updated_at.isoformat() if report_instance.updated_at else None,
+                }
+
+                generate_pdf_report_task.delay(
+                    report_data=report_data,
+                    candidate_name=resume.applicant_name,
+                    position=vacancy.get("title", "Unknown Position"),
+                    resume_file_url=resume.resume_file_url,
                )

            logger.info(
--- a/celery_worker/tasks.py
+++ b/celery_worker/tasks.py
@ -1,8 +1,12 @@
 import json
+import logging
 import os
 from typing import Any

 from celery_worker.celery_app import celery_app
+
+# Настраиваем логгер для задач
+logger = logging.getLogger(__name__)
 from celery_worker.database import (
    SyncResumeRepository,
    SyncVacancyRepository,
@ -223,7 +227,7 @@ def generate_interview_plan(
        }

    except Exception as e:
-        print(f"Ошибка генерации плана интервью: {str(e)}")
+        logger.error(f"Ошибка генерации плана интервью: {str(e)}", exc_info=True)
        return None


@ -236,24 +240,37 @@ def parse_resume_task(self, resume_id: str, file_path: str):
        resume_id: ID резюме
        file_path: Путь к PDF файлу резюме
    """
+    logger.info(f"=== НАЧАЛО ОБРАБОТКИ РЕЗЮМЕ {resume_id} ===")
+    logger.info(f"Путь к файлу: {file_path}")

    try:
        # Шаг 0: Обновляем статус в БД - начали парсинг
+        logger.info(f"Шаг 0: Обновляем статус резюме {resume_id} на 'parsing'")
        with get_sync_session() as session:
            repo = SyncResumeRepository(session)
            repo.update_status(int(resume_id), "parsing")
+        logger.info(f"Статус резюме {resume_id} успешно обновлен на 'parsing'")

        # Обновляем статус задачи
+        logger.info(f"Обновляем состояние Celery задачи на PENDING")
        self.update_state(
            state="PENDING",
            meta={"status": "Начинаем парсинг резюме...", "progress": 10},
        )
+        logger.info(f"Состояние Celery задачи обновлено")

        # Инициализируем модели из registry
+        logger.info(f"Шаг 1: Инициализируем модели из registry")
        try:
+            logger.info("Получаем chat_model из registry")
            chat_model = registry.get_chat_model()
+            logger.info("Chat model успешно получен")
+            
+            logger.info("Получаем vector_store из registry")
            vector_store = registry.get_vector_store()
+            logger.info("Vector store успешно получен")
        except Exception as e:
+            logger.error(f"ОШИБКА при инициализации моделей: {str(e)}", exc_info=True)
            # Обновляем статус в БД - ошибка инициализации
            with get_sync_session() as session:
                repo = SyncResumeRepository(session)
@ -262,17 +279,23 @@ def parse_resume_task(self, resume_id: str, file_path: str):
                    "failed",
                    error_message=f"Ошибка инициализации моделей: {str(e)}",
                )
-            raise Exception(f"Ошибка инициализации моделей: {str(e)}")
+            raise RuntimeError(f"Ошибка инициализации моделей: {str(e)}")

-        # Шаг 1: Парсинг резюме
+        # Шаг 2: Парсинг резюме
+        logger.info(f"Шаг 2: Начинаем парсинг резюме")
        self.update_state(
            state="PROGRESS",
            meta={"status": "Извлекаем текст из PDF...", "progress": 20},
        )
+        logger.info(f"Состояние Celery обновлено на PROGRESS (20%)")

+        logger.info(f"Создаем ResumeParser")
        parser = ResumeParser(chat_model)
+        logger.info(f"ResumeParser создан успешно")

+        logger.info(f"Проверяем существование файла: {file_path}")
        if not os.path.exists(file_path):
+            logger.error(f"ФАЙЛ НЕ НАЙДЕН: {file_path}")
            # Обновляем статус в БД - файл не найден
            with get_sync_session() as session:
                repo = SyncResumeRepository(session)
@ -281,61 +304,88 @@ def parse_resume_task(self, resume_id: str, file_path: str):
                    "failed",
                    error_message=f"Файл не найден: {file_path}",
                )
-            raise Exception(f"Файл не найден: {file_path}")
+            logger.info(f"Статус резюме {resume_id} обновлен на 'failed' в БД")
+            raise FileNotFoundError(f"Файл не найден: {file_path}")

+        logger.info(f"Файл существует, начинаем парсинг")
        parsed_resume = parser.parse_resume_from_file(file_path)
+        logger.info(f"Парсинг резюме завершен, получены данные: {list(parsed_resume.keys())}")

        # Получаем оригинальные данные из формы
+        logger.info(f"Шаг 3: Получаем данные резюме из БД")
        with get_sync_session() as session:
            repo = SyncResumeRepository(session)
            resume_record = repo.get_by_id(int(resume_id))
            if not resume_record:
-                raise Exception(f"Резюме с ID {resume_id} не найдено в базе данных")
+                logger.error(f"РЕЗЮМЕ С ID {resume_id} НЕ НАЙДЕНО В БД")
+                raise ValueError(f"Резюме с ID {resume_id} не найдено в базе данных")

            # Извлекаем нужные данные пока сессия активна
            applicant_name = resume_record.applicant_name
            applicant_email = resume_record.applicant_email
            applicant_phone = resume_record.applicant_phone
+            logger.info(f"Данные резюме получены: name={applicant_name}, email={applicant_email}, phone={applicant_phone}")

        # Создаем комбинированные данные: навыки и опыт из парсинга, контакты из формы
+        logger.info(f"Шаг 4: Объединяем данные из парсинга и формы")
        combined_data = parsed_resume.copy()
        combined_data["name"] = applicant_name or parsed_resume.get("name", "")
        combined_data["email"] = applicant_email or parsed_resume.get("email", "")
        combined_data["phone"] = applicant_phone or parsed_resume.get("phone", "")
+        logger.info(f"Комбинированные данные подготовлены")

-        # Шаг 2: Векторизация и сохранение в Milvus
+        # Шаг 5: Векторизация и сохранение в Milvus
+        logger.info(f"Шаг 5: Векторизация и сохранение в Milvus")
        self.update_state(
            state="PENDING",
            meta={"status": "Сохраняем в векторную базу...", "progress": 60},
        )
+        logger.info(f"Состояние Celery обновлено на 60%")

+        logger.info(f"Добавляем профиль кандидата в vector store")
        vector_store.add_candidate_profile(str(resume_id), combined_data)
+        logger.info(f"Профиль кандидата добавлен в vector store")

-        # Шаг 3: Обновляем статус в PostgreSQL - успешно обработано
+        # Шаг 6: Обновляем статус в PostgreSQL
+        logger.info(f"Шаг 6: Подготовка к обновлению статуса в БД")
        self.update_state(
            state="PENDING",
            meta={"status": "Обновляем статус в базе данных...", "progress": 85},
        )
+        logger.info(f"Состояние Celery обновлено на 85%")

-        # Шаг 4: Генерируем план интервью
+        # Шаг 7: Генерируем план интервью
+        logger.info(f"Шаг 7: Генерация плана интервью")
        self.update_state(
            state="PENDING",
            meta={"status": "Генерируем план интервью...", "progress": 90},
        )
+        logger.info(f"Состояние Celery обновлено на 90%")

+        logger.info(f"Вызываем generate_interview_plan для резюме {resume_id}")
        interview_plan = generate_interview_plan(int(resume_id), combined_data)
+        logger.info(f"План интервью сгенерирован: {interview_plan is not None}")

+        logger.info(f"Шаг 8: Обновляем статус в БД на основе плана интервью")
        with get_sync_session() as session:
            repo = SyncResumeRepository(session)

            # Проверяем результат генерации плана интервью
-            print("interview_plan", interview_plan)
+            logger.info(f"Анализируем план интервью для резюме {resume_id}")
+            logger.info(f"План интервью: {interview_plan}")
+            
            if interview_plan and interview_plan.get("is_suitable", True):
+                logger.info(f"Кандидат подходит, обновляем статус на 'parsed'")
                # Кандидат подходит - обновляем статус на parsed
                repo.update_status(int(resume_id), "parsed", parsed_data=combined_data)
+                logger.info(f"Статус резюме {resume_id} обновлен на 'parsed'")
+                
                # Сохраняем план интервью
+                logger.info(f"Сохраняем план интервью для резюме {resume_id}")
                repo.update_interview_plan(int(resume_id), interview_plan)
+                logger.info(f"План интервью сохранен")
            else:
+                logger.info(f"Кандидат НЕ подходит, отклоняем")
                # Кандидат не подходит - отклоняем
                rejection_reason = (
                    interview_plan.get(
@ -344,14 +394,17 @@ def parse_resume_task(self, resume_id: str, file_path: str):
                    if interview_plan
                    else "Ошибка анализа соответствия"
                )
+                logger.info(f"Причина отклонения: {rejection_reason}")
                repo.update_status(
                    int(resume_id),
                    "rejected",
                    parsed_data=combined_data,
                    rejection_reason=rejection_reason,
                )
+                logger.info(f"Статус резюме {resume_id} обновлен на 'rejected'")

