ezregex 2.1.0

A readable and intuitive way to generate Regular Expressions

EZRegex

A readable and intuitive way to make Regular Expressions without having to know any of the syntax

Try the frontend for this library at ezregex.org!

TLDR: This is to regular expressions what CMake is to makefiles

Table of Contents

• Usage
• Invert
• Generate
• Functions vs Methods
• Dialects
• Documentation
• Developer Docs
• Installation
• Todo
• License
• Credits

Usage

Quickstart

from ezregex import *
'foo' + number + optional(whitespace) + word
# Or, using methods
number.append(whitespace.optional).prepend('foo').append(word)
# Matches `foo123abc` and `foo123 abc`
# but not `abc123foo` or  `foo bar`

Importing as a named package is recommended

import ezregex as er

# ow is part of er already as optional whitespace
params = er.group(er.atLeastNone(er.ow + er.word + er.ow + er.optional(',') + er.ow))
# Seperate parts as variables for cleaner patterns
function = er.word + er.ow + '(' + params + ')'

# Automatically calls the re.search() function for you
function % 'some string containing func( param1 , param2)'

# The test() method is helpful for debugging, and color codes groups for you
function.test('this should match func(param1,\tparam2 ), foo(), and bar( foo,)')

.test() will print all the matches, color coded to match and group (colors not shown here):

╭───────────────────────────── Testing Regex ──────────────────────────────╮
│ Testing expression:                                                      │
│         \w+(?:\s+)?\(((?:(?:\s+)?\w+(?:\s+)?,?(?:\s+)?)*)\)              │
│ for matches in:                                                          │
│         this should match func(param1,  param2 ), foo(), and bar( foo,)  │
│                                                                          │
│ Match = "func(param1,  param2 )" (18:39)                                 │
│ Unnamed Groups:                                                          │
│         1: "param1, param2 " (23:38)                                     │
│                                                                          │
│ Match = "foo()" (41:46)                                                  │
│ Unnamed Groups:                                                          │
│         1: "" (45:45)                                                    │
│                                                                          │
│ Match = "bar( foo,)" (52:62)                                             │
│ Unnamed Groups:                                                          │
│         1: " foo," (56:61)                                               │
│                                                                          │
│                                                                          │
╰───────────────────────────────── Found  ─────────────────────────────────╯

Inverting

The invert function (available as er.invert(expression), expression.invert(), or ~expression) is useful for debugging. You pass it an expression, and it returns an example of a string that is guaranteed to match the provided expression.

Generation

In version v1.7.0 we introduced a new function: generate_regex. It takes in 2 sets of strings, and returns a regular expression that will match everything in the first set and nothing in the second set. It may be a bit crude, but it can be a good starting point if you don't know where to start. It's also really good at regex golf.

Functions vs Methods

As of v2.1.0, elemental methods were added to EZRegex objects. These shadow their function element counterparts exactly and work the same way, they're just for convenience and preference.

For example, these are all equivelent:

# Element functions
optional(whitespace) + group(either(repeat('a'), 'b')) + if_followed_by(word)
# Elemental methods
whitespace.optional.append(literal('a').repeat.or_('b').unnamed).if_followed_by(word)
# Mixed
whitespace.optional + repeat('a').or_('b').unnamed + if_followed_by(word)

Dialects

As of version v1.6.0, the concepts of dialects was introduced. Different languages often have slight variations on the regular expression syntax. As this library is meant to be language independent (even though it's written in Python), you should be able to compile regular expressions to work with other languages as well. To do that, you can simply import all the elements as a sub-package, and they should work identically, although some languages may not have the same features as others.

>>> import ezregex as er # The python dialect is the defualt dialect
>>> er.group(digit, 'name') + er.earlier_group('name')
EZRegex("(?P<name>\d)(?P=name)")
>>> import ezregex.perl as er
>>> er.group(digit, 'name') + er.earlier_group('name')
EZRegex("?P<name>\d)(\g<name>")

The currently implemented dialects are:

• Python
  • Well tested, ~99% implemented
• JavaScript
  • Under active development, the basics should work, though tests aren't in place yet
• Perl
  • Next on the roadmap, technically importable, but not implemented yet

If you know a particular flavor of regex and would like to contribute, feel free to read the developer documentation and make a pull request! If you would like one that's not implemented yet, you can also add a github issue.

Usage

• All the functions in the Python re library (search, match, sub, etc.) are implemented in the Python EZRegex dialect, and act identically to their equivalents. If you still want to use the Python re library directly, note that functions like search and sub don't accept EZRegex patterns as valid regex. Be sure to either call .str() (or cast it to a string) or .compile() (to compile to an re.Pattern) when passing to those. Using the member functions however, will be more efficient, as EZRegex caches the compiled re.Pattern internally.

