refactor(storage): inverted index support calculate the score

[b41sh]

I hereby agree to the terms of the CLA available at: https://docs.databend.com/dev/policies/cla/

Summary

• Inverted index support calculate the score
• Restore the inverted index test that was commented out

continue #16589

Tests

• Unit Test
• Logic Test
• Benchmark Test
• No Test - Explain why

Type of change

• Bug Fix (non-breaking change which fixes an issue)
• New Feature (non-breaking change which adds functionality)
• Breaking Change (fix or feature that could cause existing functionality not to work as expected)
• Documentation Update
• Refactoring
• Performance Improvement
• Other (please describe):

---

This change is 

[what-the-diff]

PR Summary

• Improved Test Variables in pruning.rs
  The test variables in pruning.rs were renamed from being prefixed with _ to their actual names (e13, e14, e15). This makes the code clearer and easier to understand.

• Activated Test Assertions
  The test assertions for the variables e13, e14, and e15 were activated to ensure they function as expected.

• Public Access Granted to TermReader in lib.rs
  The TermReader was added to the pub use list in lib.rs, meaning we've made it publicly accessible and ready for use.

• Added New Dependency tantivy-common
  The tantivy-common dependency was added to our project. This new software element enhances our project's functionality.

• Enhanced Field Normalization Handling
  By using FieldNormReader from tantivy in inverted_index_reader.rs, the handling of field normalization has been greatly improved.

• Upgraded Term ID Tracking
  We've introduced field_term_ids. This new feature allows for better tracking of term ids for each field during term info collection.

• Improved Index Slice Logic
  We've updated the way the index slice is created. It now handles score calculations and introduces new structures for managing field norms and total number tokens.

• Optimized Document Collection Process
  The DocIdsCollector instantiation now uses TermReader, which improves performance and clarity. Moreover, we've adjusted the scoring calculations, allowing for better score management.

• Enhanced Field Norms Reading
  In inverted_index_writer.rs, the reading of field norms has been streamlined, ensuring correct extraction from the segment file. This will make our data more consistent and reliable.