Swarmauri Standard

swarmauri_standard provides first-party, ready-to-use Swarmauri components built on swarmauri_core interfaces and swarmauri_base component behavior. It is the standard component bundle for agents, chains, chunkers, conversations, documents, tools, toolkits, parsers, prompts, schema converters, metrics, similarities, deprecated distance compatibility shims, vectors, vector stores, tracing utilities, STT/TTS/VLM adapters, signing helpers, and key-provider components.

Why Swarmauri Standard?

swarmauri_standard provides maintained first-party components that users can install and instantiate directly. It also demonstrates the expected Swarmauri component style: typed base classes, direct class imports, dynamic serialization, and namespace-ready registration.

FAQ

Q: What kinds of components are included?

A: The package includes agents, chains, chunkers, conversations, documents, tools, toolkits, parsers, prompts, schema converters, metrics, similarities, deprecated distance compatibility shims, vectors, vector stores, tracing utilities, STT/TTS/VLM adapters, signing helpers, and key-provider components.

Q: Is this package the same as swarmauri?

A: No. swarmauri provides the namespace importer and plugin discovery layer. swarmauri_standard provides concrete first-party implementation classes.

Q: Should examples instantiate plugins directly?

A: Yes. Direct class imports and direct instantiation are the preferred documented workflow unless a package explicitly requires another integration pattern.

Features

• Standard agents including question-answering, RAG, simple conversation, and tool-using agent patterns.
• Tool and toolkit components such as calculator, HTTP request, readability, code extraction, temperature conversion, weather, and accessibility utilities.
• Parser and document components for CSV, XML, OpenAPI specs, markdown-to-HTML, URLs, regular expressions, Python code, and phone-number extraction.
• Chunkers, conversations, prompts, prompt templates, chains, pipelines, service registries, state objects, and tracing helpers for application workflows.
• Deprecated distance compatibility shims, vectors, vector stores, measurements, metrics, norms, similarities, and related numerical components.
• Model adapter classes for LLM, tool LLM, embedding, image generation, STT, TTS, OCR, and VLM workflows where provider configuration is available.
• Signing and key-provider components that align with Swarmauri's security-oriented interfaces.
• Dynamic JSON, YAML, and TOML serialization through ComponentBase, DynamicBase, and SubclassUnion.
• Python 3.10, 3.11, 3.12, 3.13, and 3.14 support.

Installation

Install with uv:

uv add swarmauri_standard

Install with pip:

pip install swarmauri_standard

Usage

Use standard components directly from their implementation modules:

from swarmauri_standard.tools.CalculatorTool import CalculatorTool

calculator = CalculatorTool()
result = calculator("add", 5, 3)

assert result == {"operation": "add", "calculated_result": "8.0"}

Create documents and serialize them as Swarmauri components:

from swarmauri_standard.documents.Document import Document

doc = Document(content="Composable intelligence infrastructure")
payload = doc.model_dump_json()
restored = Document.model_validate_json(payload)

assert restored.type == "Document"
assert restored.content == doc.content

Hydrate registered standard components through a base-family union:

from pydantic import BaseModel

from swarmauri_base.DynamicBase import SubclassUnion
from swarmauri_base.toolkits.ToolkitBase import ToolkitBase
from swarmauri_standard.toolkits.Toolkit import Toolkit


class ToolkitEnvelope(BaseModel):
    toolkit: SubclassUnion[ToolkitBase]


payload = ToolkitEnvelope(toolkit=Toolkit()).model_dump_json()
restored = ToolkitEnvelope.model_validate_json(payload)

assert isinstance(restored.toolkit, Toolkit)

Component Families

The package includes these standard component families:

• Agents: QAAgent, RagAgent, SimpleConversationAgent, and ToolAgent.
• Chains and workflow: callable chains, chain steps, context chains, prompt context chains, pipelines, service registries, swarms, and round-robin task management.
• Text and document processing: chunkers, documents, conversations, prompts, prompt templates, parsers, messages, and schema converters.
• Tools and toolkits: calculator, arithmetic, HTTP requests, JSON requests, code extraction, code interpretation, readability scoring, weather, temperature conversion, parameters, accessibility toolkit, and generic toolkit support.
• Vector and numerical work: vectors, TF-IDF vector store, deprecated distance compatibility shims, measurements, metrics, norms, similarities, pseudometrics, seminorms, matrices, tensors, and inner products.
• AI model surfaces: LLMs, tool LLMs, embeddings, image generators, OCR, STT, TTS, and VLM classes for supported providers.
• Runtime support: decorators, tracing, transports, state, logging, and utility classes.
• Trust and security surfaces: signing and key-provider components that align with Swarmauri's core interfaces.

Distance Migration

The swarmauri_standard.distances module remains available only as a deprecated compatibility surface. Prefer the active metric and similarity families instead:

• CosineDistance -> CosineSimilarity
• EuclideanDistance -> EuclideanMetric
• LevenshteinDistance -> LevenshteinMetric
• JaccardIndexDistance -> JaccardIndexSimilarity
• SorensenDiceDistance -> DiceSimilarity
• ChebyshevDistance -> SupremumMetric
• ManhattanDistance -> LpMetric(p=1)
• MinkowskiDistance -> LpMetric(p=<order>)
• SquaredEuclideanDistance -> EuclideanMetric or GaussianRBFSimilarity, depending on whether you need metric or similarity semantics
• CanberraDistance, ChiSquaredDistance, and HaversineDistance do not yet have active drop-in replacements in the workspace

Related Packages

Foundational packages:

• swarmauri provides namespace imports and plugin discovery.
• swarmauri_core provides the interfaces implemented by these standard components.
• swarmauri_base provides component base classes, mixins, and serialization behavior.
• swarmauri_typing provides dynamic typing utilities used by the component model.

Related component packages:

• swarmauri_embedding_doc2vec for Doc2Vec embedding support.
• swarmauri_embedding_nmf for NMF embedding support.
• swarmauri_vectorstore_doc2vec for Doc2Vec vector-store workflows.
• swarmauri_standard for the active metric, similarity, and vector-store comparator surface.
• swarmauri_tool_matplotlib for Matplotlib tool support.
• swarmauri_parser_keywordextractor for keyword extraction parsing.
• swarmauri_signing_ed25519 for Ed25519 signing.

When To Use This Package

Use swarmauri_standard when you want maintained first-party implementations that are ready for examples, prototypes, tests, and production workflows where the included component behavior fits. Use smaller standalone component packages when you want a narrower dependency surface for a specific provider, algorithm, or runtime integration.

License

Apache-2.0

Contributing

When adding standard components, inherit from the matching swarmauri_base class, satisfy the swarmauri_core interface, register the component type, document direct instantiation, and follow the Swarmauri SDK contribution guide.