How do I say, in regular expressions:
Any portion of a string beginning with a capital letter, containing at least one space character, not containing the string
" _ " (space underscore space), and ending with the string "!!!" (without the quotes)?
I am having trouble with the "not containing" part.
Here is what I have so far:
[A-Z].* .*!!!
How do I modify this to also specify "Not containing ' _ '"?
It does not need to be the specific string " _ ". How can I say "not containing" ANY string? For instance not containing "dog"?
Edit: I'd like the solution to be compatible with Php's "preg_replace"
Edit: Examples:
Examples for " _ ":
Abc xyz!!! <---Matches
Hello World!!! <---Matches
Has _ Space Underscore Space!!! <--- Does Not Match
Examples for "dog":
What a dog!!! <--- Does not match, (contains "dog")
Hello World!!! <--- Matches
The x(?!y) expression matches x only if it is not immediately followed by y. So, this seems to be the thing you want:
[A-Z](?!%s)(.(?!%s))* (.(?!%s))*!!!
Where %s is your forbidden string.
Any possible regex for this would be probably much more complicated than two regexes. One like yours: [A-Z].* .*!!! and the second applied on matched strings and checking whether _ is contained.
Start with Any capital; have an optional string of everything except an underscore, a space, and then the everything except underscore again, followed by three exclamation marks.
[A-Z][^_]*[ ][^_]*!!!
First test your string for any occurance of " _ ", since that is a no match. Then check for what you want.
That's what I would do instead of spending a lot of time trying to figure out one regular expression for it.
Here's a nice site for testing your expressions: Nregex
Edit: I read some more on your question and see that my answer wasn't really good, so here's another attempt. A modification of one of the expressions above:
[A-Z](?! _ )(\w(?! _ ))* (\w(?! _ ))*!!!
[A-Z]([^ ]*(?! _ ) ?)*!!!
Edit: I missed your requirement for at least one space. The below regex includes that requirement:
[A-Z][^ ]* (?!_ )((?! _ ).)*!!!
I used to resolve it via grep's -v option (if I'm on Linux and/or can use grep).
So search for something, then skip the uninteresting parts.
grep something <INPUT> | grep -v uninteresting
Without grep (damn windows, without admin rights) but with Vim:
vim -c "v/i'm searching for this/d" -c "g/and don't need this/d" -c "w checkoutput" <INPUT>
(This opens , then deletes every line what does not match what you need, then deletes every line, what you does not need, then save the results as checkoutput. Which you should check.)
HTH
There is a nice little program in which you can built your regex together with testing it.
http://software.marioschneider-online.de/?C%23%3A_RegEx_Test
Related
I've been trying to write some python to escape 'invalid' markdown strings.
This is for use with a python library (python-telegram-bot) which requires unused markdown characters to be escaped with a \.
My aim is to match lone *,_,` characters, as well as invalid hyperlinks - eg, if no link is provided, and escape them.
An example of what I'm looking for is:
*hello* is fine and should not be changed, whereas hello* would become hello\*. On top of that, if values are nested, they should not be escaped - eg _hello*_ should remain unchanged.
My thought was to match all the doubles first, and then replace any leftover lonely characters. I managed a rough version of this using re.finditer():
def parser(txt):
match_md = r'(\*)(.+?)(\*)|(\_)(.+?)(\_)|(`)(.+?)(`)|(\[.+?\])(\(.+?\))|(?P<astx>\*)|(?P<bctck>`)|(?P<undes>_)|(?P<sqbrkt>\[)'
for e in re.finditer(match_md, txt):
if e.group('astx') or e.group('bctck') or e.group('undes') or e.group('sqbrkt'):
txt = txt[:e.start()] + '\\' + txt[e.start():]
return txt
note: regex was written to match *text*, _text_, `text`, [text](url), and then single *, _, `, [, knowing the last groups
But the issue here, is of course that the offset changes as you insert more characters, so everything shifts away. Surely there's a better way to do this than adding an offset counter?
I tried to use re.sub(), but I haven't been able to find how to replace a specific group, or had any luck with (?:) to 'not match' the valid markdown.
