kpbs / py3-pmp-wrapper Goto Github PK

They were in the repo and must have been lost after rebase across merges. They would caught the NavigableDoc.get method error.

Write tests for navigabledoc module.

When requesting the home-doc, the client returns <Navigable Doc: None>, which is confusing. It should use some logic to know that the homedoc has no navigation, or for pages without navigation, it sets the current page url to last requested.

An upcoming release to the PMP will introduce some changes to home document links. Changes have already been deployed to https://api-sandbox.pmp.io. Here's my synopsis of what needs to happen for this SDK:

Breaking changes:

• none

Cleanup changes:

• regenerate fixtures
• urn:collectiondoc:query:guids should be removed from tests/fixtures
• urn:collectiondoc:form:profilesave should be removed from tests/fixtures
• urn:collectiondoc:form:schemasave should be removed from tests/fixtures

Possible upgrades:

• urn:collectiondoc:form:documentdelete can now be used to get doc-delete endpoint/methods
• urn:collectiondoc:form:listcredentials + createcredentials + removecredentials for the cred endpoints/methods
• urn:collectiondoc:form:revoketoken for un-issuing a token (though this isn't a common workflow)

The following exception pops on a bad query:

>>> doc.query("urn:collectiondoc:query:documents", params={'tag': 'npr_api', 'profile': 'story'})
requests.exceptions.MissingSchema: Invalid URL 'None': No schema supplied. Perhaps you meant http://None?

This is not that helpful. The query urn is not present in the document. Check for None before attempting to make a request and alert the user that the URN is bad before requests creates this confusing exception.

Need test coverage for utils.json_utils.

Queries may be failing after gain_access method called if token has expired. Could be PmpAuth.get_access_token2? Review and create some edge cases with full integration tests, likely.

It says to use query_template which has been renamed to template. The class' docstring has been fixed in feature branch, but review docstrings on all methods.

Consult PMP docs for reference.

Homedoc is an unprotected endpoint in order to serve token information. However, when it is requested before gain_access method, a NoToken exception is raised even though a token is unnecessary for requesting homedoc.

Maybe bypass NoToken for requesting homedoc?

Seems like work-flow should be: instantiate Client, call gain_access. After that, everything else is possible. Would addressing this lead to a potentially confusing work-flow?

The following should be a valid query param as a per PMP docs:

cl.query('urn:collectiondoc:query:docs', params={'tag': 'npr_api,APM'})
Traceback (most recent call last):
  File "<console>", line 1, in <module>
  File "~/.virtualenvs/someproject/lib/python3.3/site-packages/py3-pmp-wrapper/pmp_api/pmp_client.py", line 124, in query
    return self.get(self.document.query(rel_type, params=params))
  File "~/.virtualenvs/someproject/lib/python3.3/site-packages/py3-pmp-wrapper/pmp_api/pmp_client.py", line 113, in get
    results = self.connector.get(endpoint)
  File "~/.virtualenvs/someproject/lib/python3.3/site-packages/py3-pmp-wrapper/pmp_api/core/conn.py", line 92, in get
    raise BadRequest(errmsg)
pmp_api.core.exceptions.BadRequest

However, it raises a BadRequest. Does it not receive a valid response? Is the status-code bad? If so, it shouldn't be listed as BadRequest, but there should be some alternate response.

Params with a semicolon pass validation, but those with a comma do not. Please address this issue so boolean values are respected in parameters or error message provides more information as to what went wrong.

Document's get method not working:

document.get('tags', '')
Traceback (most recent call last):
  File "<console>", line 1, in <module>
TypeError: get() takes 1 positional argument but 3 were given

This is because parameters are not specified in the method definition:

def get(self):
        return self.collectiondoc.get

Whereas dict's get method would accept params. Copy params from dict's get method or use_args, *_kwargs`.

It fails inconsistently, but why? Investigate and repair.

FAIL: test_paging_docs_doc (test_pager.TestPager)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "~/.virtualenvs/SOMEPROJ/lib/python3.3/site-packages/py3-pmp-wrapper/test/test_pager.py", line 38, in test_paging_docs_doc
    self.assertEqual(actual_vals['current'], self.pager.current)
nose.proxy.AssertionError: 'http://127.0.0.1:8080/docs?offset=10' != 'http://127.0.0.1:8080/docs?guid=someGUID'
- http://127.0.0.1:8080/docs?offset=10
?                             ^^^ ^^^^
+ http://127.0.0.1:8080/docs?guid=someGUID
?                            ++++++ ^ ^^^^

    """Fail immediately, with the given message."""
>>  raise self.failureException("'http://127.0.0.1:8080/docs?offset=10' != 'http://127.0.0.1:8080/docs?guid=someGUID'\n- http://127.0.0.1:8080/docs?offset=10\n?                             ^^^ ^^^^\n+ http://127.0.0.1:8080/docs?guid=someGUID\n?                            ++++++ ^ ^^^^\n")