ask-orex

ASK-Orex

ASK-Orex is a package designed to simplify regular expressions in Python. It provides a high-level interface for constructing and working with regular expressions, making it easier and more intuitive to use.

The package is heavily inspired by Richy Cotton's rebus package for the R language.

Installation

You can install Orex using pip:

pip install ask-orex

To use Orex, import the ask-orex package into your Python script:

import ask_orex as ox

Creating regular expressions

The easiest regular expression is just a literal string to be found.

s = 'Hello World'
pattern = ox.literal('Hello')
pattern.is_match(s)

True

Orex regular expressions are extended by simply using a +. A slightly more useful example is to find a hex colour

pattern = ox.literal('#') + ox.HEXDIGIT + ox.HEXDIGIT +\
 ox.HEXDIGIT + ox.HEXDIGIT + ox.HEXDIGIT + ox.HEXDIGIT

s_hex = 'This package is #a83232 hot'
pattern.is_match(s_hex)

True

pattern.findall(s_hex)

['#a83232']

On the other hand

s_nonhex = 'Just a twitter handle: #Red123'
pattern.is_match(s_nonhex)

False

pattern.findall(s_nonhex)

[]

Clearly, the pattern is somewhat burdensome. We can alternatively write it like this

pattern = ox.literal('#') + ox.repeat(ox.HEXDIGIT, 6)

or

pattern = ox.literal('#') + ox.n_or_more(ox.HEXDIGIT, min=6, max=6)

Subpatterns

Most of orex functions accept other regular expressions as characters.

Say we want to find a regex to parse emails. We can define first all email allowed characters

allowed_characters = ox.character_class(ox.ALNUM + ox.DASH + '_' + ox.DOT)

where ox.ALNUM are just all letters and numbers. Note that the _ is just used as a string. In Orex it is okay to use strings as part of the pattern, unless they are at the start, where we use ox.literal to show python that we mean pattern business. Whenever we want to use the . or _ explicitly, it is best to use ox.DOT or ox.DASH, since . and - can have a special meaning in regular expression.

Next we define

user_name = ox.one_or_more(allowed_characters)
domain_name = ox.one_or_more(allowed_characters)
extension = ox.n_or_more(ox.ALPHA,2,4)

email_pattern =  user_name + '@' + domain_name + ox.DOT + extension

email = '[e-mail]'

And indeed

email_pattern.is_match(email)

True

Capturing

If we want to extract the individual components of a match, we have to capture them first. Using the email example above, we can just change the code to

email_pattern =  ox.capture(user_name) + '@' + ox.capture(domain_name) \
                + ox.DOT + ox.capture(extension)

email_pattern.findall(email)

returns

[('captainspamalot', 'funnyspammail', 'com')]

Handy indeed! We can also name the individual components for even easier extraction

email_pattern =  ox.capture(user_name,'user') \
              + '@' + ox.capture(domain_name,'domain') \
              + ox.DOT + ox.capture(extension,'ext')

We can now use the group_dict method for even easier access

email_pattern.group_dict(email)

{'user': 'captainspamalot', 'domain': 'funnyspammail', 'ext': 'com'}

Capturing has another benefit, namely that we can use it to find repeated patterns with the ox.backreference function

tag_name = ox.one_or_more(allowed_characters)
content = ox.one_or_more(ox.ANY_CHAR)
tag_pattern = ox.literal('<') + ox.capture(tag_name, 'tag') + '>' \
          + ox.capture(content,'content') + '</' +ox.backreference(name='tag')+'>'

message = '<name>Sir Snakington</name>'
tag_pattern.group_dict(message)

{'tag': 'name', 'content': 'Sir Snakington'}

Back references can also be used to substitute patterns, though here the named backreference does not work. Instead, the backrefence uses the index of the reference in the pattern.

tag_pattern = ox.literal('<') + ox.capture(tag_name) + '>' + ox.capture(content) + '</' +ox.backreference(1)+'>'
replace_pattern = ox.literal('<') + ox.backreference(2) + '>' + ox.backreference(1) + ox.literal('<') + ox.backreference(2) + '>'
tag_pattern.sub(message, replace_pattern)

'<Sir Snakington>name<Sir Snakington>'