ARMS Agent

An LLM agent that standardizes legacy biomedical metadata records to adhere to the CEDAR template.

It fetches the live CEDAR template and queries BioPortal for canonical terms through Model Context Protocol tools, so the constraints it applies are the ones the template holds right now.

This is the agent described in Automated Standardization of Legacy Biomedical Metadata Using an Ontology-Constrained LLM Agent (arXiv:2604.08552). The evaluation harness, the experiment dataset, and the code for the data analysis between baseline vs ARMS agent in the project repository.

Install

pip install arms-agent

Configure

Three keys are required. Put them in the environment, or in a .env file in the directory you run from:

OPENAI_API_KEY=...       # LLM calls
CEDAR_API_KEY=...        # fetching CEDAR templates
BIOPORTAL_API_KEY=...    # ontology term lookups

Optional: set OPENAI_BASE_URL to route LLM calls through an OpenAI-compatible gateway.

To trace each LLM call, tool call, and agent step to Langfuse, install the extra and set both keys:

pip install 'arms-agent[tracing]'

LANGFUSE_PUBLIC_KEY=...
LANGFUSE_SECRET_KEY=...
LANGFUSE_HOST=...       # optional, defaults to Langfuse Cloud

Tracing stays off until both keys are set, and LANGFUSE_TRACING_ENABLED=false switches it off while leaving the keys in place.

Command line

arms-migrate \
  --input legacy-metadata.json \
  --target-schema https://repo.metadatacenter.org/templates/[CEDAR-TEMPLATE-UUID] \
  --output standardized-metadata.json \
  --model gpt-5-mini

--output takes a file or a directory. Given a directory, the filename comes from the input (default: the system's temp directory. --model defaults to gpt-5.6-terra. Add --debug for step-by-step logging on stderr.

Integration in Python

import asyncio, json

from langchain_core.messages import HumanMessage

from arms_agent.agent import build_migration_agent, build_response_format
from arms_agent.prompts import SYSTEM_PROMPT
from arms_agent.tools import all_tools
from arms_agent.workflow import build_workflow

template_iri = "https://repo.metadatacenter.org/templates/[CEDAR-TEMPLATE-UUID]"
legacy = json.load(open("legacy-metadata.json"))

agent = build_migration_agent(
    model="gpt-5-mini",
    system_prompt=SYSTEM_PROMPT,
    response_format=build_response_format(template_iri),
    tools=all_tools,
    reasoning_effort="high",
)

result = asyncio.run(
    build_workflow(agent).ainvoke(
        {
            "messages": [
                HumanMessage(
                    content=(
                        "Standarize the legacy metadata record to adhere to the CEDAR template.\n\n"
                        f"CEDAR Template IRI: {template_iri}\n\n"
                        f"Legacy metadata:\n```json\n{json.dumps(legacy, indent=2)}\n```"
                    )
                )
            ],
            "cedar_template_iri": template_iri,
        },
        config={"recursion_limit": 30},
    )
)
print(json.dumps(result["metadata"], indent=2))

The agent answers against a JSON schema built from the template, so the result conforms to the template's field structure. When a model answers without a validated object, a fixed extraction step parses the text into one.

Caching

CEDAR template and BioPortal term responses are cached in SQLite for 24 hours, to keep repeated runs fast and off the rate limits. Override with ARMS_CACHE_DIR and ARMS_CACHE_TTL_SECONDS in the .env file.

Other settings

Costs are local estimates from provider-reported token counts, not billed amounts.

License

BSD 2-Clause.