Documentation

Notes and Gotchas

• The different Regular Expression dialects don't all have the same features, and those features don't all work the same way. I've tried to standardize these as best I can and use reasonable names for all the elements. If you're confused by something not working as expected, be sure to understand how your language specifically handles regular expressions.
• Be careful to call functions on the entire pattern: chunk + whitespace.str() is not the same as (chunk + whitespace).str().
• In regular regex, a lot of random things capture groups for no apparent reason. All regexes in EZRegex intentionally capture passively, so to capture any groups, use group(), with the optional name parameter.
• All EZRegexs (except for raw) auto-sanitize strings given to them, so there's no need to escape characters or use r strings. This does mean, however, that you cannot pass actual regex strings to any of them, as they'll think you're talking about it literally (unless you want that, of course). To include already written regex strings, use raw
• Note that I have camelCase and snake_case versions of each of the functions, because I waver back and forth between which I like better. Both versions function identically.
• The InputType can accept strings, other EZRegexs, or entire sequences of EZRegex patterns. It can also accept things that can be cast to a string, but it will warn you when it does, so it's better to cast to a string yourself.
• The invert function can accept any regular expression, not just EZRegex expressions, if you want to use it independently of the rest of the library.

perl

Positionals

These differentiate the string starting with a sequence, and a line starting with a sequence. Do note that the startof the string is also the start of a line. These can also be called without parameters to denote the start/end of astring/line without something specific having to be next to it.

• string_starts_with
• string_ends_with
• line_starts_with
• line_ends_with
• word_boundary
  • Matches the boundary of a word, i.e. the empty space between a word character and not a word character, or the end of a string
• not_word_boundary
  • The opposite of wordBoundary

Literals

• tab
• space
• space_or_tab
• new_line
• carriage_return
• quote
  • Matches ', ", and `
• vertical_tab
• form_feed
• comma
• period
• underscore

Not Literals

• not_whitespace
• not_digit
• not_word

Catagories

• whitespace
• whitechunk
  • A "chunk" of whitespace. Just any amount of whitespace together
• digit
• number
  • Matches multiple digits next to each other. Does not match negatives or decimals
• word
• word_char
  • Matches just a single "word character", defined as any letter, number, or _
• anything
  • Matches any single character, except a newline. To also match a newline, use literally_anything
• chunk
  • A "chunk": Any clump of characters up until the next newline
• uppercase
• lowercase
• letter
  • Matches just a letter -- not numbers or _ like word_char
• hex_digit
• oct_digit
• punctuation
• controller
  • Matches a metadata ASCII characters
• printable
  • Matches printable ASCII characters
• printable_and_space
• alpha_num
• unicode
  • Matches a unicode character by name
• any_between(char: str, and_char: str)
  • Match any char between char and and_char, using the ASCII table for reference

Amounts

• match_max(input: InputType)
  • Match as many of input in the string as you can. This is equivelent to using the unary + operator. If input is not provided, it works on the previous regex pattern. That's not recommended for clarity's sake though
• match_num(num: int, input: InputType)
  • Match num amount of input in the string
• match_more_than(min: int, input: InputType)
  • Match more than min sequences of input in the string
• match_at_least(min: int, input: InputType)
  • Match at least min sequences of input in the string
• match_at_most(max: int, input: InputType)
  • Match at most max instances of input in the string
• match_range(min: int, max: int, input: InputType, greedy: bool=True, possessive: bool=False)
  • Match between min and max sequences of input in the string. This also accepts greedy and possessive parameters Max can be an empty string to indicate no maximum greedy means it will try to match as many repititions as possible non-greedy will try to match as few repititions as possible possessive means it won't backtrack to try to find any repitions see https://docs.python.org/3/library/re.html for more help
• at_least_one(input: InputType, greedy: bool=True, possessive: bool=False)
  • Match at least one of input in the string. This also accepts greedy and possessive parameters greedy means it will try to match as many repititions as possible non-greedy will try to match as few repititions as possible possessive means it won't backtrack to try to find any repitions see https://docs.python.org/3/library/re.html for more help
• at_least_none(input: InputType, greedy: bool=True, possessive: bool=False)
  • Match 0 or more sequences of input. This also accepts greedy and possessive parameters greedy means it will try to match as many repititions as possible non-greedy will try to match as few repititions as possible possessive means it won't backtrack to try to find any repitions see https://docs.python.org/3/library/re.html for more help

Choices

• optional(input: InputType, greedy: bool=True, possessive: bool=False)
  • Match input if it's there. This also accepts greedy and possessive parameters greedy means it will try to match as many repititions as possible non-greedy will try to match as few repititions as possible possessive means it won't backtrack to try to find any repitions see https://docs.python.org/3/library/re.html for more help
• either(input: InputType, or_input: InputType)
• any_of(*inputs: str, chars: bool | None=None, split: bool | None=None)
  • Match any of the given inputs. Note that inputs can be multiple parameters, or a single string. Can also accept parameters chars and split. If char is set to True, then inputs must only be a single string, it interprets inputs as characters, and splits it up to find any of the chars in the string. If split is set to true, it forces the ?(...) regex syntax instead of the [...] syntax. It should act the same way, but your output regex will look different. By default, it just optimizes it for you.
• any_char_except(*inputs: str)
  • This matches any char that is NOT in inputs. inputs can be multiple parameters, or a single string of chars to split.
• any_except(input: InputType, type: InputType='.*')
  • Matches anything other than input, which must be a single string or EZRegex chain, not a list. Also optionally accepts the type parameter, which works like this: "Match any type other than input". For example, "match any word which is not foo". Do note that this function is new, and I'm still working out the kinks.
• each(*inputs: InputType)
  • Matches if the next part of the string can match all of the given inputs. Like the + operator, but out of order.

