Parser for cubex files

cubex file format

• anchor.xml
  • contains
    • cnodes
    • metrics
    • regions
    • system_tree_nodes
      • locationgroups
        • locations
• index.0
  • the 0 stands for the metric ID
  • contains
    • a header
      • "1"
        • a "1" encoded as a
      • endianness
      • the number of cnodes in the data.0 file
    • a list of cnode indices
      • cnode_indices = [c1, c2, ...]
• data.0
  • contains
    • a header
    • data for each cnode_id in cnode_indices
      • contains metric values for all locations
      • to retrieve the value of a particular cnode with cnode_id, and a specific location_id
        • get index of cnode_index in cnode_indices
          • = the position of the cnode_id in cnode_indices
        • offset in data.0: cnode_index * num_locations + location_id
          • the locations all have incrementing location_ids
      • Important: the list is sorted!
        • not in the order of the XML

Notes

• Requires at least Python version 3.5
  • Contains typings as defined in PEP 484
• The cubex files are tar archives
  • ⚠️ ... when extracting them and parsing the extracted 0.data files using the low-level IndexParser/DataParser directly, they will create strange behaviour
    • in the most cases, parsing succeeds but some edge-cases (related to endianness?) create problems