This was my re.sub attempt:
def test(txt):
match_md = r'(?:(\*)(.+?)(\*))|' \
'(?:(\_)(.+?)(\_))|' \
'(?:(`)(.+?)(`))|' \
'(?:(\[.+?\])(\(.+?\)))|' \
'(\*)|' \
'(`)|' \
'(_)|' \
'(\[)'
return re.sub(match_md, "\\\\\g<0>", txt)
This just prefixed every match with a backslash (which was expected, but I'd hoped the ?: would stop them being matched.)
Bonus would be if \'s already in the string were escaped too, so that they wouldn't interfere with the markdown present - this could be a source of error, as the library would see it as escaped, causing it see the rest as invalid.
Thanks in advance!
You are probably looking for a regular expression like this:
def test(txt):
match_md = r'((([_*]).+?\3[^_*]*)*)([_*])'
return re.sub(match_md, "\g<1>\\\\\g<4>", txt)
Note that for clarity I just made up a sample for * and _. You can expand the list in the [] brackets easily. Now let's take a look at this thing.
The idea is to crunch through strings that look like *foo_* or _bar*_ followed by text that doesn't contain any specials. The regex that matches such a string is ([_*]).+?\1[^_*]*: We match an opening delimiter, save it in \1, and go further along the line until we see the same delimiter (now closing). Then we eat anything behind that that doesn't contain any delimiters.
Now we want to do that as long as no more delimited strings remain, that's done with (([_*]).+?\2[^_*]*)*. What's left on the right side now, if anything, is an isolated special, and that's what we need to mask. After the match we have the following sub matches:
g<0> : the whole match
g<1> : submatch of ((([_*]).+?\3[^_*]*)*)
g<2> : submatch of (([_*]).+?\3[^_*]*)
g<3> : submatch of ([_*]) (hence the \3 above)
g<4> : submatch of ([_*]) (the one to mask)
What's left to you now is to find a way how to treat the invalid hyperlinks, that's another topic.
Update:
Unfortunately this solution masks out valid markdown such as *hello* (=> \*hello\*). The work around to fix this would be to add a special char to the end of line and remove the masked special char once the substitution is done. OP might be looking for a better solution.
I've been working on this for a while, but it's beyond my understanding of regex.
I'm using Yahoo Pipes on an RSS, and I want to create hashtags from titles; so, I'd like to remove space from everything between quotes, but, if there's a colon within the quotes, I only want the space removed between the words before the colon.
And, it would be great if I could also capture the unspaced words as a group, to be able to use: #$1 to output the hashtag in one step.
So, something like:
"The New Apple: Worlds Within Worlds" Before We Begin...
Could be substituted like #$1 - with this result:
"#TheNewApple: Worlds Within Worlds" Before We Begin...
After some work, I was able to come up with, this regex:
\s(?=\s)?|(‘|’|(Review)|:.*)
("Review" was a word that often came before colons and wouldn't be stripped, if it were later in the title; that's what that's for, but I would like to not require that, to be more universal)
But, it has two problems:
I have to use multiple steps. The result of that regex would be:
"TheNewApple: Worlds Within Worlds" Before We Begin...
And I could then add another regex step, to put the hash # in front
But, it only works if the quotes are first, and I don't know how to fix that...
You can do this all in one step with regex, with a caveat. You run into problems with a repeated capturing group because only the last iteration is available in the replacement string. Searching for ( (\w+))+ and replacing with $2 will replace all the words with just the last match - not what we want.
The way around this is to repeat the pattern an arbitrary number of times that will suffice for your use. Each separate group can be referenced.
Search: "(\w+)(?: (\w+))?(?: (\w+))?(?: (\w+))?(?: (\w+))?(?: (\w+))?
Replace: "#$1$2$3$4$5$6
This will replace up to 6-word titles, exactly as you need them. First, "(\w+) matches any word following a quote. In the replacement string, it is put back as "#$1, adding the hashtag. The rest is a repeated list of (?: (\w+))? matches, each matching a possible space and word. Notice the space is part of a non-capturing group; only the word is part of the inner capture group. In the replacement string, I have $1$2$3$4$5$6, which puts back the words, without the spaces. Notice that a colon will not match any part of this, so it will stop once it hits a colon.
Examples:
"The New Apple: Worlds Within Worlds" Before We Begin...
"The New Apple" Before We Begin...
"One: Two"
only "One" word
this has "Two Words"
"The Great Big Apple Dumpling"
"The Great Big Apple Dumpling Again: Part 2"
Results:
"#TheNewApple: Worlds Within Worlds" Before We Begin...
