pipe operator doesn't work in plain python prompt

[ultrasaw]

seems that the pipe operator doesn't work when using datar in virtual anaconda environments.
here's a snippet of running the example code ran in anaconda prompt:

from datar.all import f, mutate, filter, if_else, tibble

[2021-08-03 19:57:12][datar][WARNING] Builtin name "filter" has been overriden by datar.

df = tibble(
    x=range(4),
    y=['zero', 'one', 'two', 'three']
)

C:\ProgramData\Anaconda3\envs\conda_start\lib\site-packages\pipda\utils.py:159: UserWarning: Failed to fetch the node calling the function, call it with the original function.
warnings.warn(

df >> mutate(z=f.x)

C:\ProgramData\Anaconda3\envs\conda_start\lib\site-packages\pipda\utils.py:159: UserWarning: Failed to fetch the node calling the function, call it with the original function.
warnings.warn(
Traceback (most recent call last):
File "", line 1, in
File "C:\ProgramData\Anaconda3\envs\conda_start\lib\site-packages\pipda\register.py", line 396, in wrapper
return calling_rule(generic, args, kwargs, envdata)
File "C:\ProgramData\Anaconda3\envs\conda_start\lib\site-packages\pipda_calling.py", line 93, in verb_calling_rule3
return generic(*args, **kwargs)
File "C:\ProgramData\Anaconda3\envs\conda_start\lib\functools.py", line 872, in wrapper
raise TypeError(f'{funcname} requires at least '
TypeError: _not_implemented requires at least 1 positional argument

mutate(df, z=f.x)

>         x        y       z
>   <int64> <object> <int64>
> 0       0     zero       0
> 1       1      one       1
> 2       2      two       2
> 3       3    three       3

pandas 1.2.3
python 3.8.1

ps thanks a lot for the package, hopefully the issue can be closed soon :)

[pwwang]

What's the OS? Windows? Linux?
What's the python prompt? python? ipython? bpython?
What's the version of datar? import datar; print(datar.__version__)

[ultrasaw]

What's the OS? Windows? Linux?

windows 10

What's the python prompt? python? ipython? bpython?

anaconda prompt (anaconda3) - python.exe 3.8.1

What's the version of datar? import datar; print(datar.__version__)

0.4.2

[pwwang]

pipda, the backend of datar to implement the piping syntax, relies on the source code to find the AST node to detect the environment of the function to be called (whether it's in piping syntax or regular function calling).

However, the source code is not available when the code is running in plain python prompt, thus datar couldn't tell if a verb is called in piping syntax.

There are several solutions for this:

1. Use a 3-rd party prompt (ipython, bpython) to run the code
2. Run the code in as a script file (python script.py)
3. Use a regular way to call the verbs (mutate(df, z=f.x))

[ultrasaw]

I'm using PyCharm with Anaconda + plain python prompt, will try switching to ipython then;

thanks so much for your quick answers!

[pwwang]

You are welcome! Thank you for reporting!

pipda (v0.4.4) has a new mode pipda.options.assume_all_piping = True, which handles the case where source code is not available. With this mode, the verbs are assumed to be called using piping syntax.

datar will adopt this mode in the next version.

[pwwang]

With this revision 9bd9d02, one could use the "all piping" mode (all verbs are supposed to be used in piping syntax):

from datar.all import f, mutate, filter, if_else, tibble
from pipda import options

options.assume_all_piping = True

df = tibble(
    x=range(4),
    y=['zero', 'one', 'two', 'three']
)

df >> mutate(z=f.x)
# However, without source code available, we can't have both of the above and below work:
# mutate(df, z=f.x)

[ultrasaw]

thanks a lot for making it all clear!

is there a way to pip install the dev branch of the module to already have the 9bd9d02 revision?

I'm new to github

The question is simply out of curiousity, everything works for me now that I installed ipython 🍪

[pwwang]

You can do:

pip install git+git://github.com/pwwang/datar.git@9bd9d0253a3f313e0cf5ce0fb04b6192573bb11e

But I don't recommend it. It's still in development and may need more tests.

[pwwang]

Closeing. As 0.4.3 is now able run the assume_all_piping mode.
See: #45 (comment)