fix the problem #299 in terms of using tqdm in pandas

[chengs]

• rebase on to master
• Add new test cases about pandas
• fix Remove support for *args in pandas wrappers #299 - find the correct total when using progress_apply and other functions in pandas.

#299 partially achieved its goal. With forbidding *args, one can get correct axis now, and use it to get the total (as number of iterations in progress_apply)

However, pandas sometimes runs several more iterations to optimise itself. For example, in dataframe.progress_apply(func).

• if func is a numpy.ufunc or is very slow, then it will be executed total times.
• but if it is very quick or in some other cases, it will be executed total+1 or total+shape[axis0] times.

I think there is no need to adjust such particular pandas implementation, so in the wrapper function, I write

def wrapper(*args, **kwargs):
                    # update tbar correctly
                    # it seems pandas apply calls func twice on the first column/row 
                    # to decide whether it can take a fast or slow code path.
                    # so stop when t.total==t.n
                    t.update(n=1 if t.total and t.n < t.total else 0)
                    return func(*args, **kwargs)

if total is achieved, tbar will not be updated.

[codecov-io]

Codecov Report

Merging #524 into master will increase coverage by 0.3%.
The diff coverage is 100%.

@@           Coverage Diff            @@
##           master    #524     +/-   ##
========================================
+ Coverage    99.1%   99.4%   +0.3%     
========================================
  Files           8       8             
  Lines         668     674      +6     
  Branches      118     118             
========================================
+ Hits          662     670      +8     
+ Misses          4       3      -1     
+ Partials        2       1      -1

[chengs]

I add new tests, so now, pandas.Series, pandas.Dataframe, and groupby cases are all covered. Not cover pandas.Panel because it is going to be removed. http://pandas.pydata.org/pandas-docs/version/0.20/whatsnew.html#whatsnew-0200-api-breaking-deprecate-panel

[chengs]

this ngroups is an inner attribute of pandas. It may be changed in future, but seems not happen in near future or never. So let's keep it so far.

For perfection, I will suggest to use one of pandas public apis to get group size in future. So far, it is fine.