Add wildcard/glob pattern support for exclude_paths and include_paths

[akshat62]

Adds [*] (match one segment) and [**] (match zero or more segments) wildcard support to exclude_paths and include_paths across DeepDiff, DeepHash, and DeepSearch. Patterns are detected at init time, compiled into GlobPathMatcher objects, and checked alongside exact paths in _skip_this(). Quoted keys like root['*'] are treated as literals, not wildcards.

>>> t1 = {"users": {"alice": {"name": "Alice", "pw": "s1"}, "bob": {"name": "Bob", "pw": "s2"}}}
>>> t2 = {"users": {"alice": {"name": "Alice", "pw": "x1"}, "bob": {"name": "Bob", "pw": "x2"}}}
>>> DeepDiff(t1, t2, exclude_paths=["root['users'][*]['pw']"])
{}

[seperman]

Hi @akshat62 . The existing exclude_regex_paths should cover these. What is the benefit of supporting wildcard in exclude_paths and include_paths?

[akshat62]

Hi @seperman,
exclude_regex_paths only supports exclusion. The wildcard support works in both exclude_paths and include_paths, giving users a single consistent syntax for both directions.

Users don't need to learn a separate parameter - wildcards work directly in the existing exclude_paths/include_paths they're already using. The implementation auto-separates exact paths from glob patterns internally.

The glob support targets the 80% of cases where users just want "this key at any depth" without needing to write regex.

[akshat62]

Regex operates on the string representation of paths, so a pattern like r"b" would accidentally match root['pub'] or root['dba']. Glob matching works on parsed path segments, so root[*]['b'] only matches paths where 'b' is an actual key at that exact level.