"#TheNewApple" Before We Begin...
"#One: Two"
only "#One" word
this has "#TwoWords"
"#TheGreatBigAppleDumpling"
"#TheGreatBigAppleDumplingAgain: Part 2"
You can match the text with
"([^:]*)(.*?)"(.*)
then use some programming language to output the result like this:
'"#' + removeSpace($1) + $2 + '"' + $3
I have no idea what language you're using, but this seems like a poor choice for regex. In Python I'd do this:
# Python 3
import re
titles = ['''"The New Apple: Worlds Within Worlds" Before We Begin...''',
'''"Made Up Title: For Example Only" So We Can Continue...''']
hashtagged_titles = list()
for title in titles:
hashtagme, *restofstring = title.split(":")
hashtag = '"#'+hashtagme[1:].translate(str.maketrans('', '', " "))
result = "{}:{}".format(hashtag, restofstring)
hashtagged_titles.append(result)
Do a global search for
\ (?=.*:)
Replaced with nothing. Example
You'll need a second search on the results of that if you want to capture "TheNewApple" as a single word.
Sorry in advance that this might be a little challenging to read...
I'm trying to parse a line (actually a subject line from an IMAP server) that looks like this:
=?utf-8?Q?Here is som?= =?utf-8?Q?e text.?=
It's a little hard to see, but there are two =?/?= pairs in the above line. (There will always be one pair; there can theoretically be many.) In each of those =?/?= pairs, I want the third argument (as defined by a ? delimiter) extracted. (In the first pair, it's "Here is som", and in the second it's "e text.")
Here's the regex I'm using:
=\?(.+)\?.\?(.*?)\?=
I want it to return two matches, one for each =?/?= pair. Instead, it's returning the entire line as a single match. I would have thought that the ? in the (.*?), to make the * operator lazy, would have kept this from happening, but obviously it doesn't.
Any suggestions?
EDIT: Per suggestions below to replace ".?" with "[^(\?=)]?" I'm now trying to do:
=\?(.+)\?.\?([^(\?=)]*?)\?=
...but it's not working, either. (I'm unsure whether [^(\?=)]*? is the proper way to test for exclusion of a two-character sequence like "?=". Is it correct?)
Try this:
\=\?([^?]+)\?.\?(.*?)\?\=
I changed the .+ to [^?]+, which means "everything except ?"
A good practice in my experience is not to use .*? but instead do use the * without the ?, but refine the character class. In this case [^?]* to match a sequence of non-question mark characters.
You can also match more complex endmarkers this way, for instance, in this case your end-limiter is ?=, so you want to match nonquestionmarks, and questionmarks followed by non-equals:
([^?]*\?[^=])*[^?]*
At this point it becomes harder to choose though. I like that this solution is stricter, but readability decreases in this case.
One solution:
=\?(.*?)\?=\s*=\?(.*?)\?=
Explanation:
=\? # Literal characters '=?'
(.*?) # Match each character until find next one in the regular expression. A '?' in this case.
\?= # Literal characters '?='
\s* # Match spaces.
=\? # Literal characters '=?'
(.*?) # Match each character until find next one in the regular expression. A '?' in this case.
\?= # Literal characters '?='
Test in a 'perl' program:
use warnings;
use strict;
while ( <DATA> ) {
printf qq[Group 1 -> %s\nGroup 2 -> %s\n], $1, $2 if m/=\?(.*?)\?=\s*=\?(.*?)\?=/;
}
__DATA__
=?utf-8?Q?Here is som?= =?utf-8?Q?e text.?=
Running:
perl script.pl
Results:
Group 1 -> utf-8?Q?Here is som
Group 2 -> utf-8?Q?e text.
EDIT to comment:
I would use the global modifier /.../g. Regular expression would be:
/=\?(?:[^?]*\?){2}([^?]*)/g
Explanation:
=\? # Literal characters '=?'
(?:[^?]*\?){2} # Any number of characters except '?' with a '?' after them. This process twice to omit the string 'utf-8?Q?'
([^?]*) # Save in a group next characters until found a '?'
/g # Repeat this process multiple times until end of string.
