[BUG] Stripping ECMA Regex Leading and Trailing `/` Causes Errors

Bug Report Checklist

• Have you provided a full/minimal spec to reproduce the issue?
• Have you validated the input using an OpenAPI validator (example)?
• Have you tested with the latest master to confirm the issue still exists?
• Have you searched for related issues/PRs?
• What's the actual output vs expected output?

Description

Python code generation tries to strip leading and trailing / from regex patterns, but the pattern its using to do so matches when it shouldn't.

For example this pattern in my spec: [0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12} has no leading and trailing slashes, but the code linked above would still match and end up stripping off the trailing } which causes an error from the https://github.com/curious-odd-man/RgxGen library.

import java.util.regex.Pattern;
import java.util.regex.Matcher;


public class Example {

    public static String SHOULD_NOT_MATCH = "[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}";
    public static String SHOULD_MATCH = "/^abc$/i";
    
    public static void main(String args[]) {
        Pattern valueExtractor = Pattern.compile("^/?(.+?)/?(.?)$");
        
        Matcher m = valueExtractor.matcher(SHOULD_NOT_MATCH);
        System.out.println(m.find());
        System.out.println(m.group(1));
        
        Matcher m2 = valueExtractor.matcher(SHOULD_MATCH);
        System.out.println(m2.find());
        System.out.println(m2.group(1));
    }
}

To be clear: this is causing an error for me, but it's possible this silently changing a lot of other regex and using that changed regex to generate examples.

openapi-generator version

v5.2.1

Don't think this is a regression, I got the same error on v5.1.X as well

OpenAPI declaration file content or url

openapi: 3.0.0
info:
  title: 'Example'
  version: v1
servers:
  -
    url: 'https://example.com'
    description: 'Example API API'
paths:
  '/{id}':
    get:
      operationId: example
      parameters:
        -
          name: id 
          in: path
          required: true
          schema:
            type: string
            pattern: '^[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}$'
      responses:
        '200':
          description: example
          content:
            application/json:
              schema:
                type: string

Generation Details

I ran this with the docker container:

    docker run \
        --rm \
        -it \
        --volume "$(pwd)/generator-workspace:/workspace" \
        'openapitools/openapi-generator-cli:v5.2.1' \
        generate \
        --generator-name python \
        --output "/workspace/example" \
        --remove-operation-id-prefix \
        --invoker-package "$namespace" \
        --global-property modelTests=false,apiTests=false \
        --additional-properties packageName=alli.$namespace,library=asyncio \
        --input-spec "/workspace/openapi.yml"

Steps to reproduce

1. Try to generate a schema with the example above

Related issues/PRs

n/a

Suggest a fix

I don't think the / should be optional here, you only want to match if they are present so having a ? behind each making them optional doesn't make sense.

import java.util.regex.Pattern;
import java.util.regex.Matcher;


public class Example {

    public static String SHOULD_NOT_MATCH = "[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}";
    public static String SHOULD_MATCH = "/^abc$/i";
    
    public static void main(String args[]) {
        Pattern valueExtractor = Pattern.compile("^/(.+?)/(.?)$");
        
        Matcher m = valueExtractor.matcher(SHOULD_NOT_MATCH);
        if (m.find()) {
            System.out.println("found");
            System.out.println(m.group(1));
        }
        
        Matcher m2 = valueExtractor.matcher(SHOULD_MATCH);
        if (m2.find()) {
            System.out.println("found");
            System.out.println(m2.group(1));
        }
    }
}

The current regex only works with a single flag as well, if multiple flags need to be supported: Pattern.compile("^/(.+?)/(.+)?$").

I've only experiened this with the python generator, but I suspect the same bug is present everywhere RgxGen library is being used.

Start in modules/openapi-generator/src/main/java/org/openapitools/codegen/languages/PythonClientCodegen.java at the linked regex handling around lines 1167-1172, then reproduce the issue with the supplied OpenAPI schema and Python generator Docker command. Done means regex patterns without surrounding slashes are preserved and slash-delimited patterns still generate successfully without the RgxGen error.