                # Завершаем с информацией об отклонении
+                logger.info(f"Обновляем состояние Celery на SUCCESS (отклонен)")
                self.update_state(
                    state="SUCCESS",
                    meta={
@ -362,6 +415,7 @@ def parse_resume_task(self, resume_id: str, file_path: str):
                        "rejection_reason": rejection_reason,
                    },
                )
+                logger.info(f"=== ЗАВЕРШЕНИЕ ОБРАБОТКИ РЕЗЮМЕ {resume_id} (ОТКЛОНЕН) ===")

                return {
                    "resume_id": resume_id,
@ -371,6 +425,7 @@ def parse_resume_task(self, resume_id: str, file_path: str):
                }

        # Завершено успешно
+        logger.info(f"Обновляем состояние Celery на SUCCESS (принят)")
        self.update_state(
            state="SUCCESS",
            meta={
@ -379,6 +434,7 @@ def parse_resume_task(self, resume_id: str, file_path: str):
                "result": combined_data,
            },
        )
+        logger.info(f"=== УСПЕШНОЕ ЗАВЕРШЕНИЕ ОБРАБОТКИ РЕЗЮМЕ {resume_id} ===")

        return {
            "resume_id": resume_id,
@ -387,13 +443,16 @@ def parse_resume_task(self, resume_id: str, file_path: str):
        }

    except Exception as e:
+        error_message = str(e)
+        logger.error(f"Ошибка при обработке резюме {resume_id}: {error_message}", exc_info=True)
+        
        # В случае ошибки
        self.update_state(
            state="FAILURE",
            meta={
-                "status": f"Ошибка при обработке резюме: {str(e)}",
+                "status": f"Ошибка при обработке резюме: {error_message}",
                "progress": 0,
-                "error": str(e),
+                "error": error_message,
            },
        )

@ -401,11 +460,16 @@ def parse_resume_task(self, resume_id: str, file_path: str):
        try:
            with get_sync_session() as session:
                repo = SyncResumeRepository(session)
-                repo.update_status(int(resume_id), "failed", error_message=str(e))
+                repo.update_status(int(resume_id), "failed", error_message=error_message)
        except Exception as db_error:
-            print(f"Ошибка при обновлении статуса в БД: {str(db_error)}")
+            logger.error(f"Ошибка при обновлении статуса в БД: {str(db_error)}", exc_info=True)

-        raise
+        # Возвращаем стандартное исключение вместо re-raise
+        return {
+            "resume_id": resume_id,
+            "status": "failed",
+            "error": error_message,
+        }


 # Функция больше не нужна - используем SyncResumeRepository напрямую
@ -431,7 +495,7 @@ def generate_interview_questions_task(self, resume_id: str, job_description: str
            chat_model = registry.get_chat_model()
            vector_store = registry.get_vector_store()
        except Exception as e:
-            raise Exception(f"Ошибка инициализации моделей: {str(e)}")
+            raise RuntimeError(f"Ошибка инициализации моделей: {str(e)}")

