在跳过缩写时，将camel case格式化为可在PHP中读取

Question

So i am stuck - I have looked at tons of answers in here, but none seems to resolve my last problem.

Through an API with JSON, I receive an equipment list in a camelcase format. I can not change that.

I need this camelcase to be translated into normal language -

So far i have gotten most words seperated through:

$string = "SomeEquipmentHere";

$spaced = preg_replace('/([A-Z])/', ' $1', $string);
var_dump($spaced);

string ' Some Equipment Here' (length=20)

$trimmed = trim($spaced);
var_dump($trimmed);
string 'Some Equipment Here' (length=19)

Which is working fine - But in some of the equipments consists of abbreviations

"ABSBrakes" - this would require ABS and separated from Brakes

I can't check for several uppercases next to each other since it will then keep ABS and Brakes together - there are more like these, ie: "CDRadio"

So what is want is the output to be:

"ABS Brakes"

Is there a way to format it so, if there is uppercases next to eachother, then only add a space before the last uppercase letter of that sequence?

I am not strong in regex.

EDIT

Both contributions are awesome - people coming here later should read both answers

The last problems to consists are the following patterns :

"ServiceOK" becomes "Service O K"

"ESP" becomes "ES P"

The pattern only consisting of a pure uppercased abbreviation is fixed by a function counting lowercase letter, if there is none, it will skip over the preg_replace().

But as Flying wrote in the comments on his answer, there could potentially be a lot of instances not covered by his regex, and an answer could be impossible - I don't know if this could be a challenge for the regex.

Possibly by adding some "If there is not a lowercase after the uppercase, there should not be inserted a space" rule

Answer 1

Here is a single-call pattern that doesn't use any anchors, capture groups, or references in the replacement string: /(?:[a-z]|[A-Z]+)\K(?=[A-Z]|\d+)/

Pattern&Replace Demo

Code: (Demo)

$tests = [
    'SomeEquipmentHere',
    'ABSBrakes',
    'CDRadio',
    'Valve14',
];
foreach ($tests as $test) {
    echo preg_replace('/(?:[a-z]|[A-Z]+)\K(?=[A-Z]|\d+)/',' ',$test),"
";
}

Output:

Some Equipment Here
ABS Brakes
CD Radio
Valve 14

This is a better method because there is nothing to mop up. If there are new strings to consider (that break my method), please leave them in a comment so that I can update my pattern.

Pattern Explanation:

/         #start the pattern
(?:[a-z]  #match 1 lowercase letter
|         #or
[A-Z]+)   #1 or more uppercase letters
\K        #restart the fullstring match (forget the past)
(?=[A-Z]  #look-ahead for 1 uppercase letter
|         #or
\d+)      #1 or more digits
/         #end the pattern

---

Edit:

There are some other patterns that may provide better accuracy including:

/(?:[a-z]|\B[A-Z]+)\K(?=[A-Z]\B|\d+)/

Granted, the above pattern will not properly handle ServiceOK

Demo Link Word Boundaries Link

---

or this pattern with an anchor:

/(?!^)(?=[A-Z][a-z]+|(?<=\D)\d)/

The above pattern will accurately split: SomeEquipmentHere, ABSBrakes, CDRadio, Valve14, ServiceOK, ESP as requested by the OP.

Demo Link

*Note: Pattern accuracy can be improved as more sample strings are provided.