feat: add PageRank analyzer with power-iteration algorithm

Add PageRankAnalyzer implementing Google's PageRank algorithm for
computing recursive node importance in graphs:

- Power-iteration method with configurable damping factor (default 0.85),
  convergence tolerance, and max iterations
- Dangling node handling (distributes rank uniformly)
- PageRankResult with raw rank, normalized rank (relative to average),
  and human-readable importance labels
- Gini coefficient for rank inequality measurement
- Shannon entropy ratio for rank distribution uniformity
- Rank distribution tiers (Very Low to Very High)
- Hidden influencer detection (nodes with disproportionate importance
  relative to their degree)
- Degree centrality comparison with position shift tracking
- Top-N and bottom-N node queries

Also:
- Add JUnit 4.13.2 test dependency to pom.xml
- Add testSourceDirectory config pointing to Gvisual/test

77 new tests covering: constructors, empty/single/two-node graphs,
star/triangle/chain/complete/barbell/disconnected topologies,
convergence, rank-sum-to-1 invariant, damping factor effects,
Gini/entropy metrics, hidden influencers, degree comparison,
result immutability, and all result type methods.

Total tests: 271 -> 348 (all passing)

Author: sauravbhattacharya001