Conditionals

These can only be used once in a given expression. They only match a given expression if the expression is/ins'tfollowed/preceeded by a the given pattern

• if_proceded_by(input: InputType)
  • Matches the pattern if it has input coming after it. Can only be used once in a given pattern, as it only applies to the end
• if_not_proceded_by(input: InputType)
  • Matches the pattern if it does not have input coming after it. Can only be used once in a given pattern, as it only applies to the end
• if_preceded_by(input: InputType)
  • Matches the pattern if it has input coming before it. Can only be used once in a given pattern, as it only applies to the beginning
• if_not_preceded_by(input: InputType)
  • Matches the pattern if it does not have input coming before it. Can only be used once in a given pattern, as it only applies to the beginning
• if_enclosed_with(open: str, stuff: InputType, close: str | None=None)
  • Matches if the string has open, then stuff, then close, but only "matches" stuff. Just a convenience combination of ifProceededBy and ifPreceededBy.

Grouping

• group(input: InputType, name: str | None=None)
  • Causes input to be captured as an unnamed group. Only useful when replacing regexs
• passive_group(input: InputType)
  • As all regexs in EZRegex capture passively, this is entirely useless. But if you really want to, here it is
• earlier_group(num_or_name: int | str)
  • Matches whatever the group referenced by num_or_name matched earlier. Must be after a group which would match num_or_name
• if_exists(num_or_name: int | str, does: InputType, doesnt: InputType | None=None)
  • Matches does if the group num_or_name exists, otherwise it matches doesnt

Replacement

In the intrest of "I don't want to think about any syntax at all", I have included replace members. Do note that theyare not interoperable with the other EZRegexs, and can only be used with other strings and each other.

• rgroup(num_or_name: str | int)
  • Puts in its place the group specified, either by group number (for unnamed groups) or group name (for named groups). Named groups are also counted by number, I'm pretty sure. Groups are numbered starting from 1
• replace_entire
  • Puts in its place the entire match
• replace(string: str, rtn_str: bool=True) -> str | EZRegex
  • Generates a valid regex replacement string, using Python f-string like syntax.

Example: replace("named: {group}, numbered: {1}, entire: {0}")

Like Python f-strings, use {{ and }} to specify { and }

Set the rtn_str parameter to True to have it return an EZRegex type instead of a string

Note: Remember that index 0 is the entire match

There's a few of advantages to using this instead of just the regular regex replacement syntax:

• It's consistent between dialects
• It's closer to Python f-string syntax, which is cleaner and more familiar
• It handles numbered, named, and entire replacement types the same

Misc

• is_exactly(input: InputType)
  • This matches the string if and only if the entire string is exactly equal to input
• literal(input: InputType)
  • This is a redundant function. You should always be able to use ... + 'stuff' just as easily as ... + literal('stuff')
• raw(regex: str)
  • If you already have some regular regex written, and you want to incorperate it, this will allow you to include it without sanatizing all the backslaches and such, which all the other EZRegexs do automatically

Premade

These are some useful combinations that may be commonly used. They are not as stable, and may be changed and added to inlater versions to make them more accurate

• literally_anything
  • Any character, include newline
• signed
  • a signed number, including 123, -123, and +123
• unsigned
  • Same as number. Will not match +123
• plain_float
  • Will match 123.45 and 123.
• full_float
  • Will match plain_float as well as things like 1.23e-10 and 1.23e+10
• int_or_float
• ow
  • "Optional Whitechunk"
• email
  • Matches an email
• version
  • The official regex for matching version numbers from https://semver.org/. It includes 5 groups that can bematched/replaced: major, minor, patch, prerelease, and buildmetadata
• version_numbered
  • Same as version, but it uses numbered groups for each version number instead of named groups

Flags

These shadow python regex flags, and can just as easily be specified directly to the re library instead. They're providedhere for compatibility with other regex dialects. See https://docs.python.org/3/library/re.html#flags for details

• ASCII
  • Make matching words, word boundaries, digits, and whitespace perform ASCII-only matching instead of full Unicodematching. This is only meaningful for Unicode (str) patterns, and is ignored for bytes patterns
• IGNORECASE
  • Perform case-insensitive matching, including expressions that explicitly use uppercase members. Full Unicode matching(such as Ü matching ü) also works unless the ASCII flag is used to disable non-ASCII matches. The current locale does notchange the effect of this flag unless the LOCALE flag is also used
• DOTALL
  • Make the '.' special character match any character at all, including a newline. It's recommended you simply useliterally_anything instead
• LOCALE
  • Try not to use this, and rely on unicode matching instead
• MULTILINE
  • This is automatically inserted when using line_start and line_end, you shouldn't need to add it manually
• UNICODE
  • Match using the full unicode standard, instead of just ASCII. Enabled by default, and therefore redundant.

javascript

Positionals

These differentiate the string starting with a sequence, and a line starting with a sequence. Do note that the startof the string is also the start of a line. These can also be called without parameters to denote the start/end of astring/line without something specific having to be next to it.