Tested in a Perl script:
use warnings;
use strict;
while ( <DATA> ) {
printf qq[Group -> %s\n], $1 while m/=\?(?:[^?]*\?){2}([^?]*)/g;
}
__DATA__
=?utf-8?Q?Here is som?= =?utf-8?Q?e text.?= =?utf-8?Q?more text?=
Running and results:
Group -> Here is som
Group -> e text.
Group -> more text
Thanks for everyone's answers! The simplest expression that solved my issue was this:
=\?(.*?)\?.\?(.*?)\?=
The only difference between this and my originally-posted expression was the addition of a ? (non-greedy) operator on the first ".*". Critical, and I'd forgotten it.
How do I get the substring " It's big \"problem " using a regular expression?
s = ' function(){ return " It\'s big \"problem "; }';
/"(?:[^"\\]|\\.)*"/
Works in The Regex Coach and PCRE Workbench.
Example of test in JavaScript:
var s = ' function(){ return " Is big \\"problem\\", \\no? "; }';
var m = s.match(/"(?:[^"\\]|\\.)*"/);
if (m != null)
alert(m);
This one comes from nanorc.sample available in many linux distros. It is used for syntax highlighting of C style strings
\"(\\.|[^\"])*\"
As provided by ePharaoh, the answer is
/"([^"\\]*(\\.[^"\\]*)*)"/
To have the above apply to either single quoted or double quoted strings, use
/"([^"\\]*(\\.[^"\\]*)*)"|\'([^\'\\]*(\\.[^\'\\]*)*)\'/
Most of the solutions provided here use alternative repetition paths i.e. (A|B)*.
You may encounter stack overflows on large inputs since some pattern compiler implements this using recursion.
Java for instance: http://bugs.java.com/bugdatabase/view_bug.do?bug_id=6337993
Something like this:
"(?:[^"\\]*(?:\\.)?)*", or the one provided by Guy Bedford will reduce the amount of parsing steps avoiding most stack overflows.
/(["\']).*?(?<!\\)(\\\\)*\1/is
should work with any quoted string
"(?:\\"|.)*?"
Alternating the \" and the . passes over escaped quotes while the lazy quantifier *? ensures that you don't go past the end of the quoted string. Works with .NET Framework RE classes
/"(?:[^"\\]++|\\.)*+"/
Taken straight from man perlre on a Linux system with Perl 5.22.0 installed.
As an optimization, this regex uses the 'posessive' form of both + and * to prevent backtracking, for it is known beforehand that a string without a closing quote wouldn't match in any case.
This one works perfect on PCRE and does not fall with StackOverflow.
"(.*?[^\\])??((\\\\)+)?+"
Explanation:
Every quoted string starts with Char: " ;
It may contain any number of any characters: .*? {Lazy match}; ending with non escape character [^\\];
Statement (2) is Lazy(!) optional because string can be empty(""). So: (.*?[^\\])??
Finally, every quoted string ends with Char("), but it can be preceded with even number of escape sign pairs (\\\\)+; and it is Greedy(!) optional: ((\\\\)+)?+ {Greedy matching}, bacause string can be empty or without ending pairs!
An option that has not been touched on before is:
Reverse the string.
Perform the matching on the reversed string.
Re-reverse the matched strings.
This has the added bonus of being able to correctly match escaped open tags.
Lets say you had the following string; String \"this "should" NOT match\" and "this \"should\" match"
Here, \"this "should" NOT match\" should not be matched and "should" should be.
On top of that this \"should\" match should be matched and \"should\" should not.
First an example.
// The input string.
const myString = 'String \\"this "should" NOT match\\" and "this \\"should\\" match"';
// The RegExp.
const regExp = new RegExp(
// Match close
'([\'"])(?!(?:[\\\\]{2})*[\\\\](?![\\\\]))' +
'((?:' +
// Match escaped close quote
'(?:\\1(?=(?:[\\\\]{2})*[\\\\](?![\\\\])))|' +
// Match everything thats not the close quote
'(?:(?!\\1).)' +
'){0,})' +
// Match open
'(\\1)(?!(?:[\\\\]{2})*[\\\\](?![\\\\]))',
'g'
);
// Reverse the matched strings.
matches = myString
// Reverse the string.
