Module compression 

Context Compression Engine (Phase 2 - ENG-34)

Provides token counting and context budget management for LLM interactions. Uses tiktoken-rs for accurate token counting with explicit error handling.

Structs

ContextBudgetInput

Input for context budget checking

ContextBudgetResult

Result of context budget check

MemoryTokenCount

Token count for a single memory

Enums

CompressionStrategy

Compression strategy for memory content

TokenEncoding

Supported encoding types for token counting

Functions

check_context_budget

Check token budget for a set of memories

count_tokens

Count tokens in text using the specified model or encoding.

detect_encoding

Detect the appropriate encoding for a model name

parse_encoding

Parse encoding string to TokenEncoding