PT-2516 - enhance pt-mongodb-index-check unused and duplicate analysis

[jenunes]

Multi-signal unused index scoring:

• Collect $indexStats, listIndexes, collStats, and serverStatus
• Score indexes into tiers: SAFE_TO_DROP, LIKELY_UNUSED, LOW_USAGE, MONITOR, KEEP
• Hard guards for id, unique, TTL, hidden, partial, and system indexes
• Warmup period detection (configurable --warmup-days)
• Cross-reference duplicate prefix indexes with container usage

Duplicate detection improvements:

• keyToken distinguishes hashed/text/2dsphere/2d from btree directions
• Handle primitive.Symbol decoding (MongoDB 8.x)
• Skip id as a droppable prefix candidate
• Property-aware: partialFilterExpression, sparse, collation checked
• Unique prefix warning when dropping would lose uniqueness
• Rich sectioned text report with Reason, Action, optional index sizes

Sharded cluster support:

• Aggregate $indexStats across shards (sum ops, oldest since)
• NormalizeIndexStat for mongos responses missing top-level name/key
• DeduplicateIndexRecords safety net

Connection and UX:

• Ping after connect for immediate auth/network errors
• Require --databases or --all-databases (or DB in URI)
• Skip system databases (admin, config, local) and system.* collections
• collStats IndexSizes uses dynamic map for arbitrary index names

Documentation:

• README.rst rewritten with new flags, sample output, sharded behavior, edge cases (warmup, read preference, background builds), PTDEBUG usage

Tests:

• 41 unit/regression tests covering scoring, duplicate detection, report rendering, system DB filtering, URI parsing, and sharded aggregation

• The contributed code is licensed under GPL v2.0

• Contributor Licence Agreement (CLA) is signed

• util/update-modules has been ran

• Documentation updated

• Test suite update

[it-percona-cla]

All committers have signed the CLA.