        # Шаг 1: Получить parsed резюме из базы данных
        self.update_state(
--- a/docker-compose.yml
+++ b/docker-compose.yml
@ -0,0 +1,133 @@
+
+services:
+  # PostgreSQL Database
+  postgres:
+    image: postgres:15
+    environment:
+      POSTGRES_DB: hr_ai
+      POSTGRES_USER: hr_user
+      POSTGRES_PASSWORD: hr_password
+    volumes:
+      - postgres_data:/var/lib/postgresql/data
+    ports:
+      - "5432:5432"
+    healthcheck:
+      test: ["CMD-SHELL", "pg_isready -U hr_user -d hr_ai"]
+      interval: 30s
+      timeout: 10s
+      retries: 5
+
+  # Redis for Celery and caching
+  redis:
+    image: redis:7-alpine
+    ports:
+      - "6379:6379"
+    volumes:
+      - redis_data:/data
+    healthcheck:
+      test: ["CMD", "redis-cli", "ping"]
+      interval: 30s
+      timeout: 10s
+      retries: 3
+
+  # LiveKit Server
+  livekit:
+    image: livekit/livekit-server:latest
+    ports:
+      - "7880:7880"
+      - "7881:7881"
+      - "7882:7882/udp"
+      - "3478:3478/udp"
+    volumes:
+      - caddy_data:/certs
+    depends_on:
+      - caddy
+    restart: unless-stopped
+    environment:
+      LIVEKIT_CONFIG: |
+        keys:
+          devkey: devkey_secret_32chars_minimum_length
+        webhook:
+          api_key: devkey
+        turn:
+          enabled: true
+          tls_port: 5349
+          domain: hr.aiquity.xyz
+          cert_file: /certs/certificates/acme-v02.api.letsencrypt.org-directory/${DOMAIN:-localhost}/${DOMAIN:-localhost}.crt
+          key_file: /certs/certificates/acme-v02.api.letsencrypt.org-directory/${DOMAIN:-localhost}/${DOMAIN:-localhost}.key
+        port: 7880
+        rtc:
+          tcp_port: 7881
+          port_range_start: 50000
+          port_range_end: 60000
+          use_external_ip: true
+        redis:
+          address: redis:6379
+        
+
+  # HR AI Backend
+  backend:
+    image: cr.yandex/crp9p5rtbnbop36duusi/hr-ai-backend:latest
+    expose:
+      - "8000"
+    env_file:
+      - .env
+    environment:
+      - DATABASE_URL=postgresql+asyncpg://hr_user:hr_password@postgres:5432/hr_ai
+      - REDIS_CACHE_URL=redis
+      - REDIS_CACHE_PORT=6379
+      - REDIS_CACHE_DB=0
+      - LIVEKIT_URL=ws://livekit:7880
+      - LIVEKIT_API_KEY=devkey
+      - LIVEKIT_API_SECRET=devkey_secret_32chars_minimum_length
+      - APP_ENV=development
+      - DEBUG=true
+    volumes:
+      - ./agent_commands:/tmp/agent_commands
+      - backend_uploads:/app/uploads
+    depends_on:
+      postgres:
+        condition: service_healthy
+      redis:
+        condition: service_healthy
+      livekit:
+        condition: service_started
+    healthcheck:
+      test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
+      interval: 30s
+      timeout: 10s
+      retries: 3
+
+
+  # Caddy reverse proxy with automatic HTTPS
+  caddy:
+    image: caddy:2-alpine
+    ports:
+      - "80:80"
+      - "443:443"
+    volumes:
+      - ./Caddyfile:/etc/caddy/Caddyfile:ro
+      - caddy_data:/data
+      - caddy_config:/config
+    depends_on:
+      - backend
+      - livekit
+    environment:
+      - DOMAIN=${DOMAIN:-localhost}
+    restart: unless-stopped
+
+  frontend:
+    image: cr.yandex/crp9p5rtbnbop36duusi/hr-ai-frontend:latest
+    expose:
+      - "3000"
+    environment:
+      - NODE_ENV=production
+    restart: unless-stopped
+    volumes:
+      - ./.env.local:/app/.env.local:ro
+volumes:
+  postgres_data:
+  redis_data:
+  backend_uploads:
+  caddy_data:
+  caddy_config:
--- a/pyproject.toml
+++ b/pyproject.toml
@ -36,6 +36,7 @@ dependencies = [
    "greenlet>=3.2.4",
    "xhtml2pdf>=0.2.17",
    "playwright>=1.55.0",
+    "celery-types==0.23.0",
 ]

