neuromorpho-api

Querying the neuromorpho.org database in Python with requests.

Installation

pip install neuromorpho-api

Quickstart

>>> from neuromorpho_api import requestor as requests

>>> neuron_endpoint = "https://neuromorpho.org/api/neuron/"
>>> resp = requests.get(neuron_endpoint + "fields")
>>> resp.json()
{'Neuron Fields': ['neuron_id',
  'neuron_name',
  'archive',
  'age_scale',
  'gender',
  ...

See the tutorial for more detailed examples, including selecting neurons by attribute pattern matching and downloading SWC traces.

What is this package?

The neuromorpho-api package provides a requests.Session instance with a custom SSLContext needed for interacting with neuromorpho.org.

Why can’t I just use requests directly?

You may see something like the following:

>>> import requests
>>> requests.get("https://neuromorpho.org/api/neuron/id/1")
Traceback (most recent call last)
   ...
SSLError: HTTPSConnectionPool(host='neuromorpho.org', port=443): Max retries exceeded with url: /api/neuron/id/1 (Caused by SSLError(SSLError(1, '[SSL: DH_KEY_TOO_SMALL] dh key too small (_ssl.c:1006)')))

It appears the key provided by neuromorpho.org is incompatible with the default Python SSL cipher. This package works around this issue using Robin De Schepper’s solution posted on Stack Overflow. At some point in the future, the neuromorpho certificate may (hopefully) be updated, at which point this package will be archived.

Future compatibility

The requestor provided by this package is intended as a temporary workaround for the current DH key issue with neuromorpho.org. If the neuromorpho.org certificate is updated, the requestor with the custom SSLContext will no longer be necessary, and users can switch to using requests directly. Therefore, the following import alias is recommended:

from neuromorpho-api import requestor as requests

With an updated certificate, the above can be changed to:

import requests

and all code that depends on requests.get should continue to work as expected.