studyPipe: Convenient pipe in python | Pipe pratique en python

Based On sspipe/sspipe itself based on JulienPalard/Pipe
Based On /pytoolz/toolz (toolz.curried)
Based On kieferk/dfply

Table of Contents

• studyPipe: Convenient pipe in python | Pipe pratique en python
  • studyPipe's vocabulary:
  • Install:
  • Usage:
  • Imported pipes:
  • Curried Functions
  • Comments

Example: functions in the re module (regex)
Without studyPipe

import re
list(filter(lambda e:re.match("^[^_]",e),map(str.lower,dir(re))))

With studyPipe

dir(re) | _ftools_.map(  __.lower() ) | _ftools_.filterl( _fun_.re.match("^[^_]") )
#OR
re |_fun_| dir | _ftools_.map(  __.lower() ) | _ftools_.filterl(("^[^_]",__) |_funs_| re.match )

Example: from sspipe

np.linspace(0, 2*np.pi, 100) | px[np.cos(px) < 0] | p(plt.plot, px, np.sin(px), 'r')

with studyPipe

np.linspace(0, 2*np.pi, 100) | __[np.cos(__) < 0] | _fun_.plt.plot(__, np.sin(__), 'r');
#OR
np.linspace(0, 2*np.pi, 100) | __[np.cos(__) < 0] | ((__, np.sin(__), 'r') |_funs_| plt.plot);
#OR
np.linspace(0, 2*np.pi, 100) | __[np.cos(__) < 0] | ( plt.plot |_funsInv_| (__, np.sin(__), 'r') );

Exemple 2: from sspipe
with sspipe

(
  [0, pi/6, pi/2]
  | p(map, lambda x: '{:.3f}'.format(sin(x)))
  | p(list) 
  | p(' '.join) 
  | p(print, "Example 5: using builtin map transform and lambda", px)
)
>>> Example 5: using builtin map transform and lambda 0.000 0.500 1.000

with studyPipe (and without |_fun_|)

([0, pi/6, pi/2] 
    | _ftools_.mapl(_fun_.str('{:.3f}').format(sin(__)))
    | _fun_.str(" ").join
    | _fun_.print("Example 5: using builtin map transform and lambda", __)
)
>>> Example 5: using builtin map transform and lambda 0.000 0.500 1.000

with studyPipe (and with |_fun_|)

([0, pi/6, pi/2] 
    | _ftools_.mapl(sin(__) |_fun_| '{:.3f}'.format )
    |_fun_| " ".join
    | _fun_.print("Example 5: using builtin map transform and lambda", __)
)
#OR
([0, pi/6, pi/2] 
    | _ftools_.mapl( sin(__) )
    | _ftools_.format('{:.3f}')
    | _ftools_.join(" ")
    | _fun_.print("Example 5: using builtin map transform and lambda", __)
)
>>> Example 5: using builtin map transform and lambda 0.000 0.500 1.000

studyPipe's vocabulary:

__: placeholder (px in sspipe) (two underscore)

•   np.array(range(100)) | (__-__.mean())/__.std()

_fun_: before a function (replace p in sspipe) (one underscore before, one underscore after)

•   "Hello World" | _fun_.print  

  (same as `print("Hello World")

•   "Hello World" | _fun_.print(__,end=' !!!!\n')

  (same as print("Hello World",end=' !!!!\n'))

• "Hello World" | _fun_.print(__.lower(),end='!' * _fun_.len(__) + '\n')

  (same as print("Hello World".lower(),end='!' * len("Hello World") + '\n'))

|_fun_|: between pipes when the next element is an function (without parameters)

• "Hello World" |_fun_| print

  (same as print("Hello World"))

• "studyPipe" |_fun_| "Hello World {} !!".format |_fun_| print

  (same as print("Hello World {} !!".format("studyPipe")))

•   __ |_fun_| "Hello World {} !!".format |_fn_| print

  (same as lambda x: print("Hello World {} !!".format(x)))

_ftools_: curried's functions from toolz.curried and JulienPalard/Pipe (one underscore before, one underscore after)

•   ["Helloee","World", "blablabla"] | _ftools_.mapl( __ |_fun_| len )

  (same as list(map(len,["Helloee","World", "blablabla"])))

•   ["Helloee","World", "blablabla", "dd", "22h"] | _ftools_.filterl( __ |_fun_| len > 4 )

  (same as list(filter(lambda x:len(x) > 4,["Helloee","World", "blablabla", "dd", "22h"])))
• More on: JulienPalard/Pipe
• More on: /pytoolz/toolz

|_funs_|: when the next element is a function and you want to expand an iterable as function's parameters (next function without parameters)