 [build-system]
--- a/scripts/build-and-push.sh
+++ b/scripts/build-and-push.sh
@ -0,0 +1,71 @@
+#!/bin/bash
+
+# Build and push script for Yandex Cloud Container Registry
+# Usage: ./scripts/build-and-push. [tag]
+
+set -e
+
+# Configuration
+REGISTRY_ID="${YANDEX_REGISTRY_ID:-your-registry-id}"
+IMAGE_NAME="hr-ai-backend"
+TAG="${1:-latest}"
+FULL_IMAGE_NAME="cr.yandex/${REGISTRY_ID}/${IMAGE_NAME}:${TAG}"
+
+# Colors for output
+RED='\033[0;31m'
+GREEN='\033[0;32m'
+YELLOW='\033[1;33m'
+NC='\033[0m' # No Color
+
+echo -e "${YELLOW}Building and pushing HR AI Backend to Yandex Cloud Container Registry${NC}"
+
+# Check if required environment variables are set
+if [ -z "$REGISTRY_ID" ] || [ "$REGISTRY_ID" = "your-registry-id" ]; then
+    echo -e "${RED}Error: YANDEX_REGISTRY_ID environment variable is not set${NC}"
+    echo "Please set it to your Yandex Cloud Container Registry ID"
+    echo "Example: export YANDEX_REGISTRY_ID=crp1234567890abcdef"
+    exit 1
+fi
+
+# Check if yc CLI is installed and authenticated
+if ! command -v yc &> /dev/null; then
+    echo -e "${RED}Error: Yandex Cloud CLI (yc) is not installed${NC}"
+    echo "Please install it from: https://cloud.yandex.ru/docs/cli/quickstart"
+    exit 1
+fi
+
+# Check authentication
+if ! yc config list | grep -q "token:"; then
+    echo -e "${RED}Error: Not authenticated with Yandex Cloud${NC}"
+    echo "Please run: yc init"
+    exit 1
+fi
+
+echo -e "${YELLOW}Configuring Docker for Yandex Cloud Container Registry...${NC}"
+yc container registry configure-docker
+
+echo -e "${YELLOW}Building Docker image: ${FULL_IMAGE_NAME}${NC}"
+docker build -t "${FULL_IMAGE_NAME}" .
+
+echo -e "${YELLOW}Pushing image to registry...${NC}"
+docker push "${FULL_IMAGE_NAME}"
+
+echo -e "${GREEN}✓ Successfully built and pushed: ${FULL_IMAGE_NAME}${NC}"
+
+# Also tag as latest if a specific tag was provided
+if [ "$TAG" != "latest" ]; then
+    LATEST_IMAGE_NAME="cr.yandex/${REGISTRY_ID}/${IMAGE_NAME}:latest"
+    echo -e "${YELLOW}Tagging as latest...${NC}"
+    docker tag "${FULL_IMAGE_NAME}" "${LATEST_IMAGE_NAME}"
+    docker push "${LATEST_IMAGE_NAME}"
+    echo -e "${GREEN}✓ Also pushed as: ${LATEST_IMAGE_NAME}${NC}"
+fi
+
+echo -e "${GREEN}Build and push completed successfully!${NC}"
+echo ""
+echo "Image is available at:"
+echo "  ${FULL_IMAGE_NAME}"
+echo ""
+echo "To use in production, update your docker-compose.prod.yml:"
+echo "  backend:"
+echo "    image: ${FULL_IMAGE_NAME}"
--- a/scripts/deploy.sh
+++ b/scripts/deploy.sh
@ -0,0 +1,301 @@
+#!/bin/bash
+
+# SSH Deploy script for HR AI Backend
+# Usage: ./scripts/deploy.sh [environment] [image_tag]
+
+set -e
+
+# Configuration
+ENVIRONMENT="${1:-production}"
+IMAGE_TAG="${2:-latest}"
+REGISTRY_ID="${YANDEX_REGISTRY_ID:-your-registry-id}"
+IMAGE_NAME="hr-ai-backend"
+FULL_IMAGE_NAME="cr.yandex/${REGISTRY_ID}/${IMAGE_NAME}:${IMAGE_TAG}"
+
+# Server configuration (set these as environment variables)
+SERVER_HOST="${DEPLOY_HOST:-your-server.com}"
+SERVER_USER="${DEPLOY_USER:-deploy}"
+SERVER_PORT="${DEPLOY_PORT:-22}"
+DEPLOY_PATH="${DEPLOY_PATH:-/opt/hr-ai-backend}"
+
+# Colors for output
+RED='\033[0;31m'
+GREEN='\033[0;32m'
+YELLOW='\033[1;33m'
+BLUE='\033[0;34m'
+NC='\033[0m' # No Color
+
