numberly / flask-stupe Goto Github PK

Trying to run our README's example without PyMongo installed throws:

Traceback (most recent call last):
  File "run.py", line 2, in <module>
    from flask_stupe import paginate, schema_required
ImportError: cannot import name 'paginate'

This error is completely normal, because we only expose things that are usable based on the different modules available. But it's not very developer-friendly and an error message explaining that paginate is not available because pymongo isn't installed would be much better.

See pallets/flask#3111. Some changes should be made here, starting with the README example, which is no longer up to date.

https://gitlab.com/pgjones/quart

Using our @paginate decorator 😏

As we can see here: http://flask.pocoo.org/docs/0.12/api/

Today flask prettify the json output unless:

This function’s response will be pretty printed if it was not requested with X-Requested-With: XMLHttpRequest to simplify debugging unless the JSONIFY_PRETTYPRINT_REGULAR config parameter is set to false. Compressed (not pretty) formatting currently means no indents and no spaces after separators.

We don't want to pass that http header each time!

When doing pagination it is nice to have http headers see https://www.vinaysahni.com/best-practices-for-a-pragmatic-restful-api#pagination and https://tools.ietf.org/html/rfc5988#page-6

These make life much easier than having to roll things yourself which can lead to mistakes

I ran the example in README.md and caught this error while POST-ing data to /user endpoint with the payload:

{
  "username": "itsjef",
  "password": "itsme"
}

It was this line: https://github.com/numberly/flask-stupe/blob/master/flask_stupe/validation.py#L19 that caused the bug, since schema.load() was an instance method.

My workaround is to replace

@schema_required(UserSchema)

with

@schema_required(UserSchema())

Additional Info:

• python version: 3.6.5
• flask-stupe version: 3.6.4
• pymongo: 3.7.1
• marshmallow: 2.15.4

https://docs.travis-ci.com/user/deployment/pypi/