Skip to main content

Command-line pipes in Python

Project description

Usage

Basics

At the command prompt, use pype to act on each item in the file with python commands:

$ pype map x.upper() <<<'abc'
ABC

Chain python functions together with !:

$ pype map 'x.upper() ! len(x)' <<<hello
5

Use x as a placeholder for the input at each stage:

$ pype map ' x.split() ! x[0].upper() + "!"' <<<'Hello world'
HELLO!

$ pype map 'x.split() ! x[0].upper() + "!" ! x.replace("H", "J")' <<<'Hello world'
JELLO!

Automatically import modules you need:

$ pype stack 'itertools.repeat(x, 2) ! "".join' <<<hello,world!
hello,world!
hello,world!

Commands

map

Use map to act on each input item.

$ pype map 'x * 2' <<<'a\nbb\n'
aa
bbbb

filter

Use filter to evaluate a condition on each line of input and exclude false values.

$  pype filter 'len(x) > 1' <<<'a\nbb\nccc\n'
bb
ccc

apply

Use apply to act on the sequence of items.

$   pype apply 'len(x)' <<<'a\nbb\n'
2

stack

Use stack to treat the input as a single string, including newlines.

$  pype stack 'len(x)' <<<'a\nbb\n'
5

Use eval to evaluate a python expression without any input.

$ pype eval 1+1
2

reduce

Use reduce to evaluate a function of two arguments successively over a sequence, like functools.reduce

$ pype map int reduce operator.mul <<EOF
1
2
3
4
EOF

24

Autocall

You don’t need to explicitly call the function with f(x); just use f. For example, instead of

$ pype map 'len(x)' <<<'a\nbb'
5

try

$ pype map len <<<'a\nbb'
5

Async

Making sequential requests is slow. These requests take 18 seconds to complete.

$ time pype map 'requests.get ! x.text ! len' apply max <<EOF
http://httpbin.org/delay/5
http://httpbin.org/delay/1
http://httpbin.org/delay/4
http://httpbin.org/delay/3
http://httpbin.org/delay/4
EOF

302

0.61s user
0.06s system
19.612 total

Concurrent requests can go much faster. The same requests now take only 5 seconds. Just use await async_function to get concurrency out of the box.

$ time pype map 'await asks.get ! x.text ! len' apply max <<EOF
http://httpbin.org/delay/5
http://httpbin.org/delay/1
http://httpbin.org/delay/4
http://httpbin.org/delay/3
http://httpbin.org/delay/4
EOF

297

0.57s user
0.08s system
5.897 total

Streaming

map and filter values are handled in streaming fashion, while retaining order.

Making concurrent requests, each response is printed one at a time, as soon as (1) it is ready and (2) all of the preceding requests have already been handled.

For example, the 3 seconds item is ready before the preceding 4 seconds item, but it is held until the 4 seconds is ready because 4 seconds was started first, so the ordering is maintained.

$ time pype --exec-before 'import datetime; now=datetime.datetime.utcnow; START_TIME=now(); print("Elapsed time | Response size")' map 'await asks.get !  f"{(now() - START_TIME).seconds} seconds    | {len(x.content)} bytes"'  <<EOF
http://httpbin.org/delay/1
http://httpbin.org/delay/2
http://httpbin.org/delay/4
http://httpbin.org/delay/3
EOF
Elapsed time | Response size
1 seconds    | 297 bytes
2 seconds    | 297 bytes
4 seconds    | 297 bytes
3 seconds    | 297 bytes

Configuration

Add code to automatically execute, into your config file.

For example:

# ~/.config/pype/config.toml

exec_before = """

from itertools import *
from collections import Counter

"""

Then you can directly use the imported objects without referencing the module.

$ printf 'hello\nworld\n' | pype --autocall map 'Counter ! json.dumps'

{"h": 1, "e": 1, "l": 2, "o": 1}
{"w": 1, "o": 1, "r": 1, "l": 1, "d": 1}

You can set any of the pype options in your config. For example, to make --no-autocall the default, add

# ~/.config/pype/config.toml

autocall = false

then just use pype as normal

$ printf 'a\nbb\nccc\n' | pype map 'len'
<built-in function len>
<built-in function len>
<built-in function len>

Aliases

Define new commands in your config file which provide aliases to other commands. For example, this config adds a jsonl command for reading jsonlines streams into Python objects, by calling calling out to the map traversal.

[[alias]]

name = "jsonl"
short_help = "Load jsonlines into python objects."

[[alias.stage]]

name= "map"
options = []
arguments = [ "json.loads ! attr.make_class('X', list(x.keys()))(**x)"]

Now we can use it like a regular command:

$ pype jsonl  <<< $'{"a":1, "b":2}\n{"a": 5, "b":9}'
X(a=1, b=2)
X(a=5, b=9)

The new command jsonl can be used in pipelines as well. To get the maximum value in a sequence of jsonlines objects.

$ pype jsonl map 'x.a' apply max <<< $'{"a":1, "b":2}\n{"a": 5, "b":9}'
5

Installation

Get it with pip:

pip install python-pype

Caveats

  • pype assumes trusted command arguments and untrusted input stream data. It uses eval on your commands, not on the input stream data. If you use exec, eval, subprocess, or similar commands, you can execute arbitrary code from the input stream, like in regular python.

Status

  • Check the issues page for open tickets.

  • This package is experimental pre-alpha and is subject to change.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

python-pype-0.0.77.tar.gz (13.5 kB view hashes)

Uploaded Source

Built Distribution

python_pype-0.0.77-py3-none-any.whl (36.4 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page