jaat-api 0.2.0

Official lightweight client wrapper library for the JAAT API.

jaat-api

The official, lightweight Python client library for the JAAT suite of tools.

jaat-api is an easy-to-use, intuitive, session-managed SDK for programtically accessing the open-source JAAT modules. Securely send text batches to the openly hosted JAAT modules.

---

Installation

Install directly from PyPI:

pip install jaat-api

Quick Start

Initialize the JAATClient with your API key (provided by us) and execute JAAT on your batch texts seamlessly:

from jaat_api import JAATClient

client = JAATClient(api_key="jaat_app_private_key_here")

job_ads = [
    "[JOB AD TEXT 1]",
    "[JOB AD TEXT 2]",
    "..."
]

task_results = client.run_batch("task", job_ads) # alternatively, client.get_tasks(job_ads)

# continue to run other jobs
skill_results = client.run_batch("skill", job_ads)

Helpful Tips

Handling Exceptions

The SDK provides a structured JAATError instance whenever an operation breaks down or fails. Ideally, wrap jobs within try/except blocks to handle these conditions gracefully:

from jaat_api import JAATError

try:
    ## as before
except JAATError as e:
    print(f"Extraction failed (Status Code: {e.status_code})")
    print(f"Error Context: {e.message}")

Sandbox Public Evaluation

Anyone looking to evaluate the platform programmatically before contacting us for a dedicated token can use the public sandbox key:

client = JAATClient(api_key="jaat_public_sandbox_testing_key")

try:
    results = client.get_tasks(["JOB AD TEXT"])
    print(results)
except JAATError as e:
    if e.status_code == 429:
        print("Sandbox daily evaluation volume exhausted. Please rotate to a private key.")

As can be seen, the public API key is highly throttled and is strictly for testing purposes. The current usage of this key is limited to one request per day, per IP.

License

jaat-api (and JAAT) are distributed under the MIT License. See LICENSE for more information.

Getting a Key

We will soon be releasing a contact form to let us know your interest in receiving a private API key. Stay tuned!

Citation

This code and the overarching JAAT package is part of ongoing work to leverage lightweight, but powerful NLP tools to unlock structured, coded data from job posting data.

If you find this software useful, please consider citing:

@article{meisenbacher2025extracting,
  title={Extracting O* NET Features from the NLx Corpus to Build Public Use Aggregate Labor Market Data},
  author={Meisenbacher, Stephen and Nestorov, Svetlozar and Norlander, Peter},
  journal={arXiv preprint arXiv:2510.01470},
  year={2025}
}