wharris / dougrain Goto Github PK

Wondering if you've considered using collections.OrderedDict to make key order somewhat predictable.

JSON Hypertext Application Language Internet Draft 4 has been published but Dougrain is still stuck on Draft 3.

Draft 4: http://tools.ietf.org/html/draft-kelly-json-hal-04
Diffs between Draft 3 and Draft 4: http://tools.ietf.org/rfcdiff?url2=draft-kelly-json-hal-04.txt

I don't think the spec has a rule for this, but I could easily be wrong on this. Currently, when using Document.embed, you get a different data structure depending on whether you call it once or more often. If you iterate over, for example, a database result and want to provide a unified interface independent of the result length, you would have to have make a check like this on the server:

if len(result) == 1:
    doc.embed('rel', [result])
else:
    [doc.embed('rel', r) for r in result]

I'd like to have a flag on embed to always wrap the object in a list, even on the first call. What do you think?

Dougrain does not support draft 5 of the HAL spec.

The main change in draft 5 is to make explicit the Hypertext Cache Pattern, embedded resources are merely advice to the client and should not replace links.

Servers SHOULD NOT entirely "swap out" a link for an embedded
resource (or vice versa) because client support for this technique is
OPTIONAL.

Dougrain should ensure that an equivalent link exists whenever a resource is embedded in a document.

Draft 5: http://tools.ietf.org/html/draft-kelly-json-hal-05
Diffs between draft 4 and draft 5: http://tools.ietf.org/rfcdiff?url2=draft-kelly-json-hal-05.txt

_links often contain URI templates[http://tools.ietf.org/html/rfc6570],

This is what I do at the moment:

the_url = uritemplate.expand( mydoc.links['myrel'].url(), { "a": b, "c": d, })

I'd like to have this functionality built into dougrain, so I can write

the_url = mydoc.links['myrel'].url(a=b, c=d)

Also, I'd have to re-read the HAL spec to figure out if there are situations where the values can be extrapolated from the document itself, so the url could expand automatically.

https://pypi.python.org/pypi/uritemplate is a great URI template library I've worked with.

I have the impression, that CURIE does actually not work, they only expand to the correct url if you have a "base_uri" specified which prefixes everything.