logpoller: core changes and persistent store

[jadepark-dev]

This PR adds persistent database store in TON logpoller, improves query interface, and log ingestion pipeline

Primary Changes

• Added PostgreSQL persistent store with advanced query capabilities including byte-level and bit-level filtering
• Implemented stream-based event processing pipeline with concurrent transaction loading and log processing
• Enhanced query interface with chainlink-common integration for standardized pagination and sorting
• Refactored architecture from synchronous to asynchronous processing with channel-based data flow

Database & Storage

• Created PostgreSQL schema for filters and logs with proper indexing and constraints
• Added transaction support with configurable batch insertion and retry logic for timeout handling
• Implemented cursor-based pagination with address+msgLT composite cursors for efficient large dataset traversal
• Enhanced in-memory store compatibility to maintain test infrastructure during migration

Query System Improvements

• Replaced monolithic query builder with modular system supporting field, byte, and bit-level filtering
• Added TON-specific BOC data filtering with proper header offset handling for PostgreSQL alignment
• Implemented typed log decoding helpers with TLB parsing and application-level filtering support

Others

• Implemented memory-efficient log batching with configurable thresholds and database transaction support
• Extended configuration with database-specific settings while maintaining backward compatibility
• Enhanced service initialization with proper dependency injection and lifecycle management
• Added validation and default value application for missing configuration fields

Next Steps

• e2e behavior validation with core TON2EVM and EVM2TON messaging test
• add TON schema migration file in core
• initialize postgreSQL database when relayer is started, replacing memory store