.split('').reverse().join('')
// '"hctam "\dluohs"\ siht" dna "\hctam TON "dluohs" siht"\ gnirtS'
// Match the quoted
.match(regExp)
// ['"hctam "\dluohs"\ siht"', '"dluohs"']
// Reverse the matches
.map(x => x.split('').reverse().join(''))
// ['"this \"should\" match"', '"should"']
// Re order the matches
.reverse();
// ['"should"', '"this \"should\" match"']
Okay, now to explain the RegExp.
This is the regexp can be easily broken into three pieces. As follows:
# Part 1
(['"]) # Match a closing quotation mark " or '
(?! # As long as it's not followed by
(?:[\\]{2})* # A pair of escape characters
[\\] # and a single escape
(?![\\]) # As long as that's not followed by an escape
)
# Part 2
((?: # Match inside the quotes
(?: # Match option 1:
\1 # Match the closing quote
(?= # As long as it's followed by
(?:\\\\)* # A pair of escape characters
\\ #
(?![\\]) # As long as that's not followed by an escape
) # and a single escape
)| # OR
(?: # Match option 2:
(?!\1). # Any character that isn't the closing quote
)
)*) # Match the group 0 or more times
# Part 3
(\1) # Match an open quotation mark that is the same as the closing one
(?! # As long as it's not followed by
(?:[\\]{2})* # A pair of escape characters
[\\] # and a single escape
(?![\\]) # As long as that's not followed by an escape
)
This is probably a lot clearer in image form: generated using Jex's Regulex
Image on github (JavaScript Regular Expression Visualizer.)
Sorry, I don't have a high enough reputation to include images, so, it's just a link for now.
Here is a gist of an example function using this concept that's a little more advanced: https://gist.github.com/scagood/bd99371c072d49a4fee29d193252f5fc#file-matchquotes-js
here is one that work with both " and ' and you easily add others at the start.
("|')(?:\\\1|[^\1])*?\1
it uses the backreference (\1) match exactley what is in the first group (" or ').
http://www.regular-expressions.info/backref.html
One has to remember that regexps aren't a silver bullet for everything string-y. Some stuff are simpler to do with a cursor and linear, manual, seeking. A CFL would do the trick pretty trivially, but there aren't many CFL implementations (afaik).
A more extensive version of https://stackoverflow.com/a/10786066/1794894
/"([^"\\]{50,}(\\.[^"\\]*)*)"|\'[^\'\\]{50,}(\\.[^\'\\]*)*\'|“[^”\\]{50,}(\\.[^“\\]*)*”/
This version also contains
Minimum quote length of 50
Extra type of quotes (open “ and close ”)
If it is searched from the beginning, maybe this can work?
\"((\\\")|[^\\])*\"
I faced a similar problem trying to remove quoted strings that may interfere with parsing of some files.
I ended up with a two-step solution that beats any convoluted regex you can come up with:
line = line.replace("\\\"","\'"); // Replace escaped quotes with something easier to handle
line = line.replaceAll("\"([^\"]*)\"","\"x\""); // Simple is beautiful
Easier to read and probably more efficient.
If your IDE is IntelliJ Idea, you can forget all these headaches and store your regex into a String variable and as you copy-paste it inside the double-quote it will automatically change to a regex acceptable format.
example in Java:
String s = "\"en_usa\":[^\\,\\}]+";
now you can use this variable in your regexp or anywhere.
(?<="|')(?:[^"\\]|\\.)*(?="|')
" It\'s big \"problem "
match result:
It\'s big \"problem
("|')(?:[^"\\]|\\.)*("|')
" It\'s big \"problem "
match result:
" It\'s big \"problem "
Messed around at regexpal and ended up with this regex: (Don't ask me how it works, I barely understand even tho I wrote it lol)
"(([^"\\]?(\\\\)?)|(\\")+)+"
I need to clip out all the occurances of the pattern '--' that are inside single quotes in long string (leaving intact the ones that are outside single quotes).
Is there a RegEx way of doing this?
(using it with an iterator from the language is OK).
