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

or by adding another command like map <pipeline>

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

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

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

$ pype map 'x.split()[0] ! x.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

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

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

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

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

$ pype eval 1+1
2

$ 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 20 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 6 seconds. Use amap, or afilter, or reduce with await some_async_function to get concurrency out of the box.

$ time pype amap '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

Async streaming

amap and afilter values are handled in streaming fashion, while retaining the order of the input items in the output. The order of function calls is not constrained – if you need the function to be called with items in a specific order, use the synchronous version.

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 of the input items is maintained in the output.

$ 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.

$ pype map 'Counter ! json.dumps' <<<'hello\nworld\n'
{"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 set a different default value for the concurrency maximum pype --max-concurrent, add max_concurrent to your config file (note the underscore):

# ~/.config/pype/config.toml

max_concurrent = 10

then just use pype as normal.

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]]

command = "map"
options = []
arguments = [ "json.loads ! types.SimpleNameSpace(**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

Plugins

Add new commands like map and reduce by installing pype plugins. You can try them out without installing by adding them to any .py file in your ~/.config/pype/modules/.