swarmauri_tool_entityrecognition 0.9.1.dev2

Swarmauri Community Entity Recognition Tool

Swarmauri Tool Entity Recognition

Named-entity recognition tool for Swarmauri based on Hugging Face transformers. Uses the default pipeline("ner") model to detect tokens labeled as PERSON, ORG, LOC, etc., and returns a JSON-encoded dictionary of entities grouped by label.

Features

• Wraps the transformers NER pipeline in a Swarmauri ToolBase component.
• Auto-downloads the default model on first run (usually dslim/bert-base-NER).
• Aggregates entity tokens by label and returns them as a JSON string in the entities key.

Prerequisites

• Python 3.10 or newer.
• transformers, torch, and associated dependencies (installed automatically). Ensure GPU/CPU compatibility for PyTorch according to your environment.
• Internet access on first run to download model weights.

Installation

# pip
pip install swarmauri_tool_entityrecognition

# poetry
poetry add swarmauri_tool_entityrecognition

# uv (pyproject-based projects)
uv add swarmauri_tool_entityrecognition

Quickstart

import json
from swarmauri_tool_entityrecognition import EntityRecognitionTool

text = "Apple Inc. is an American multinational technology company."
tool = EntityRecognitionTool()
result = tool(text=text)

entities = json.loads(result["entities"])
print(entities)

Example output:

{"B-ORG": ["Apple", "Inc."], "B-MISC": ["American"], "I-MISC": ["multinational"], ...}

Tips

• The default pipeline tokenizes into subwords; reconstruct phrases by joining consecutive tokens with the same label when needed.
• Specify a different model by subclassing and passing pipeline("ner", model="<model>") if you require language-specific NER.
• Cache Hugging Face models (HF_HOME) in CI or container builds to avoid repeated downloads.

Want to help?

If you want to contribute to swarmauri-sdk, read up on our guidelines for contributing that will help you get started.