In Python client, produce URIs to search results

[danielballan]

A long-standing aim of Tiled is to provide URLs to arbitrary units of data. The URLs can be passed to curl or a data analysis program that speaks HTTP, shared with colleagues, associated with DOIs in publications, etc.

It would be nice if the Python client provided an easy way to obtain URLs that point to search results. We produce URLs to nodes, but not to their contents, filtered or otherwise:

In [14]: c['nmc_sim']
Out[14]: <Container {'22a5JHY7z5w', '22ACa4ELm3h', '22eMren2rdQ', ...} ~16467 entries>

In [15]: c['nmc_sim'].uri
Out[15]: 'https://aimm.lbl.gov/api/v1/metadata/nmc_sim'

In [16]: c['nmc_sim'].search(Key('sim_code') == 'fdmnes')
Out[16]: <Container {'22a5JHY7z5w', '22ACa4ELm3h', '22eMren2rdQ', ...} ~14740 entries>

In [17]: c['nmc_sim'].search(Key('sim_code') == 'fdmnes').uri
Out[17]: 'https://aimm.lbl.gov/api/v1/metadata/nmc_sim' <--- not what I wanted!

[danielballan]

It's worth pointing out that you can handle the URIs to from_uri which will "navigate" directly to that node:

from_uri('https://aimm.lbl.gov/api/v1/metadata/nmc_sim')

I think the task here, then, would be two-fold:

1. Expose a contents_uri (name negotiable) that points to a /search/ endpoint, which may have filtering query parameters.
2. Teach from_uri what to do with a /search/ URL.