Escape user input going to xmp metadata

[epuronta]

When generating XMP metadata, the data is embedded in the XML template string unescaped. Now, if any non-XML-safe data comes in, the output XMP is invalid and will make the output PDF unparseable as Factur-X.

This will produce something like this when reading the file back:
pypdf.errors.PdfReadError: XML in XmpInformation was invalid: not well-formed (invalid token)

There are two easy cases to make this occur in practice with metadata automatically extracted from Factur-X payload:

• Selling company name
• Invoice number in ExchangedDocument/ID

All metadata going to XMP generation should be escaped.

This PR does a couple of things:

• Add profile autodetection for minimum and basicwl - just a mechanical extension to the existing mechanism
• Run isort as expected on pdf.py
• Extend the roundtrip tests by having it first attach the generated XML to a PDF, and then reading the PDF back. This raises errors for a few existing tests cases
• Fix the XMP generation with escaping logic

Fixes #79

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 91.18%. Comparing base (5ed87a6) to head (eeda872).
Report is 35 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master      #80      +/-   ##
==========================================
+ Coverage   90.95%   91.18%   +0.22%     
==========================================
  Files          18       18              
  Lines        1360     1395      +35     
==========================================
+ Hits         1237     1272      +35     
  Misses        123      123              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.