Add @idata(iterable) decorator-function.

[ankostis]

The purpose is to replace the construct below:

@data(*[a for a in foo if a > 0])

With this typing & memory-efficient alternative:

@data(a for a in foo if a > 0)

(*) It is a minor change that should not interfere with the possible re-write of the library in other branches.

[ankostis]

The flake8 fails in the TC class is unrelated to my modifications to the code.

[codecov-io]

Current coverage is 100% (diff: 100%)

Merging #40 into master will not change coverage

@@           master   #40   diff @@
===================================
  Files           1     1          
  Lines         104   106     +2   
  Methods         0     0          
  Messages        0     0          
  Branches        0     0          
===================================
+ Hits          104   106     +2   
  Misses          0     0          
  Partials        0     0          

Powered by Codecov. Last update 2d0a689...347fe29

[txels]

Sorry, this fell off my radar.
Looks good. The flake8 issue is solved in master, so a rebase would fix it.

I'd love to see some test and documentation.

[pradyunsg]

May I take charge of this PR and submit a new PR the same feature with tests and documentation?

[ankostis]

Of course!

[txels]

BTW in terms of memory efficiency: since ddt is not (nor wants to be) a test runner, in order for the test runner to work properly it has to make the class look like it has one different test method per data value - these extra methods are created dynamically at class load time, not at test execution time. Meaning that the whole generator data will have to be materialised in memory anyway, hence there will not be the typical memory savings of other scenarios where generators are used.

If you need to test with large datasets, ddt is not going to be a memory efficient approach.

[txels]

@ankostis FYI this is my current thinking about where to take this library, that use iterable test data as the base concept #44 (comment)

[ankostis]

@txels rebased, and all tests pass, as promised.
Thanks for the explanation of the no-memory-gains for this PR,
so we are only left with the typing gains :-)

[txels]

Released in 1.1.1

Sorry about the delay in releasing.