Configuration Guide¶

This document provides detailed information about configuring Ragatanga for your specific needs.

Configuration Methods¶

Ragatanga can be configured in several ways:

Environment Variables: Set configuration through environment variables
Configuration Files: Use .env files for persistent configuration
Programmatic Configuration: Configure components directly in code
Configuration Objects: Use configuration classes for fine-grained control

Environment Variables¶

Core Configuration¶

Variable	Description	Default	Example
`RAGATANGA_LOG_LEVEL`	Logging level	`INFO`	`DEBUG`
`RAGATANGA_CACHE_DIR`	Directory for caching	`~/.ragatanga/cache`	`/tmp/ragatanga_cache`
`RAGATANGA_MAX_RESULTS`	Maximum results to return	`5`	`10`
`RAGATANGA_CONFIDENCE_THRESHOLD`	Minimum confidence score	`0.6`	`0.8`

Embedding Configuration¶

Variable	Description	Default	Example
`RAGATANGA_EMBEDDING_PROVIDER`	Embedding provider	`openai`	`sentence_transformers`
`RAGATANGA_EMBEDDING_MODEL`	Embedding model name	`text-embedding-ada-002`	`all-MiniLM-L6-v2`
`RAGATANGA_EMBEDDING_BATCH_SIZE`	Batch size for embeddings	`32`	`64`
`RAGATANGA_EMBEDDING_CACHE`	Enable embedding cache	`true`	`false`

LLM Configuration¶

Variable	Description	Default	Example
`RAGATANGA_LLM_PROVIDER`	LLM provider	`openai`	`anthropic`
`RAGATANGA_LLM_MODEL`	LLM model name	`gpt-3.5-turbo`	`claude-2`
`RAGATANGA_LLM_TEMPERATURE`	LLM temperature	`0.0`	`0.7`
`RAGATANGA_LLM_MAX_TOKENS`	Maximum tokens for LLM	`1024`	`2048`

API Keys¶

Variable	Description	Example
`OPENAI_API_KEY`	OpenAI API key	`sk-...`
`ANTHROPIC_API_KEY`	Anthropic API key	`sk-ant-...`
`HUGGINGFACE_API_KEY`	HuggingFace API key	`hf_...`

Configuration File¶

You can create a .env file in your project root with the following format:

# Core Configuration
RAGATANGA_LOG_LEVEL=INFO
RAGATANGA_CACHE_DIR=~/.ragatanga/cache
RAGATANGA_MAX_RESULTS=5
RAGATANGA_CONFIDENCE_THRESHOLD=0.6

# Embedding Configuration
RAGATANGA_EMBEDDING_PROVIDER=openai
RAGATANGA_EMBEDDING_MODEL=text-embedding-ada-002
RAGATANGA_EMBEDDING_BATCH_SIZE=32
RAGATANGA_EMBEDDING_CACHE=true

# LLM Configuration
RAGATANGA_LLM_PROVIDER=openai
RAGATANGA_LLM_MODEL=gpt-3.5-turbo
RAGATANGA_LLM_TEMPERATURE=0.0
RAGATANGA_LLM_MAX_TOKENS=1024

# API Keys
OPENAI_API_KEY=your-openai-api-key

Programmatic Configuration¶

Configuring the Ontology Manager¶

from ragatanga.core.ontology import OntologyManager
from ragatanga.core.config import OntologyConfig

# Create configuration
ontology_config = OntologyConfig(
    reasoning_level="RDFS",
    cache_materialized=True,
    prefixes={
        "ex": "http://example.org/",
        "schema": "http://schema.org/"
    }
)

# Initialize with configuration
ontology_manager = OntologyManager(
    "path/to/ontology.ttl",
    config=ontology_config
)

Configuring the Embedding Manager¶

from ragatanga.core.embeddings import EmbeddingManager
from ragatanga.core.config import EmbeddingConfig

# Create configuration
embedding_config = EmbeddingConfig(
    provider="sentence_transformers",
    model_name="all-MiniLM-L6-v2",
    batch_size=64,
    cache_embeddings=True
)

# Initialize with configuration
embedding_manager = EmbeddingManager(config=embedding_config)

Configuring the LLM Manager¶

from ragatanga.core.llm import LLMManager
from ragatanga.core.config import LLMConfig

# Create configuration
llm_config = LLMConfig(
    provider="openai",
    model_name="gpt-4",
    temperature=0.2,
    max_tokens=2048,
    system_prompt="You are a helpful assistant specialized in knowledge retrieval."
)

# Initialize with configuration
llm_manager = LLMManager(config=llm_config)

Configuring the Retriever¶

from ragatanga.core.retrieval import AdaptiveRetriever
from ragatanga.core.config import RetrievalConfig

# Create configuration
retrieval_config = RetrievalConfig(
    semantic_search_weight=0.7,
    ontology_search_weight=0.3,
    max_results=10,
    confidence_threshold=0.6,
    use_query_enhancement=True,
    use_multilingual=True
)

# Initialize with configuration
retriever = AdaptiveRetriever(
    ontology_manager,
    embedding_manager=embedding_manager,
    llm_manager=llm_manager,
    config=retrieval_config
)

Configuration Classes¶

Ragatanga provides several configuration classes for fine-grained control:

OntologyConfig: Configuration for the ontology manager
EmbeddingConfig: Configuration for embedding generation and storage
LLMConfig: Configuration for language model interactions
RetrievalConfig: Configuration for retrieval operations
APIConfig: Configuration for the API server

See the API Reference for detailed documentation of these classes.