[python-client] Add model default values

[spacether]

PR checklist

• Read the contribution guidelines.
• Ran the shell script under ./bin/ to update Petstore sample so that CIs can verify the change. (For instance, only need to run ./bin/python-petstore.sh and ./bin/security/python-petstore.sh if updating the {LANG} (e.g. php, ruby, python, etc) code generator or {LANG} client's mustache templates). Windows batch files can be found in .\bin\windows\.
• Filed the PR against the correct branch: master, 3.4.x, 4.0.x. Default: master.
• Copied the technical committee to review the pull request if your PR is targeting a particular programming language.
  • @taxpon
  • @frol
  • @mbohlool
  • @cbornet
  • @kenjones-cisco
  • @tomplus
  • @wing328

Description of the PR

This PR lets python client users set model defaults on integers, numbers, and arrays when they are not currently able to in a swagger2.0 spec. The new code loads model variable examples into the model variable default value if the default is not already set and an example exists for the variable.
Fixes this issue: #1748

Reasons to do this:
The current swagger-parser is not including default values that are float, integer, or array when processing a swagger2.0 spec. So in a swagger 2.0 spec, there is no way to set defaults on:

• integers
• numbers
• arrays

This PR adds in fallback code which uses and assigns a spec example value as the default value if one exists in the spec.
So it uses examples to set defaults on the types

• string
• boolean
• integers
• numbers
• arrays

Tests

• test_type_holder_example.py added to verify that examples are loaded into defaults
• test_type_holder_default.py added to verify what default-only definition does and does not work
• samples updated to pass CircleCi tests (173 files)

[wing328]

What about using Boolean.valueOf instead of using equalsIgnoreCase?

[spacether]

Boolean.valueOf(string) will always return True or False,
so Boolean.valueOf("5") would return False.
When I pass in "5" as a boolean the code should return null. Code updated use equalsIgnoreCase to check for "true" and "false" otherwise return null.
I'd prefer matching on true false strings rather than truthy "0" "5" "1.234" "[1,2,3]" because truthiness differs from language.

equalsIgnoreCase lets us only look for a string match to true or false ignoring case.

[wing328]

When I pass in "5" as a boolean the code should return null

If someone puts down 5 as the example value and assume it will be treated false in boolean, it's the user's issue (not something we need to handle) and I don't think it should return null (clearly there's a value defined)

We've seen spec using 1 or 0 to as example value for boolean type (not entirely correct as they should be using "true" or "false" instead) but still we want to able to handle these cases correctly as some languages (e.g. Perl) treat 1 as true.

[spacether]

Coercing it to a boolean gives the user back some boolean which is usefull. I will change it to Boolean.valueOf. This behavior may surprise python users though because they will be using a python client with java casting of string truthiness.
"1" -> True in Python, false in Java
"0" -> True in Python, false in Java

[wing328]

@spacether the change looks good to me. Thanks for your contribution.

[wing328]

@spacether If the default value is not defined, it shouldn't fall back to the example value.

I'll file a PR to remove the fallback.

[spacether]

Good to know. Thank you you @wing328 sorry for the breaking change.

[wing328]

@spacether no worries. In the templates, you can do something similar to the following to have the default value fall back to example value (not something I recommend but ok if there's a valid use case)

{{{defaultValue}}}{{^defaultValue}}{{{example}}}{{/defaultValue}}

[wing328]

FYI. Filed #2213