• string_starts_with
• string_ends_with
• line_starts_with
• line_ends_with
• word_boundary
  • Matches the boundary of a word, i.e. the empty space between a word character and not a word character, or the end of a string
• not_word_boundary
  • The opposite of wordBoundary

Literals

• tab
• space
• space_or_tab
• new_line
• carriage_return
• quote
  • Matches ', ", and `
• vertical_tab
• form_feed
• comma
• period
• underscore

Not Literals

• not_whitespace
• not_digit
• not_word

Catagories

• whitespace
• whitechunk
  • A "chunk" of whitespace. Just any amount of whitespace together
• digit
• number
  • Matches multiple digits next to each other. Does not match negatives or decimals
• word
• word_char
  • Matches just a single "word character", defined as any letter, number, or _
• anything
  • Matches any single character, except a newline. To also match a newline, use literally_anything
• chunk
  • A "chunk": Any clump of characters up until the next newline
• uppercase
• lowercase
• letter
  • Matches just a letter -- not numbers or _ like word_char
• hex_digit
• oct_digit
• punctuation
• controller
  • Matches a metadata ASCII characters
• printable
  • Matches printable ASCII characters
• printable_and_space
• alpha_num
• unicode
  • Matches a unicode character by name
• any_between(char: str, and_char: str)
  • Match any char between char and and_char, using the ASCII table for reference

Amounts

• match_max(input: InputType)
  • Match as many of input in the string as you can. This is equivelent to using the unary + operator. If input is not provided, it works on the previous regex pattern. That's not recommended for clarity's sake though
• match_num(num: int, input: InputType)
  • Match num amount of input in the string
• match_more_than(min: int, input: InputType)
  • Match more than min sequences of input in the string
• match_at_least(min: int, input: InputType)
  • Match at least min sequences of input in the string
• match_at_most(max: int, input: InputType)
  • Match at most max instances of input in the string
• match_range(min: int, max: int, input: InputType, greedy: bool=True, possessive: bool=False)
  • Match between min and max sequences of input in the string. This also accepts greedy and possessive parameters Max can be an empty string to indicate no maximum greedy means it will try to match as many repititions as possible non-greedy will try to match as few repititions as possible possessive means it won't backtrack to try to find any repitions see https://docs.python.org/3/library/re.html for more help
• at_least_one(input: InputType, greedy: bool=True, possessive: bool=False)
  • Match at least one of input in the string. This also accepts greedy and possessive parameters greedy means it will try to match as many repititions as possible non-greedy will try to match as few repititions as possible possessive means it won't backtrack to try to find any repitions see https://docs.python.org/3/library/re.html for more help
• at_least_none(input: InputType, greedy: bool=True, possessive: bool=False)
  • Match 0 or more sequences of input. This also accepts greedy and possessive parameters greedy means it will try to match as many repititions as possible non-greedy will try to match as few repititions as possible possessive means it won't backtrack to try to find any repitions see https://docs.python.org/3/library/re.html for more help

Choices

• optional(input: InputType, greedy: bool=True, possessive: bool=False)
  • Match input if it's there. This also accepts greedy and possessive parameters greedy means it will try to match as many repititions as possible non-greedy will try to match as few repititions as possible possessive means it won't backtrack to try to find any repitions see https://docs.python.org/3/library/re.html for more help
• either(input: InputType, or_input: InputType)
• any_of(*inputs: str, chars: bool | None=None, split: bool | None=None)
  • Match any of the given inputs. Note that inputs can be multiple parameters, or a single string. Can also accept parameters chars and split. If char is set to True, then inputs must only be a single string, it interprets inputs as characters, and splits it up to find any of the chars in the string. If split is set to true, it forces the ?(...) regex syntax instead of the [...] syntax. It should act the same way, but your output regex will look different. By default, it just optimizes it for you.
• any_char_except(*inputs: str)
  • This matches any char that is NOT in inputs. inputs can be multiple parameters, or a single string of chars to split.
• any_except(input: InputType, type: InputType='.*')
  • Matches anything other than input, which must be a single string or EZRegex chain, not a list. Also optionally accepts the type parameter, which works like this: "Match any type other than input". For example, "match any word which is not foo". Do note that this function is new, and I'm still working out the kinks.
• each(*inputs: InputType)
  • Matches if the next part of the string can match all of the given inputs. Like the + operator, but out of order.

