Python Regex for XML like data string [duplicate]

This question already has an answer here:

(The links provided in the asked before do not provide answers applicable to this question. So it has been incorrectly flagged).

I have a string like so:

s = '"<w s='<v1>'><m><v2>v3</m><v name='x'><v4></v></w>"'


The the single quotes around attribute s of node w, have to be escaped for Python. I know 'normal' XML would escape chevrons with < and >, but I am dealing with a data conversion, so I have to deal with the data I am given.

<

>

I have escaped the single quotes like so ' just so I can try out some test cases.

'

What I am trying to achieve is find the bits in chevrons that don't have a paired closing tag (</ >) (i.e I want v1, v2 and v4, (v3 -isn't required, since it's not in chevrons - it's the inner text).

(</ >)

I have to embellish those by putting extra characters around them, leaving out the chevrons.

So I'd have $$v1 etc.

$$v1

I need to leave the original XML parts (paired opening/closing tags) intact.

The final converted string should look like so:

'"<w s='$$v1'><m>$$v2v3</m><v name='x'>$$v4</v></w>"'


Python's XML ElementTree couldn't deal with the chevron in element attributes.

lxml although better would also give me issues.

Either way I'd end up with malformed XML or mismatched tags.

So I have resorted to trying a number of things like using Regex's or parsing string. Things quickly descend into a mess.

One of the obstacles I am facing, is if the closing chevron is within a quoted string I have to ignore it and go on to the next chevron, to get to the end of the w XML node.

w

Overall, I don't actually care if Regex is not used.

But, I've been using regex101.com to try out some Regex's:

This didn't seem to work:

<(?P<tag>S)(.*?)>(.*?)</(P=tag)>


This seems better:

</?(?P<tag>S)(.*?)>


But, the quoted chevron should be ignored, and the regex isn't quite greedy enough...

Is there a way to do this with Regex?

EDIT: I will continue to add my findings here... And I will post a complete custom parser solution for others following in similar footsteps. Have had some time to research the Regexes further over course of a weekend...

This Regex for example also gives left chevrons that aren't preceded by a single quote:

(Negative Lookbehind)

((?<!['])<)


This Regex gives right chevrons that aren't followed by single quote:

(Negative Lookahead)

(>(?![']))


Here is a pattern that can be applied recursively when creating a parser and accommodates the quoting rules. You can check it out here:

(?P<opentag>((?P<oc>(?<!['])<)(?P<tag>S)s*(?P<attr>.*?)(?P<cc>>(?![']))))(?P<inner>.*?)(?P<closetag>(?P=oc)/(?P=tag)(?P=cc))


(?P<name>) - names a capture group, you can later reference with this syntax (?P=name).

(?P<name>)

(?P=name)

You have to escape / symbol with the Regex so you get /? - the ? makes it optional

/

/?

?

S means non space

S

.*? means any char, zero or more times. The ? on the end prevents greedy expansion.

.*?

?

These are other types of scenario you have to deal with:

p2 = '<a><b></b><c></c></a>'
p3 = '<a><b><c></c></b></a>'
p4 = '<a name='<var>'><b></b></a>'
p5 = '<a><var><var></a>'
p6 = '<a></var><var></a>'
p7 = '<a><b></b><b></b></a>'


For p6, it should leave it as <a></var>$$var</a>.

p6

<a></var>$$var</a>

Here's a link to a Folder on Google Drive, with 4 other versions I've tried

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

2 Answers
2

Check on the next regex <(?P<tag>[^<s/>]+)>(?!.*</(?P=tag)>).

<(?P<tag>[^<s/>]+)>(?!.*</(?P=tag)>)

To try the regex online and get an explanation, please click here.

Regex is not recommended for XML or "XML like" data:

Instead, consider any of the following alternative approaches: