Improve file matching logic (Boite filename - S3 filename)

[PascalEgn]

Description

Extend the basic matcher from #24 to handle more complexity, more file types, multiple roots and better observability for Airflow operation.

Work involved

• Add a parameter to configure the types that should get matched. Examples are: PDF, PDF_LATEX, TIFF, PDF_OCR, PDF_TRANSMIS
• Add a parameter to configure multiple roots. So for example the default root raw/<TYPE>/... but more roots can be passed, for example raw/CORRECTIONS_2/<TYPE>/.... In this case the order of the list should determine which root gets preferred. Which means for example only if a file could not be found in the CORRECTIONS_2 root, the function should search for it in the basic raw/ root.
• Add run summary metrics (total matched, total unmatched)
• Add optional report mode. This should lead to a detailed output in the written logs (Bascially display summary metrics, and the matches for all listed files in the boite files)
• Add optional dry-run mode, this should stop the script execution after the matching happend. So no XML etc get created after...

Acceptance criteria

• Matcher supports all configured file types
• Files are found whether they are under primary roots or corrections roots
• Report mode with detailed logs exist
• Dry-run mode is implemented

Screenshots(Optional)