(
    [["Hello",[0,1,2,3,4]], ["World",[5,6,7,8]]]  
    | _ftools_.mapl((  __[0].upper(), 
                       __[1] | _ftools_.concat2(', '),
                       __[1] |_fun_| sum  ) 
                    |_funs_| "{} : {} (sum: {})".format
                   ) 
    |_fun_| '\n'.join |_fun_|  print
)
>>> HELLO : 0, 1, 2, 3, 4 (sum: 10)
... WORLD : 5, 6, 7, 8 (sum: 26)

same as

    print(
        '\n'.join(
            map(lambda x:"{} : {} (sum: {})".format(x[0].upper(),
                                                    ', '.join(map(str,x[1])),
                                                    sum(x[1])),
                [["Hello",[0,1,2,3,4]], ["World",[5,6,7,8]]] 
               )
        )
    )
>>> HELLO : 0, 1, 2, 3, 4 (sum: 10)
... WORLD : 5, 6, 7, 8 (sum: 26)

|_funsInv_| : when the next element is an iterable (tuple, list, ...) and you want to expand them as function's parameters (the element before is a function without parameters)

(
    os.listdir() 
    | _ftools_.where(__.endswith(".py")) 
    | _ftools_.sorted(key=len)
    | _ftools_.enumerate
    | _ftools_.mapl( "{}. {}".format |_funsInv_|  __ )  
    | _ftools_.join("\n") 
    | _fun_.print
)
#OR
(
    os.listdir() 
    | _ftools_.where(__.endswith(".py")) 
    | _ftools_.sorted(key=len)
    | _ftools_.enumerate
    | _ftools_.mapl( "{}. {}" )  
    | _ftools_.join("\n") 
    | _fun_.print
)
#OR
(
    os.listdir() 
    | _ftools_.where(__.endswith(".py")) 
    | _ftools_.sorted(key=len)
    | _ftools_.enumerate
    | _ftools_.format( "{}. {}" )  
    | _ftools_.join_n
    | _fun_.print
)
>>> 0. 03_classification_script.py
... 1. 06_decision_trees_script.py
... 2. 03_classification-Copy1_script.py
... 3. 04_training_linear_models_script.py
... 4. 01_the_machine_learning_landscape_script.py
... 5. 02_end_to_end_machine_learning_project_script.py
... 6. 07_ensemble_learning_and_random_forests_script.py
... 7. 07_ensemble_learning_and_random_forests_modif_script.py

(
    os.listdir() 
    | _ftools_.where(__.endswith(".py")) 
    | _ftools_.sorted(key=len)
    | _ftools_.enumerate
    | _ftools_.mapl( "{}. {} ({})".format |_funsInv_| (__[0], __[1], len |_funsInv_| (__[1],) ) )  
    | _ftools_.join_n
    | _fun_.print
)
#OR
(
    os.listdir() 
    | _ftools_.where(__.endswith(".py")) 
    | _ftools_.sorted(key=len)
    | _ftools_.enumerate
    | _ftools_.mapl( "{}. {} ({})".format |_funsInv_| ( __[0], __[1], _fun_.len(__[1]) ))  
    | _ftools_.join_n
    | _fun_.print
)
>>> 0. 03_classification_script.py (27)
... 1. 06_decision_trees_script.py (27)
... 2. 03_classification-Copy1_script.py (33)
... 3. 04_training_linear_models_script.py (35)
... 4. 01_the_machine_learning_landscape_script.py (43)
... 5. 02_end_to_end_machine_learning_project_script.py (48)
... 6. 07_ensemble_learning_and_random_forests_script.py (49)
... 7. 07_ensemble_learning_and_random_forests_modif_script.py (55)

Install:

pip install git+https://github.com/luluperet/studyPipe.git

Usage:

#Configuration:  
from studyPipe import config
config.globalsFn=lambda:globals() # Alow studyPipe to know your variables, modules, ....

#Import
from studyPipe import __, _fun_, _funs_, _ftools_, _funsInv_

Imported pipes:

px and p from sspipe (also as _px and _p, also as px_ and p_)
X from dfply (also as _X and X_)
dfply as df (also as _df and df_)
curried from toolz (also as _c or c_)
T and F: for True and False

Curried Functions

All functions available in JulienPalard/Pipe have a twin function that calls the list function on the result of the original function: theses functions ends by l
wherel, selectl, ....
More: mapl, zipl, filterl, rangel, enumeratel, sorted, join, enumeratel, enumerate, sorted, join_n, join_comma, format_sep, format_n, format_comma

Comments

when _funs_(_funsInv_) is not between pipes (|_funs_|, |_funsInv_|), it act same as _fun_. The only difference between _fun_ and _funs (_funsInv) it's when they are between pipes, otherwise they are the same

When __ is alone before |_funs_| they are some problems..., you have to use |_funsInv_| Ex (with |_funs_|):

(
    [["Hello",[0,1,2,3,4]], ["World",[5,6,7,8]]]  
    | _ftools_.mapl( __ |_funs_| "{} : {} ".format ) 
)
>>> [<sspipe.pipe.Pipe at 0x7fa233c63618>, <sspipe.pipe.Pipe at 0x7fa233c63348>]

(with |_funsInv_|)

(
    [["Hello",[0,1,2,3,4]], ["World",[5,6,7,8]]]  
    | _ftools_.mapl(  "{} : {} ".format |_funsInv_| __ ) 
)
>>> ['Hello : [0, 1, 2, 3, 4] ', 'World : [5, 6, 7, 8] ']