Conditionals

These can only be used once in a given expression. They only match a given expression if the expression is/ins'tfollowed/preceeded by a the given pattern

• if_proceded_by(input: InputType)
  • Matches the pattern if it has input coming after it. Can only be used once in a given pattern, as it only applies to the end
• if_not_proceded_by(input: InputType)
  • Matches the pattern if it does not have input coming after it. Can only be used once in a given pattern, as it only applies to the end
• if_preceded_by(input: InputType)
  • Matches the pattern if it has input coming before it. Can only be used once in a given pattern, as it only applies to the beginning
• if_not_preceded_by(input: InputType)
  • Matches the pattern if it does not have input coming before it. Can only be used once in a given pattern, as it only applies to the beginning
• if_enclosed_with(open: str, stuff: InputType, close: str | None=None)
  • Matches if the string has open, then stuff, then close, but only "matches" stuff. Just a convenience combination of ifProceededBy and ifPreceededBy.

Grouping

• group(input: InputType, name: str | None=None)
  • Causes input to be captured as an unnamed group. Only useful when replacing regexs
• passive_group(input: InputType)
  • As all regexs in EZRegex capture passively, this is entirely useless. But if you really want to, here it is
• earlier_group(num_or_name: int | str)
  • Matches whatever the group referenced by num_or_name matched earlier. Must be after a group which would match num_or_name
• if_exists(num_or_name: int | str, does: InputType, doesnt: InputType | None=None)
  • Matches does if the group num_or_name exists, otherwise it matches doesnt

Replacement

In the intrest of "I don't want to think about any syntax at all", I have included replace members. Do note that theyare not interoperable with the other EZRegexs, and can only be used with other strings and each other.

• rgroup(num_or_name: str | int)
  • Puts in its place the group specified, either by group number (for unnamed groups) or group name (for named groups). Named groups are also counted by number, I'm pretty sure. Groups are numbered starting from 1
• replace_entire
  • Puts in its place the entire match
• replace(string: str, rtn_str: bool=True) -> str | EZRegex
  • Generates a valid regex replacement string, using Python f-string like syntax.

Example: replace("named: {group}, numbered: {1}, entire: {0}")

Like Python f-strings, use {{ and }} to specify { and }

Set the rtn_str parameter to True to have it return an EZRegex type instead of a string

Note: Remember that index 0 is the entire match

There's a few of advantages to using this instead of just the regular regex replacement syntax:

• It's consistent between dialects
• It's closer to Python f-string syntax, which is cleaner and more familiar
• It handles numbered, named, and entire replacement types the same

Misc

• is_exactly(input: InputType)
  • This matches the string if and only if the entire string is exactly equal to input
• literal(input: InputType)
  • This is a redundant function. You should always be able to use ... + 'stuff' just as easily as ... + literal('stuff')
• raw(regex: str)
  • If you already have some regular regex written, and you want to incorperate it, this will allow you to include it without sanatizing all the backslaches and such, which all the other EZRegexs do automatically

Premade

These are some useful combinations that may be commonly used. They are not as stable, and may be changed and added to inlater versions to make them more accurate

• literally_anything
  • Any character, include newline
• signed
  • a signed number, including 123, -123, and +123
• unsigned
  • Same as number. Will not match +123
• plain_float
  • Will match 123.45 and 123.
• full_float
  • Will match plain_float as well as things like 1.23e-10 and 1.23e+10
• int_or_float
• ow
  • "Optional Whitechunk"
• email
  • Matches an email
• version
  • The official regex for matching version numbers from https://semver.org/. It includes 5 groups that can bematched/replaced: major, minor, patch, prerelease, and buildmetadata
• version_numbered
  • Same as version, but it uses numbered groups for each version number instead of named groups

Flags

These shadow python regex flags, and can just as easily be specified directly to the re library instead. They're providedhere for compatibility with other regex dialects. See https://docs.python.org/3/library/re.html#flags for details

• ASCII
  • Make matching words, word boundaries, digits, and whitespace perform ASCII-only matching instead of full Unicodematching. This is only meaningful for Unicode (str) patterns, and is ignored for bytes patterns
• IGNORECASE
  • Perform case-insensitive matching, including expressions that explicitly use uppercase members. Full Unicode matching(such as Ü matching ü) also works unless the ASCII flag is used to disable non-ASCII matches. The current locale does notchange the effect of this flag unless the LOCALE flag is also used
• DOTALL
  • Make the '.' special character match any character at all, including a newline. It's recommended you simply useliterally_anything instead
• LOCALE
  • Try not to use this, and rely on unicode matching instead
• MULTILINE
  • This is automatically inserted when using line_start and line_end, you shouldn't need to add it manually
• UNICODE
  • Match using the full unicode standard, instead of just ASCII. Enabled by default, and therefore redundant.

python

Positionals

These differentiate the string starting with a sequence, and a line starting with a sequence. Do note that the startof the string is also the start of a line. These can also be called without parameters to denote the start/end of astring/line without something specific having to be next to it.