For example, starting with
"xxxx rt / $ 'dfdf--fggh-dfgdfg' ghgh- dddd -- 'dfdf' ghh-g '--ggh--' vcbcvb"
I should end up with:
"xxxx rt / $ 'dfdffggh-dfgdfg' ghgh- dddd -- 'dfdf' ghh-g 'ggh' vcbcvb"
So I am looking for a regex that could be run from the following languages as shown:
+-------------+------------------------------------------+
| Language | RegEx |
+-------------+------------------------------------------+
| JavaScript | input.replace(/someregex/g, "") |
| PHP | preg_replace('/someregex/', "", input) |
| Python | re.sub(r'someregex', "", input) |
| Ruby | input.gsub(/someregex/, "") |
+-------------+------------------------------------------+
I found another way to do this from an answer by Greg Hewgill at Qn138522
It is based on using this regex (adapted to contain the pattern I was looking for):
--(?=[^\']*'([^']|'[^']*')*$)
Greg explains:
"What this does is use the non-capturing match (?=...) to check that the character x is within a quoted string. It looks for some nonquote characters up to the next quote, then looks for a sequence of either single characters or quoted groups of characters, until the end of the string. This relies on your assumption that the quotes are always balanced. This is also not very efficient."
The usage examples would be :
JavaScript: input.replace(/--(?=[^']*'([^']|'[^']*')*$)/g, "")
PHP: preg_replace('/--(?=[^\']*'([^']|'[^']*')*$)/', "", input)
Python: re.sub(r'--(?=[^\']*'([^']|'[^']*')*$)', "", input)
Ruby: input.gsub(/--(?=[^\']*'([^']|'[^']*')*$)/, "")
I have tested this for Ruby and it provides the desired result.
This cannot be done with regular expressions, because you need to maintain state on whether you're inside single quotes or outside, and regex is inherently stateless. (Also, as far as I understand, single quotes can be escaped without terminating the "inside" region).
Your best bet is to iterate through the string character by character, keeping a boolean flag on whether or not you're inside a quoted region - and remove the --'s that way.
If bending the rules a little is allowed, this could work:
import re
p = re.compile(r"((?:^[^']*')?[^']*?(?:'[^']*'[^']*?)*?)(-{2,})")
txt = "xxxx rt / $ 'dfdf--fggh-dfgdfg' ghgh- dddd -- 'dfdf' ghh-g '--ggh--' vcbcvb"
print re.sub(p, r'\1-', txt)
Output:
xxxx rt / $ 'dfdf-fggh-dfgdfg' ghgh- dddd -- 'dfdf' ghh-g '-ggh-' vcbcvb
The regex:
( # Group 1
(?:^[^']*')? # Start of string, up till the first single quote
[^']*? # Inside the single quotes, as few characters as possible
(?:
'[^']*' # No double dashes inside theses single quotes, jump to the next.
[^']*?
)*? # as few as possible
)
(-{2,}) # The dashes themselves (Group 2)
If there where different delimiters for start and end, you could use something like this:
-{2,}(?=[^'`]*`)
Edit: I realized that if the string does not contain any quotes, it will match all double dashes in the string. One way of fixing it would be to change
(?:^[^']*')?
in the beginning to
(?:^[^']*'|(?!^))
Updated regex:
((?:^[^']*'|(?!^))[^']*?(?:'[^']*'[^']*?)*?)(-{2,})
Hm. There might be a way in Python if there are no quoted apostrophes, given that there is the (?(id/name)yes-pattern|no-pattern) construct in regular expressions, but it goes way over my head currently.
Does this help?
def remove_double_dashes_in_apostrophes(text):
return "'".join(
part.replace("--", "") if (ix&1) else part
for ix, part in enumerate(text.split("'")))
Seems to work for me. What it does, is split the input text to parts on apostrophes, and replace the "--" only when the part is odd-numbered (i.e. there has been an odd number of apostrophes before the part). Note about "odd numbered": part numbering starts from zero!
You can use the following sed script, I believe:
:again
s/'\(.*\)--\(.*\)'/'\1\2'/g
t again
Store that in a file (rmdashdash.sed) and do whatever exec magic in your scripting language allows you to do the following shell equivalent:
sed -f rmdotdot.sed < file containing your input data
What the script does is:
:again <-- just a label
s/'\(.*\)--\(.*\)'/'\1\2'/g
substitute, for the pattern ' followed by anything followed by -- followed by anything followed by ', just the two anythings within quotes.
t again <-- feed the resulting string back into sed again.
Note that this script will convert '----' into '', since it is a sequence of two --'s within quotes. However, '---' will be converted into '-'.
Ain't no school like old school.