+echo -e "${YELLOW}Deploying HR AI Backend to ${ENVIRONMENT} environment${NC}"
+
+# Check if required environment variables are set
+missing_vars=""
+if [ -z "$SERVER_HOST" ] || [ "$SERVER_HOST" = "your-server.com" ]; then
+    missing_vars="$missing_vars DEPLOY_HOST"
+fi
+if [ -z "$SERVER_USER" ] || [ "$SERVER_USER" = "deploy" ]; then
+    missing_vars="$missing_vars DEPLOY_USER"
+fi
+if [ -z "$REGISTRY_ID" ] || [ "$REGISTRY_ID" = "your-registry-id" ]; then
+    missing_vars="$missing_vars YANDEX_REGISTRY_ID"
+fi
+
+if [ -n "$missing_vars" ]; then
+    echo -e "${RED}Error: Required environment variables are not set:${NC}"
+    for var in $missing_vars; do
+        echo "  - $var"
+    done
+    echo ""
+    echo "Example configuration:"
+    echo "export DEPLOY_HOST=your-server.example.com"
+    echo "export DEPLOY_USER=deploy"
+    echo "export YANDEX_REGISTRY_ID=crp1234567890abcdef"
+    echo "export DEPLOY_PATH=/opt/hr-ai-backend  # optional"
+    echo "export DEPLOY_PORT=22  # optional"
+    exit 1
+fi
+
+# Test SSH connection
+echo -e "${BLUE}Testing SSH connection to ${SERVER_USER}@${SERVER_HOST}:${SERVER_PORT}...${NC}"
+if ! ssh -p "${SERVER_PORT}" -o ConnectTimeout=10 -o StrictHostKeyChecking=no "${SERVER_USER}@${SERVER_HOST}" "echo 'SSH connection successful'"; then
+    echo -e "${RED}Error: Cannot connect to server via SSH${NC}"
+    echo "Please check your SSH key configuration and server details"
+    exit 1
+fi
+
+# Create deployment directory structure on server
+echo -e "${BLUE}Creating deployment directories on server...${NC}"
+ssh -p "${SERVER_PORT}" "${SERVER_USER}@${SERVER_HOST}" "
+    sudo mkdir -p ${DEPLOY_PATH}/{config,logs,data,agent_commands}
+    sudo mkdir -p ${DEPLOY_PATH}/data/{postgres,redis,uploads,caddy_data,caddy_config}
+    sudo mkdir -p ${DEPLOY_PATH}/logs/{caddy}
+    sudo chown -R ${SERVER_USER}:${SERVER_USER} ${DEPLOY_PATH}
+"
+
+# Copy configuration files to server
+echo -e "${BLUE}Copying configuration files to server...${NC}"
+scp -P "${SERVER_PORT}" docker-compose.yml "${SERVER_USER}@${SERVER_HOST}:${DEPLOY_PATH}/"
+scp -P "${SERVER_PORT}" Caddyfile "${SERVER_USER}@${SERVER_HOST}:${DEPLOY_PATH}/"
+
+# Create frontend environment file
+echo -e "${BLUE}Creating frontend environment configuration...${NC}"
+ssh -p "${SERVER_PORT}" "${SERVER_USER}@${SERVER_HOST}" "
+cat > ${DEPLOY_PATH}/.env.local << 'EOF'
+# Frontend Environment Configuration
+NODE_ENV=production
+
+# API URL (adjust based on your setup)
+NEXT_PUBLIC_API_URL=https://\${SERVER_HOST}/api
+REACT_APP_API_URL=https://\${SERVER_HOST}/api
+VUE_APP_API_URL=https://\${SERVER_HOST}/api
+
+# LiveKit Configuration for frontend
+NEXT_PUBLIC_LIVEKIT_URL=ws://\${SERVER_HOST}/rtc
+REACT_APP_LIVEKIT_URL=ws://\${SERVER_HOST}/rtc
+VUE_APP_LIVEKIT_URL=ws://\${SERVER_HOST}/rtc
+
+# For localhost development (no HTTPS)
+# NEXT_PUBLIC_API_URL=http://\${SERVER_HOST}/api
+# REACT_APP_API_URL=http://\${SERVER_HOST}/api
+# VUE_APP_API_URL=http://\${SERVER_HOST}/api
+# NEXT_PUBLIC_LIVEKIT_URL=ws://\${SERVER_HOST}/rtc
+# REACT_APP_LIVEKIT_URL=ws://\${SERVER_HOST}/rtc
+# VUE_APP_LIVEKIT_URL=ws://\${SERVER_HOST}/rtc
+
+# Add your frontend-specific environment variables here
+EOF
+"
+
+# Create production environment file on server
