Haystack components

Reusable component APIs for Haystack RAG pipelines

Reusable components for building custom Haystack RAG pipelines.

AgenticRouter

Route and orchestrate RAG with agent-like behavior for tool selection and self-reflection.

Constructor

AgenticRouter(
    model: str = "llama-3.3-70b-versatile",
    api_key: str | None = None,
    api_base_url: str = "https://api.groq.com/openai/v1"
)

model · str · default: "llama-3.3-70b-versatile"

LLM model name for routing decisions

api_key · str

API key for the LLM provider. Falls back to GROQ_API_KEY environment variable

api_base_url · str · default: "https://api.groq.com/openai/v1"

Base URL for OpenAI-compatible API endpoint

Methods

select_tool

Select the best tool for a given query.

select_tool(query: str) -> str

query · str · required

User query to route

tool · str

Selected tool name: “retrieval”, “web_search”, “calculation”, or “reasoning”

evaluate_answer_quality

Evaluate generated answer quality with relevance, completeness, and grounding scores.

evaluate_answer_quality(
    query: str,
    answer: str,
    context: str = ""
) -> dict[str, Any]

query · str · required

Original user query

answer · str · required

Generated answer to evaluate

context · str · default: ""

Retrieved context used to generate the answer (for grounding check)

evaluation · dict[str, Any]

Dictionary containing:

• relevance (0-100): Does it answer the query?
• completeness (0-100): Is it sufficiently detailed?
• grounding (0-100): Is it grounded in the context?
• issues (list): List of identified problems (max 3)
• suggestions (list): List of improvement suggestions (max 2)

self_reflect_loop

Run self-reflection loop to iteratively improve answer quality.

self_reflect_loop(
    query: str,
    answer: str,
    context: str = "",
    max_iterations: int = 2,
    quality_threshold: int = 75
) -> str

query · str · required

Original user query

answer · str · required

Initial answer to refine

context · str · default: ""

Retrieved context for grounding

max_iterations · int · default: "2"

Maximum refinement iterations

quality_threshold · int · default: "75"

Target quality score (0-100) to stop refinement early

refined_answer · str

Final refined answer after iterative improvement

---

ContextCompressor

Compress and summarize retrieved context to reduce token usage and improve answer quality.

Constructor

ContextCompressor(
    model: str = "llama-3.3-70b-versatile",
    api_key: str | None = None
)

model · str · default: "llama-3.3-70b-versatile"

LLM model for compression (Groq API)

api_key · str

Groq API key. Falls back to GROQ_API_KEY environment variable

Methods

compress

Compress context using specified compression technique.

compress(
    context: str,
    query: str,
    compression_type: str = "abstractive",
    **kwargs: Any
) -> str

context · str · required

Retrieved context to compress

query · str · required

Original query for relevance filtering

compression_type · str · default: "abstractive"

Compression technique: “abstractive”, “extractive”, or “relevance_filter”

**kwargs · Any

Additional parameters:

• max_tokens (int): For abstractive compression (default: 2048)
• num_sentences (int): For extractive compression (default: 5)
• relevance_threshold (float): For relevance filtering (default: 0.5)

compressed · str

Compressed context text

compress_abstractive

Abstractive compression using LLM summarization.

compress_abstractive(
    context: str,
    query: str,
    max_tokens: int = 2048
) -> str

context · str · required

Context to compress

query · str · required

Query for relevance

max_tokens · int · default: "2048"

Maximum tokens in compressed output

summary · str

Compressed context summary

compress_extractive

Extractive compression by selecting key sentences.

compress_extractive(
    context: str,
    query: str,
    num_sentences: int = 5
) -> str

context · str · required

Context to compress

query · str · required

Query for relevance

num_sentences · int · default: "5"

Number of sentences to extract

selected · str

Selected sentences joined together

filter_by_relevance

Filter context chunks by relevance threshold.

filter_by_relevance(
    context: str,
    query: str,
    relevance_threshold: float = 0.5
) -> str

context · str · required

Context to filter

query · str · required

Query for relevance scoring

relevance_threshold · float · default: "0.5"

Minimum relevance score (0-1) to keep chunks

filtered · str

Filtered context containing only relevant chunks

---

QueryEnhancer

Enhance queries using multi-query generation and query expansion techniques.

Constructor

QueryEnhancer(
    model: str = "llama-3.3-70b-versatile",
    api_key: str | None = None
)

model · str · default: "llama-3.3-70b-versatile"

LLM model for query enhancement

api_key · str

API key. Falls back to GROQ_API_KEY environment variable

Methods

generate_multi_queries

Generate multiple query variations for better retrieval coverage.

generate_multi_queries(
    query: str,
    num_queries: int = 3
) -> list[str]

query · str · required

Original query

num_queries · int · default: "3"

Number of query variations to generate

queries · list[str]

List of query variations including the original

---

ResultMerger

Merge and deduplicate results from multiple retrieval sources.

Methods

merge_results

Merge document results using specified strategy.

merge_results(
    results_list: list[list[Document]],
    strategy: str = "rrf",
    top_k: int = 10
) -> list[Document]

results_list · list[list[Document]] · required

List of result lists from different sources

strategy · str · default: "rrf"

Merging strategy: “rrf” (Reciprocal Rank Fusion) or “interleave”

top_k · int · default: "10"

Number of results to return after merging

merged · list[Document]

Merged and deduplicated document list