PRegEx: Write Human-Readable Regular Expressions

Motivation

Imagine you are trying to find all URLs inside a text. Each of these URLs must:

• Start with either http:// or https:// or the domain of the URL
• End with either .com or .org

You might end up writing a complicated regular expression(RegEx) like the below:

This RegEx is difficult to read and create. Is there a way that you can write a more human-readable RegEx with ease?

That is when PRegEx comes in handy.

What is PRegEx?

PRegEx is a Python package that allows you to construct RegEx patterns in a more human-friendly way.

To install PRegEx, type:

pip install pregex

The version of PRegEx that will be used in this article is 2.0.1:

pip install pregex==2.0.1

To learn how to use PRegEx, let’s start with some examples.

Capture URLs

Get a Simple URL

First, we will try to get a URL in a text using PRegEx.

from pregex.core.classes import AnyButWhitespace
from pregex.core.quantifiers import OneOrMore
from pregex.core.operators import Either

text = "You can find me through GitHub https://github.com/khuyentran1401"

pre = (
    "https://"
    + OneOrMore(AnyButWhitespace())
    + Either(".com", ".org")
    + OneOrMore(AnyButWhitespace())
)

Output:

['https://github.com/khuyentran1401']

In the code above, we use:

• AnyButWhitespace() to match any character except for whitespace characters
• OneOrMore() to match the provided pattern one or more times.
• Either to match either one of the provided patterns.

Specifically,

• OneOrMore(AnyButWhitespace()) matches one or more characters that are not whitespace characters.
• Either(".com", ".org) matches either .com or .org .

HTTP or HTTPS

Sometimes, a URL might use the scheme http instead of https . Let’s make the character s optional by using Optional() :

from pregex.core.quantifiers import Optional

text = "You can find me through GitHub http://github.com/khuyentran1401"

pre = (
    "http"
    + Optional("s")
    + "://"
    + OneOrMore(AnyButWhitespace())
    + Either(".com", ".org")
    + OneOrMore(AnyButWhitespace())
)
pre.get_matches(text)

Output:

['http://github.com/khuyentran1401']

Match URL without a Scheme

Some URLs in a text might not include a scheme such at https or http . Let’s make the scheme optional with Optional .

To make our code more readable, we will assign PRegrex’s methods to a variable.

text = "You can find me through my website codecut.ai/ or GitHub https://github.com/khuyentran1401"

at_least_one_character_except_white_space = OneOrMore(AnyButWhitespace())
optional_scheme = Optional("http" + Optional("s") + "://")
domain_choice = Either(".com", ".org", ".ai")

pre = (
    optional_scheme
    + at_least_one_character_except_white_space
    + domain_choice
    + at_least_one_character_except_white_space
)
pre.get_matches(text)

Output:

['codecut.ai/', 'https://github.com/khuyentran1401']

Let’s take a look at the underlying RegEx pattern:

pre.get_pattern()

(?:https?\:\/\/)?[^\s]+(?:\.com|\.org)[^\s]+

We have just avoided creating a complicated pattern with some human-readable lines of code!

Capture Time

AnyDigit() matches any numeric character. Let’s use this to match a time in a text.

from pregex.classes import AnyDigit

text = "It is 6:00 pm now"
pre = AnyDigit() + ":" + AnyDigit()
pre.get_matches(text)

['6:0']

Right now, we only match one digit on either side of : . Let’s make this more general by wrapping AddLeastOnce() around AnyDigit() :

pre = OneOrMore(AnyDigit()) + : + AtLeastOnce(AnyDigit())
pre.get_matches(text)

['6:00']

Capture Phone Numbers

Common formats for a phone number are:

##########
###-###-####
### ### ####
###.###.####

These formats either have punctuation or nothing between numbers. We can use AnyFrom("-", " ", ".") to match either - , . , or space.

We also use Optional() to make punctuation optional.

text = "My phone number is 3452352312 or 345-235-2312 or 345 235 2312 or 345.235.2312"

punctuation = AnyFrom("-", " ", ".")
optional_punctuation = Optional(punctuation)
at_least_one_digit = OneOrMore(AnyDigit())

pre = (
    at_least_one_digit
    + optional_punctuation
    + at_least_one_digit
    + optional_punctuation
    + at_least_one_digit
)
pre.get_matches(text)

['3452352312', '345-235-2312', '345 235 2312', '345.235.2312']

Capture an Email Address

Now let’s utilize what we have learned so far to capture an email address from a text.

text = "My email is abcd@gmail.com"

pre = (
    OneOrMore(AnyButWhitespace())
    + "@"
    + OneOrMore(AnyButWhitespace())
    + Either(".com", ".org", ".io", ".net")
)

pre.get_matches(text)

Output:

['abcd@gmail.com']

Next Step

This article gives you an overview of how to use PRegEx to match complicated patterns without spending hours on them.

I encouraged you to check out PRegEx’s documentation for other useful methods.

Feel free to play and fork the source code of this article here.