• string_starts_with
• string_ends_with
• line_starts_with
• line_ends_with
• word_boundary
  • Matches the boundary of a word, i.e. the empty space between a word character and not a word character, or the end of a string
• not_word_boundary
  • The opposite of wordBoundary

Literals

• tab
• space
• space_or_tab
• new_line
• carriage_return
• quote
  • Matches ', ", and `
• vertical_tab
• form_feed
• comma
• period
• underscore

Not Literals

• not_whitespace
• not_digit
• not_word

Catagories

• whitespace
• whitechunk
  • A "chunk" of whitespace. Just any amount of whitespace together
• digit
• number
  • Matches multiple digits next to each other. Does not match negatives or decimals
• word
• word_char
  • Matches just a single "word character", defined as any letter, number, or _
• anything
  • Matches any single character, except a newline. To also match a newline, use literally_anything
• chunk
  • A "chunk": Any clump of characters up until the next newline
• uppercase
• lowercase
• letter
  • Matches just a letter -- not numbers or _ like word_char
• hex_digit
• oct_digit
• punctuation
• controller
  • Matches a metadata ASCII characters
• printable
  • Matches printable ASCII characters
• printable_and_space
• alpha_num
• unicode
  • Matches a unicode character by name
• any_between(char: str, and_char: str)
  • Match any char between char and and_char, using the ASCII table for reference

Amounts

• match_max(input: InputType)
  • Match as many of input in the string as you can. This is equivelent to using the unary + operator. If input is not provided, it works on the previous regex pattern. That's not recommended for clarity's sake though
• match_num(num: int, input: InputType)
  • Match num amount of input in the string
• match_more_than(min: int, input: InputType)
  • Match more than min sequences of input in the string
• match_at_least(min: int, input: InputType)
  • Match at least min sequences of input in the string
• match_at_most(max: int, input: InputType)
  • Match at most max instances of input in the string
• match_range(min: int, max: int, input: InputType, greedy: bool=True, possessive: bool=False)
  • Match between min and max sequences of input in the string. This also accepts greedy and possessive parameters Max can be an empty string to indicate no maximum greedy means it will try to match as many repititions as possible non-greedy will try to match as few repititions as possible possessive means it won't backtrack to try to find any repitions see https://docs.python.org/3/library/re.html for more help
• at_least_one(input: InputType, greedy: bool=True, possessive: bool=False)
  • Match at least one of input in the string. This also accepts greedy and possessive parameters greedy means it will try to match as many repititions as possible non-greedy will try to match as few repititions as possible possessive means it won't backtrack to try to find any repitions see https://docs.python.org/3/library/re.html for more help
• at_least_none(input: InputType, greedy: bool=True, possessive: bool=False)
  • Match 0 or more sequences of input. This also accepts greedy and possessive parameters greedy means it will try to match as many repititions as possible non-greedy will try to match as few repititions as possible possessive means it won't backtrack to try to find any repitions see https://docs.python.org/3/library/re.html for more help

Choices

• optional(input: InputType, greedy: bool=True, possessive: bool=False)
  • Match input if it's there. This also accepts greedy and possessive parameters greedy means it will try to match as many repititions as possible non-greedy will try to match as few repititions as possible possessive means it won't backtrack to try to find any repitions see https://docs.python.org/3/library/re.html for more help
• either(input: InputType, or_input: InputType)
• any_of(*inputs: str, chars: bool | None=None, split: bool | None=None)
  • Match any of the given inputs. Note that inputs can be multiple parameters, or a single string. Can also accept parameters chars and split. If char is set to True, then inputs must only be a single string, it interprets inputs as characters, and splits it up to find any of the chars in the string. If split is set to true, it forces the ?(...) regex syntax instead of the [...] syntax. It should act the same way, but your output regex will look different. By default, it just optimizes it for you.
• any_char_except(*inputs: str)
  • This matches any char that is NOT in inputs. inputs can be multiple parameters, or a single string of chars to split.
• any_except(input: InputType, type: InputType='.*')
  • Matches anything other than input, which must be a single string or EZRegex chain, not a list. Also optionally accepts the type parameter, which works like this: "Match any type other than input". For example, "match any word which is not foo". Do note that this function is new, and I'm still working out the kinks.
• each(*inputs: InputType)
  • Matches if the next part of the string can match all of the given inputs. Like the + operator, but out of order.

Conditionals

These can only be used once in a given expression. They only match a given expression if the expression is/ins'tfollowed/preceeded by a the given pattern

• if_proceded_by(input: InputType)
  • Matches the pattern if it has input coming after it. Can only be used once in a given pattern, as it only applies to the end
• if_not_proceded_by(input: InputType)
  • Matches the pattern if it does not have input coming after it. Can only be used once in a given pattern, as it only applies to the end
• if_preceded_by(input: InputType)
  • Matches the pattern if it has input coming before it. Can only be used once in a given pattern, as it only applies to the beginning
• if_not_preceded_by(input: InputType)
  • Matches the pattern if it does not have input coming before it. Can only be used once in a given pattern, as it only applies to the beginning
• if_enclosed_with(open: str, stuff: InputType, close: str | None=None)
  • Matches if the string has open, then stuff, then close, but only "matches" stuff. Just a convenience combination of ifProceededBy and ifPreceededBy.

