Isurp petta porting

[eyobderese]

This pull request introduces several improvements and refactorings across the empirical probability, surprisingness, and truth value estimation code. The changes focus on standardizing utility function names (e.g., replacing min/max with min-of-two/max-of-two), improving code clarity, updating type signatures, and enhancing test coverage. Additionally, the pull request updates test files to use local utility imports and adds a comprehensive new test suite for the surprisingness module.

Utility function standardization and usage updates:

• Replaced all occurrences of min/max with min-of-two/max-of-two (and corresponding min-atom/max-atom in some places) throughout the codebase for consistency, including in emp-tv.metta, isurp-old.metta, isurp.metta, TruthValue.metta, and beta-dist.metta. [1] [2] [3] [4] [5]

Empirical probability and surprisingness logic improvements:

• Refactored empirical probability and bootstrapping logic in emp-prob.metta:

  • Updated function signatures and types for clarity.
  • Improved conditional logic in emp_prob_subsmp and emp_prob_bs.
  • Added debug print statements in emp-prob-pbs for better traceability. [1] [2] [3] [4] [5] [6]

• Changed partitioning logic in surprisingness and truth value estimation:

  • Replaced calls to generet-partition-without-pattern with the more general partitions function in isurp.metta and est-tv.metta. [1] [2]

Type signature and documentation clean-up:

• Removed or commented out unused and redundant type signatures and documentation blocks in emp-prob.metta, isurp.metta, and emp-tv.metta to streamline code and improve readability. [1] [2] [3] [4] [5]

Test infrastructure improvements:

• Added a comprehensive test file petta-test.metta for the surprisingness module, including extensive commented-out test cases and utility imports for future expansion.
• Updated test-blk-abstractness.metta to use local utility imports instead of commented-out global imports, improving modularity and maintainability.

These changes collectively improve code consistency, maintainability, and testability across the surprisingness and truth value modules.