+echo -e "${BLUE}Creating production environment configuration...${NC}"
+ssh -p "${SERVER_PORT}" "${SERVER_USER}@${SERVER_HOST}" "
+cat > ${DEPLOY_PATH}/.env << 'EOF'
+# Production Environment Configuration
+DATABASE_URL=postgresql+asyncpg://hr_user:hr_password@postgres:5432/hr_ai
+REDIS_CACHE_URL=redis
+REDIS_CACHE_PORT=6379
+REDIS_CACHE_DB=0
+
+# LiveKit Configuration
+LIVEKIT_URL=ws://livekit:7880
+LIVEKIT_API_KEY=devkey
+LIVEKIT_API_SECRET=devkey_secret_32chars_minimum_length
+
+# Caddy Domain Configuration (set your domain for automatic HTTPS)
+DOMAIN=${SERVER_HOST:-localhost}
+
+# App Configuration
+APP_ENV=production
+DEBUG=false
+
+# Add your production API keys here:
+# OPENAI_API_KEY=your-openai-api-key
+# DEEPGRAM_API_KEY=your-deepgram-api-key
+# CARTESIA_API_KEY=your-cartesia-api-key
+# ELEVENLABS_API_KEY=your-elevenlabs-api-key
+
+# S3 Storage Configuration (optional)
+# S3_ENDPOINT_URL=https://s3.storage.selcloud.ru
+# S3_ACCESS_KEY_ID=your_s3_access_key
+# S3_SECRET_ACCESS_KEY=your_s3_secret_key
+# S3_BUCKET_NAME=your-bucket-name
+# S3_REGION=ru-1
+
+# Milvus Vector Database Configuration (optional)
+# MILVUS_URI=http://milvus:19530
+# MILVUS_COLLECTION=hr_candidate_profiles
+EOF
+"
+
+# Create production docker compose override
+echo -e "${BLUE}Creating production docker compose configuration...${NC}"
+ssh -p "${SERVER_PORT}" "${SERVER_USER}@${SERVER_HOST}" "
+cat > ${DEPLOY_PATH}/docker-compose.prod.yml << 'EOF'
+
+services:
+  backend:
+    image: ${FULL_IMAGE_NAME}
+    env_file:
+      - .env
+    restart: unless-stopped
+    volumes:
+      - ./agent_commands:/tmp/agent_commands
+      - ./data/uploads:/app/uploads
+      - ./logs:/app/logs
+
+  postgres:
+    restart: unless-stopped
+    volumes:
+      - ./data/postgres:/var/lib/postgresql/data
+
+  redis:
+    restart: unless-stopped
+    volumes:
+      - ./data/redis:/data
+
+  livekit:
+    restart: unless-stopped
+    ports:
+      - \"3478:3478/udp\"
+
+  caddy:
+    env_file:
+      - .env
+    restart: unless-stopped
+    volumes:
+      - ./data/caddy_data:/data
+      - ./data/caddy_config:/config
+      - ./logs/caddy:/var/log/caddy
+
+  frontend:
+    restart: unless-stopped
+EOF
+"
+
+# Pull latest image and deploy
+echo -e "${BLUE}Pulling latest image and starting services...${NC}"
+ssh -p "${SERVER_PORT}" "${SERVER_USER}@${SERVER_HOST}" "
+    cd ${DEPLOY_PATH}
+    
+    # Configure Docker for Yandex Cloud Registry
+    echo 'Configuring Docker for Yandex Cloud Registry...'
+    
+    # Completely reset Docker config to fix credential helper issues
+    echo 'Resetting Docker configuration...'
+    mkdir -p ~/.docker
+    cat > ~/.docker/config.json << 'DOCKER_CONFIG'
+{
+    \"auths\": {},
+    \"HttpHeaders\": {
+        \"User-Agent\": \"Docker-Client/20.10.0 (linux)\"
+    }
+}
+DOCKER_CONFIG
+    
+    # Install yc CLI if not found
+    if ! command -v yc &> /dev/null; then
+        echo 'Installing Yandex Cloud CLI...'
+        curl -sSL https://storage.yandexcloud.net/yandexcloud-yc/install.sh | bash
+        source ~/.bashrc || source ~/.bash_profile || export PATH=\"\$HOME/yandex-cloud/bin:\$PATH\"
+    fi
+    
+    # Use manual login instead of yc configure-docker
+    if command -v yc &> /dev/null; then
