Add APPROX_PERCENTILE aggregation function

[FearfulTomcat27]

Add comprehensive implementation of the APPROX_PERCENTILE function for efficient percentile calculation using T-Digest algorithm:

Core Components

• TDigest algorithm implementation with memory optimization and reset functionality
• Abstract base classes for both regular and grouped accumulator patterns
• Support for both basic and weighted percentile calculations

Features Added

• Efficient quantile estimation with configurable compression
• Memory usage tracking with RamUsageEstimator integration
• Type-safe conversions for int, long, float, and double data types
• Support for grouped aggregation operations
• Reset functionality for memory reuse

Performance Optimizations

• Code structure improvements and method extraction for better maintainability
• Reduced code duplication through inheritance hierarchy
• Memory-efficient array management with TDigestBigArray

[Copilot]

Pull Request Overview

This PR implements the APPROX_PERCENTILE aggregation function in IoTDB, providing efficient percentile calculation using the T-Digest algorithm. The implementation includes comprehensive support for both regular and weighted percentile calculations with memory optimizations.

• Adds T-Digest algorithm implementation with configurable compression and reset functionality
• Introduces abstract base classes for both regular and grouped accumulator patterns
• Provides support for all numeric data types (int, long, float, double) and grouped aggregations

Reviewed Changes

Copilot reviewed 16 out of 16 changed files in this pull request and generated 9 comments.

Show a summary per file

File	Description
common.thrift	Adds APPROX_PERCENTILE to TAggregationType enum
TableBuiltinAggregationFunction.java	Registers APPROX_PERCENTILE as builtin function
SqlConstant.java	Defines APPROX_PERCENTILE constant
AstBuilder.java	Adds validation for APPROX_PERCENTILE function arguments
TableMetadataImpl.java	Implements type validation and return type logic
TDigest.java	Core T-Digest algorithm implementation with serialization
TDigestBigArray.java	Memory-efficient array management for grouped operations
AbstractApproxPercentileAccumulator.java	Base class for single accumulator pattern
ApproxPercentileAccumulator.java	Regular percentile accumulator implementation
ApproxPercentileWithWeightAccumulator.java	Weighted percentile accumulator implementation
AbstractGroupedApproxPercentileAccumulator.java	Base class for grouped accumulator pattern
GroupedApproxPercentileAccumulator.java	Grouped regular percentile accumulator
GroupedApproxPercentileWithWeightAccumulator.java	Grouped weighted percentile accumulator
TDigestStateFactory.java	Factory for creating T-Digest state objects
AccumulatorFactory.java	Factory integration for accumulator creation
IoTDBTableAggregationIT.java	Integration tests for APPROX_PERCENTILE functionality