python3 support ?

[stonebig]

import pandas as pd
from ply import install_ply, X, sym_call
install_ply(pd)

gives

---------------------------------------------------------------------------
ImportError                               Traceback (most recent call last)
<ipython-input-1-f35c480251ef> in <module>()
      1 import pandas as pd
----> 2 from ply import install_ply, X, sym_call
      3 
      4 install_ply(pd)

D:\result_tests\WinPython-64bit-3.4.2.3_build3\python-3.4.2.amd64\lib\site-packages\ply\__init__.py in <module>()
----> 1 from methods import install_ply
      2 from symbolic import X, sym_call

ImportError: No module named 'methods'

[stonebig]

it seems a "."-like thing

from .methods import install_ply
from .symbolic import X, sym_call

and

from . import symbolic

and then a few parenthesitizing

print ()

[stonebig]

see #2 and #3

After that I still have something I dont understand when I use "where"

(flights
  .groupby(['Year', 'Month', 'DayofMonth'])
  .ply_select(
    arr = X.ArrTime.mean(),
    dep = X.DepTime.mean())
  .ply_where(X.Year >= X.Year))
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-13-58a798c52b76> in <module>()
      4     arr = X.ArrTime.mean(),
      5     dep = X.DepTime.mean())
----> 6   .ply_where(X.Year >= X.Year))

TypeError: unorderable types: GetAttr() >= GetAttr()

No clue there

[stonebig]

after a little reduce effort ply_where alone works :

flights.ply_where(flights.ArrTime < 1500, flights.ArrTime >1400).head()

this works:

(flights
  .groupby(['Year', 'Month', 'DayofMonth'])
  .ply_select(
    arr = X.ArrTime.mean(),
    dep = X.DepTime.mean())
  )

but still no chance in combination:

(flights
  .groupby(['Year', 'Month', 'DayofMonth'])
  .ply_select(
    arr = X.ArrTime.mean(),
    dep = X.DepTime.mean())
  .ply_where(X.arr < 1500, X.dep >1400))

gives:

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-3-0cbf68ca069f> in <module>()
      4     arr = X.ArrTime.mean(),
      5     dep = X.DepTime.mean())
----> 6   .ply_where(X.arr < 1500, X.dep >1400))

TypeError: unorderable types: GetAttr() < int()

[jhorowitz-coursera]

Yep, use of operators on symbolic expressions (like symbolicExpression + 3) fails in Python 3. It turns out that this is because the current system needs Python to call symbolicExpression.__getattr__('__add__') at some point in the evaluation of symbolicExpression + 3, and for new-style classes (the only ones available in Python 3), this call is skipped. See https://docs.python.org/2/reference/datamodel.html#new-style-special-lookup for details on that.

The solution is to add the special methods (like __add__) directly to ply.symbolic.Expression. Kinda ugly, but that's life.

I'll take care of this. Thanks for all your debugging work!

[jhorowitz-coursera]

@stonebig -- I just pushed a new version which should fix these issues. (I decided to drop in a version of https://pypi.python.org/pypi/six.) The relevant changes are in 115c79e. Please run

pip install pandas-ply --upgrade

and let me know if everything works ok.

(Thanks again for your help.)

[stonebig]

everything works ok.