Configuration

AIECS uses environment variables for configuration. You can set these in a .env file or as system environment variables.

Environment Variables

LLM Provider Configuration

OpenAI

OPENAI_API_KEY=your_openai_api_key
OPENAI_API_BASE=https://api.openai.com/v1  # Optional
OPENAI_MODEL=gpt-4  # Default model

Google Vertex AI

VERTEX_PROJECT_ID=your_gcp_project_id
VERTEX_LOCATION=us-central1  # Default location
GOOGLE_APPLICATION_CREDENTIALS=/path/to/credentials.json

xAI

XAI_API_KEY=your_xai_api_key
XAI_API_BASE=https://api.x.ai/v1

Database Configuration

PostgreSQL

POSTGRES_HOST=localhost
POSTGRES_PORT=5432
POSTGRES_DB=aiecs
POSTGRES_USER=aiecs_user
POSTGRES_PASSWORD=your_password
POSTGRES_POOL_SIZE=10  # Connection pool size
POSTGRES_MAX_OVERFLOW=20  # Max overflow connections

Redis Configuration

REDIS_HOST=localhost
REDIS_PORT=6379
REDIS_DB=0
REDIS_PASSWORD=  # Optional
REDIS_MAX_CONNECTIONS=50

Celery Configuration

CELERY_BROKER_URL=redis://localhost:6379/0
CELERY_RESULT_BACKEND=redis://localhost:6379/0
CELERY_TASK_SERIALIZER=json
CELERY_RESULT_SERIALIZER=json
CELERY_ACCEPT_CONTENT=["json"]
CELERY_TIMEZONE=UTC
CELERY_ENABLE_UTC=true

Google Cloud Storage

GCS_BUCKET_NAME=your_bucket_name
GCS_PROJECT_ID=your_gcp_project_id
GOOGLE_APPLICATION_CREDENTIALS=/path/to/credentials.json

Server Configuration

HOST=0.0.0.0
PORT=8000
DEBUG=false
LOG_LEVEL=INFO
WORKERS=4  # Number of worker processes

Tool Configuration

Document Parser Tool

DOC_PARSER_GCS_PROJECT_ID=your_project_id
DOC_PARSER_GCS_BUCKET_NAME=your_bucket_name
DOC_PARSER_TEMP_DIR=/tmp/doc_parser
DOC_PARSER_MAX_FILE_SIZE=104857600  # 100MB in bytes

Document Creator Tool

DOC_CREATOR_TEMPLATES_DIR=/path/to/templates
DOC_CREATOR_OUTPUT_DIR=/path/to/output
DOC_CREATOR_DEFAULT_FORMAT=markdown
DOC_CREATOR_DEFAULT_STYLE=default
DOC_CREATOR_AUTO_BACKUP=true

Web Scraper Tool

WEB_SCRAPER_USER_AGENT=AIECS/1.5.3
WEB_SCRAPER_TIMEOUT=30
WEB_SCRAPER_MAX_RETRIES=3
WEB_SCRAPER_ENABLE_JAVASCRIPT=false

Configuration File

You can also use a configuration file for more complex setups.

Example `.env` File

# LLM Providers
OPENAI_API_KEY=sk-...
VERTEX_PROJECT_ID=my-project-123

# Database
POSTGRES_HOST=localhost
POSTGRES_PORT=5432
POSTGRES_DB=aiecs
POSTGRES_USER=aiecs_user
POSTGRES_PASSWORD=secure_password

# Redis
REDIS_HOST=localhost
REDIS_PORT=6379

# Celery
CELERY_BROKER_URL=redis://localhost:6379/0
CELERY_RESULT_BACKEND=redis://localhost:6379/0

# Google Cloud
GCS_BUCKET_NAME=my-aiecs-bucket
GOOGLE_APPLICATION_CREDENTIALS=/path/to/credentials.json

# Server
HOST=0.0.0.0
PORT=8000
DEBUG=false
LOG_LEVEL=INFO

Programmatic Configuration

You can also configure AIECS programmatically:

from aiecs.config.config import Settings

# Create custom settings
settings = Settings(
    openai_api_key="your_key",
    postgres_host="localhost",
    postgres_port=5432,
    redis_host="localhost"
)

# Use settings in your application
from aiecs.llm.llm_integration import LLMIntegration

llm = LLMIntegration(
    provider="openai",
    api_key=settings.openai_api_key
)

Best Practices

Never commit secrets: Keep your .env file out of version control
Use environment-specific configs: Maintain separate configs for dev, staging, and production
Validate configuration: Use the built-in validation tools to check your configuration
Secure credentials: Use secret management services in production
Monitor resources: Configure appropriate pool sizes and timeouts based on your workload