+        echo 'Getting Yandex Cloud token and logging in manually...'
+        YC_TOKEN=\$(yc iam create-token 2>/dev/null)
+        if [ ! -z \"\$YC_TOKEN\" ]; then
+            echo \"\$YC_TOKEN\" | docker login --username oauth --password-stdin cr.yandex
+            echo 'Docker login successful'
+        else
+            echo 'Error: Could not get YC token. Please run: yc init'
+            echo 'You need to authenticate yc CLI first on the server'
+            exit 1
+        fi
+    else
+        echo 'Error: yc CLI installation failed'
+        exit 1
+    fi
+    
+    echo 'Current Docker config:'
+    cat ~/.docker/config.json
+    
+    # Pull only our custom images from Yandex Registry first
+    echo 'Pulling custom images from Yandex Registry...'
+    docker pull ${FULL_IMAGE_NAME} || echo 'Failed to pull backend image'
+    docker pull cr.yandex/crp9p5rtbnbop36duusi/hr-ai-frontend:latest || echo 'Failed to pull frontend image'
+    
+    # Reset Docker config to default for pulling public images
+    echo 'Resetting Docker config for public images...'
+    cat > ~/.docker/config.json << 'DOCKER_CONFIG'
+{
+    \"auths\": {}
+}
+DOCKER_CONFIG
+    
+    # Stop old containers
+    echo 'Stopping existing services...'
+    docker compose -f docker-compose.yml -f docker-compose.prod.yml down --remove-orphans
+    
+    # Start new containers
+    echo 'Starting services with new image...'
+    docker compose -f docker-compose.yml -f docker-compose.prod.yml up -d
+    
+    # Wait for services to start
+    echo 'Waiting for services to start...'
+    sleep 10
+    
+    # Run database migrations
+    echo 'Running database migrations...'
+    docker compose -f docker-compose.yml -f docker-compose.prod.yml exec -T backend uv run alembic upgrade head || echo 'Migration failed or already up to date'
+    
+    # Show service status
+    echo 'Service status:'
+    docker compose -f docker-compose.yml -f docker-compose.prod.yml ps
+"
+
+# Health check
+echo -e "${BLUE}Performing health check...${NC}"
+sleep 20
+if ssh -p "${SERVER_PORT}" "${SERVER_USER}@${SERVER_HOST}" "curl -f http://localhost/health" >/dev/null 2>&1; then
+    echo -e "${GREEN}✓ Deployment successful! Service is healthy.${NC}"
+else
+    echo -e "${YELLOW}⚠ Service deployed but health check failed. Check logs:${NC}"
+    echo "ssh -p ${SERVER_PORT} ${SERVER_USER}@${SERVER_HOST} 'cd ${DEPLOY_PATH} && docker compose logs backend caddy'"
+fi
+
+echo -e "${GREEN}Deployment completed!${NC}"
+echo ""
+echo "Service URLs:"
+if [ "\$DOMAIN" != "localhost" ]; then
+    echo "  Main site: https://${SERVER_HOST}"
+    echo "  API: https://${SERVER_HOST}/api"
+    echo "  LiveKit: https://${SERVER_HOST}/livekit"
+else
+    echo "  Main site: http://${SERVER_HOST}"
+    echo "  API: http://${SERVER_HOST}/api"
+    echo "  LiveKit: http://${SERVER_HOST}/livekit"
+fi
+echo ""
+echo "Useful commands:"
+echo "  Check logs: ssh ${SERVER_USER}@${SERVER_HOST} 'cd ${DEPLOY_PATH} && docker compose logs -f'"
+echo "  Service status: ssh ${SERVER_USER}@${SERVER_HOST} 'cd ${DEPLOY_PATH} && docker compose ps'"
+echo "  Restart: ssh ${SERVER_USER}@${SERVER_HOST} 'cd ${DEPLOY_PATH} && docker compose restart'"
--- a/uv.lock
+++ b/uv.lock