Grouping

• group(input: InputType, name: str | None=None)
  • Causes input to be captured as an unnamed group. Only useful when replacing regexs
• passive_group(input: InputType)
  • As all regexs in EZRegex capture passively, this is entirely useless. But if you really want to, here it is
• earlier_group(num_or_name: int | str)
  • Matches whatever the group referenced by num_or_name matched earlier. Must be after a group which would match num_or_name
• if_exists(num_or_name: int | str, does: InputType, doesnt: InputType | None=None)
  • Matches does if the group num_or_name exists, otherwise it matches doesnt

Replacement

In the intrest of "I don't want to think about any syntax at all", I have included replace members. Do note that theyare not interoperable with the other EZRegexs, and can only be used with other strings and each other.

• rgroup(num_or_name: str | int)
  • Puts in its place the group specified, either by group number (for unnamed groups) or group name (for named groups). Named groups are also counted by number, I'm pretty sure. Groups are numbered starting from 1
• replace_entire
  • Puts in its place the entire match
• replace(string: str, rtn_str: bool=True) -> str | EZRegex
  • Generates a valid regex replacement string, using Python f-string like syntax.

Example: replace("named: {group}, numbered: {1}, entire: {0}")

Like Python f-strings, use {{ and }} to specify { and }

Set the rtn_str parameter to True to have it return an EZRegex type instead of a string

Note: Remember that index 0 is the entire match

There's a few of advantages to using this instead of just the regular regex replacement syntax:

• It's consistent between dialects
• It's closer to Python f-string syntax, which is cleaner and more familiar
• It handles numbered, named, and entire replacement types the same

Misc

• is_exactly(input: InputType)
  • This matches the string if and only if the entire string is exactly equal to input
• literal(input: InputType)
  • This is a redundant function. You should always be able to use ... + 'stuff' just as easily as ... + literal('stuff')
• raw(regex: str)
  • If you already have some regular regex written, and you want to incorperate it, this will allow you to include it without sanatizing all the backslaches and such, which all the other EZRegexs do automatically

Premade

These are some useful combinations that may be commonly used. They are not as stable, and may be changed and added to inlater versions to make them more accurate

• literally_anything
  • Any character, include newline
• signed
  • a signed number, including 123, -123, and +123
• unsigned
  • Same as number. Will not match +123
• plain_float
  • Will match 123.45 and 123.
• full_float
  • Will match plain_float as well as things like 1.23e-10 and 1.23e+10
• int_or_float
• ow
  • "Optional Whitechunk"
• email
  • Matches an email
• version
  • The official regex for matching version numbers from https://semver.org/. It includes 5 groups that can bematched/replaced: major, minor, patch, prerelease, and buildmetadata
• version_numbered
  • Same as version, but it uses numbered groups for each version number instead of named groups

Flags

These shadow python regex flags, and can just as easily be specified directly to the re library instead. They're providedhere for compatibility with other regex dialects. See https://docs.python.org/3/library/re.html#flags for details

• ASCII
  • Make matching words, word boundaries, digits, and whitespace perform ASCII-only matching instead of full Unicodematching. This is only meaningful for Unicode (str) patterns, and is ignored for bytes patterns
• IGNORECASE
  • Perform case-insensitive matching, including expressions that explicitly use uppercase members. Full Unicode matching(such as Ü matching ü) also works unless the ASCII flag is used to disable non-ASCII matches. The current locale does notchange the effect of this flag unless the LOCALE flag is also used
• DOTALL
  • Make the '.' special character match any character at all, including a newline. It's recommended you simply useliterally_anything instead
• LOCALE
  • Try not to use this, and rely on unicode matching instead
• MULTILINE
  • This is automatically inserted when using line_start and line_end, you shouldn't need to add it manually

Operators

• +, <<, >>
  • These all do the same thing: combine expressions
• *
  • Multiplies an expression a number of times. expr * 3 is equivelent to expr + expr + expr. Can also be used like expr * ... is equivalent to any_amt(expr)
• +
  • A unary + operator acts exactly as a match_max() does, or, if you're familiar with regex syntax, the + operator
• []
  • expr[2, 3] is equivalent to match_range(2, 3, expr)
  • expr[2, ...] or expr[2,] is equivalent to at_least(2, expr)
  • expr[... , 2] is equivalent to at_most(2, expr)
  • expr[...] or expr[0, ...] is equivelent to at_least_0(expr)
  • expr[1, ...] is equivalent to at_least_1(expr)
• &
  • Coming soon! This will work like the + operator, but they can be out of order. Like an and operation.
• |
  • Coming soon! This will work like an or operation, which will work just like any_of()
• %
  • This automatically calls re.search() for you and returns the match object (or None). Use like this: (digit * 2) % '99 beers on the wall'
• ~
  • This inverts the expression, equivalent to calling the .invert() method

Developer Documentation

The EZRegex class

Everything relies on the EZRegex class. EZRegex shouldn't be instantiated by the user, as each dialect subclasses the EZRegex class and defines their own elements specific to that dialect (more on that later). Each element represents a fundamental part of the Regular Expression syntax for that language, as well as less-fundemental common combinations for convenience (like email and float).

