red

Python regex command-line tool, to replace functionality akin to 'perl -ne'.

Installation

From the project root directory::

$ python setup.py install

Usage

Use --help/-h to view info on the arguments::

$ red --help

The special variables available:

:line: for the entire line that was matched, without the trailing newline :g: for groups, which you index to get that match group, as g[0] in "(\w+) ." :ag: for aggregated groups with the -a option, which is a list of all match groups :d: for the groupdict, which you can index by named group as d["foo"] in "(?P\w+) ." :ad: for the aggregated group dicts, a list of dicts

Example usage::

$ cat test.txt 
foo 1 bar 2
fiz 5 baz 10
funk 10 bunk 9
funk a bunk b
a b c d
aaaaa
bbbb
cc

Use it like grep::

$ cat test.txt | red "\w+ (\d+) \w+ (\d+)" 
foo 1 bar 2
fiz 5 baz 10
funk 10 bunk 9

It works with path as argument as well, but the path must appear directly after the regex (oddity of argparse).::

$ red "\w+ (\d+) \w+ (\d+)" test.txt

Use it to evaluate Python code on groups stored in variable g::

$ red "\w+ (\d+) \w+ (\d+)" test.txt -e "int(g[0]) + int(g[1])"
3
15
19

Import an arbitrary library and do absolutely anything::

$ red '\w+ (\d+) \w+ (\d+)' test.txt -i json -e '"{} => {}".format(line, json.dumps(g))'
foo 1 bar 2 => ["1", "2"]
fiz 5 baz 10 => ["5", "10"]
funk 10 bunk 9 => ["10", "9"]

If you want to execute a few python statements before the eval, you can do that as well with -x::

$ red "(\w+) (\d+).*" test.txt -x 'x = int(g[1]) ; y = g[0][::-1]' -e '(x, y)'
(1, 'oof')
(5, 'zif')
(10, 'knuf')

Use it to aggregate across all of stdin, into list ag::

$ cat test.txt | red "\w+ (\d+) \w+ (\d+)" -a "sum([int(x[0]) for x in ag])"
16

Aggregate has its own -X for an exec before the aggregate too::

$ red "(\w+) (\d+).*" test.txt -X 's = sum(int(g[1]) for g in ag)' -a 's'
16

Evaluate on each match, and aggregate against all matches::

$ cat test.txt | red "\w+ (\d+) \w+ (\d+)" -a "sum([int(x[0]) for x in ag])" -e "'adding {}'.format(g[0])"
adding 1
adding 5
adding 10
16

You can use named groups as well, stored in variables d and aggregated into ad::

$ cat test.txt | red "\w+ (?P<first>\d+) \w+ \d+" -e "'first value is {first}'.format(**d)"
first value is 1
first value is 5
first value is 10

Even multiline will work, since whatever is passed into -x is just exec'd::

$ red '.*' test.txt -x '
# if you want multiline, just hit apostrophe and press enter
# and start typing
if line.startswith("foo"):
     print("Line started with foo: {}".format(line))
'

Line started with foo: foo 1 bar 2

Get creative!::

$ cat urls.txt 
https://www.google.com/
http://www.yahoo.com/
http://www.example.com/foo

$ red "(.*)" urls.txt -i requests -x 'response = requests.get(line)' -e '[response.status_code, response.content[:20]]'
[404, '<!doctype html>\n<htm']
[200, '<!doctype html><html']
[200, '<?xml version="1.0" ']

$ cat somelogfile.log 
INFO: visited http://www.example.com/foo
ERROR: visited https://www.google.com/
ERROR: visited http://www.yahoo.com/

$ cat somelogfile.log | red '(?P<debuglevel>ERROR|INFO): \S+ (?P<url>.*)' -i requests,collections -x 'response = requests.get(d["url"])' -e '(response.status_code, response.content[0:20])' -a 'collections.Counter([d["debuglevel"] for d in ad]).items()'
(404, '<!doctype html>\n<htm')
(200, '<!doctype html><html')
(200, '<?xml version="1.0" ')
[('INFO', 1), ('ERROR', 2)]

Disclaimer

Code in -e and -a are run through eval, so if you paste in code from someone, it will run. Just sayin'. Only run code you trust, or have inspected personally. The same goes for any library you import with -i.

Release Notes

:0.2.1: Added -x for pre-eval exec, added -i for importing libraries, and -X for aggregate exec :0.2.0: Added -i/--import functionality :0.1.1: Path works as CLI arg :0.1.0: Version is available on pypi, with functionality of evaluation and aggregation :0.0.1: Project created