PDFData Python client

Standalone client for https://pdfdata.org/api/v1. Requires Python 3.9+ and requests; Flask, MongoDB and scientific packages are not client dependencies. Version 0.2.0 source and build configuration are in this repository. A public PyPI release has not been made as part of this change.

Install from a local checkout (repository root):

python -m pip install ./src/pdfdb_backend

Or install the built wheel supplied with a release:

python -m pip install /path/to/pdfdata_api-0.2.0-py3-none-any.whl

Generate a key on the website's API page after signing in. Set it in an environment variable and query:

import os
from pdfdata_api import PDFData_API

with PDFData_API(api_key=os.environ["PDFDATA_API_KEY"]) as client:
    page = client.get_materials(elements="Si,O", per_page=20)
    material = client.get_material("mp-10009-X")
    curve = client.get_pdf_data("mp-10009-X")
    all_matches = client.get_all_materials(elements="Si,O")

get_materials() and search() return one page by default. Page size is 1–100. get_all_materials() and search_all() collect successive pages. Legacy client calls with per_page=0 also collect successive pages; direct HTTP requests must use page sizes 1–100. Collection is not a database snapshot while data changes.

Advanced search:

result = client.search({
    "elements": {"all": ["Si"], "none": ["Au"]},
    "qmax": "[10,30]",
    "page": 1,
    "per_page": 20,
})

For applications already holding a website access JWT, PDFData_API() can call generate_api_key(jwt_token), get_api_keys(jwt_token) and delete_api_key(jwt_token, "Key 1"). The generated key is set on the client. Key management requires the login JWT; material queries require the API key. generate_api_key creates a new key on each call and is not automatically retried.

MaterialsAPI remains an alias of PDFData_API. The old import from PDFData_API import PDFData_API remains supported. Backend source imports from app.api.client import MaterialsAPI forward to the same implementation. The default timeout is 30 seconds and can be set with timeout=60. HTTP errors raise requests.HTTPError; inspect exc.response.status_code and JSON code.

Build from this directory:

python -m build

Only pdfdata_api and the legacy PDFData_API module are included in the distribution. Production environment files and backend code are excluded.