EZRegex can accept a string or a function to define how it's supposed to interact with the current "chain" of elements. If it's a string, it just adds it to the end. If it's a function, it can accept any positional or named parameters, but has to accept cur=... as the last parameter (it's complicated). The cur parameter is the current regular expression chain, as a string. What's returned becomes the new cur parameter of the next element, or, if there is no next element, the final regex. That way you can add to the front or back of an expression, and you can change what exactly gets added to the current expression based on other parameters.

The EZRegex class has operators overloaded so you can combine them in intuitive ways and call them by intuitive names.

Typing & Linting

The updated method of doing this is to define all the EZRegex elements of a dialect in elements.py, and then add type hints and doc strings in the elements.pyi file. EZRegex elements that accept parameters are typed as functions (even though they're not), for both convenience for the user when using linter, and to give documentation in an easier way. EZRegex elements that don't accept parameters should be typed as EZRegex, and given documentation as a string on the line below it. This is slightly non-standard, but linters support it, as well as my documentation generator script, which parses the .pyi files. The elements can also be seperated into groups in the .pyi files by using "Group: \<group name\>\n\<group description\>", which also gets parsed by the documentation script. The groups aren't used in the actual library, but are helpful in seperating the documentation, as well as used in ezregex.org

Dialects

Because most regex dialects are 90% identical, a parent EZRegex class implements most of the applicable logic, and a hidden "base" dialect is implemented, but works a bit differently. It has an elements.py file, but it defines all the elements as a dict in the form of {"element_name": {"keyword": "arguements"}}. It then has a load_dialect() function, which is the only thing importable from it. The reason it's done this way is because most elements, though identical in different dialects, have to be the appropriate dialect subclass. load_dialect() takes the dialect type as a parameter, and instantiates the base elements from it's dict and returns a new dict of initialized elements to be dumped into the global scope of the dialect. The elements.py file of a specific dialect can then remove any elements that it doesn't support (using the del keyword) and add/overwrite any it does support, or that work differently.

Each subclass of EZRegex must implement a few options to describe the dialect-specific behavior of the EZRegex class, for example, in the JavaScript dialect, /'s are added to the beginning and end of the pattern, and flags are handled differently in each dialect. This has to be implemented directly into the EZRegex subclass.

There's 4 parts that are required:

• beginning and end
  • Plain strings which describe what to tack onto the beginning and end of the compiled pattern (but before flags are added)
• flag_func
  • An abstract function that gets called with final, which is the final compiled pattern with beginning and end attached, and flags, which is a string of all the flags applied to the pattern. Internally, the flags are single digits, because flags usually are. They get passed to this function as a single string, which can be parsed and modified if necissary (it usually isn't)
• escape_chars
  • The characters that need to be escaped. Should be a byte string (i.e. b'...')

Inverting

There's actually 2 algorithms implemented for "inverting" regexs. The old algorithm regexs the regexs in a specific order to replace parts one at a time. This is just as nasty and horrifying as it sounds. Dispite it being a terrible, terrible solution, I actually got it to work decently well.

Later, when I was reading up on abstract syntax trees, and scrolling around on PyPi, I realized that Python has one built in, and that it's available to use. I reimplemented the whole algorithm to instead parse the AST given by the built-in re lexer, and wrote my own parser on top of it, which works much better.

Along the way, I also discovered, deep in the corners of the internet, 2 other Python libraries which do almost the same thing: xeger (regex backwards), and sre_yield. xeger technically works, however it tends to include unprintable characters, so it's output isn't very readable. sre_yeild is better, but it can be very slow, and is not quite the use case I'm going for. My invert algorithm is meant to be a debugging tool (though it doubles well for a testing tool), so it does things like detecting words (as opposed to seperate word characters) and inserts actual words, and doing the same for numbers and inserting 12345..., as well as a couple other enhancements.

Tests

Tests are implemented using pytest. Dependancies required for testing are:

pytest jstyleson py_js_runner rich Cope pydantic

Installation

EZRegex is distributed on PyPI as a universal wheel and is available on Linux, macOS and Windows and supports Python 3.10+ and PyPy.

pip install ezregex

The import name is the same as the package name:

import ezregex as er

Todo

See the GitHub Issue Page

License

EZRegex is distributed under the MIT License

Credits

This library was written from scratch entirely by Copeland Carter. Inspirations for this project include:

• PyParsing
  • I stole a bunch of the operators (especially the [] operator) from them, though we happened upon the same basic structure independantly (convergent evolution, anyone?)
• regular-expressions.info
  • Their reference is where I got a lot of the other regex flavors
• human-regex
  • Gave me the idea for including element methods, instead of solely element functions