Struct IntegratedLongContextProcessor 

pub struct IntegratedLongContextProcessor { /* private fields */ }

Integrated long context processor that orchestrates all modules.

This processor coordinates:

1. Unified extraction (one AI call for all information)
2. AI-driven focus tracking with intelligent analysis
3. Coherence-based segmentation
4. Progressive compression with focus priority
5. Focus message injection
6. Pattern learning

Implementations

impl IntegratedLongContextProcessor

pub fn new( provider: Box<dyn Provider>, model: String, config: ProcessorConfig, ) -> Self

Create a new integrated processor.

Arguments

• provider - Provider for AI extraction.
• model - Model name for extraction.
• config - Processor configuration.

pub fn with_defaults(provider: Box<dyn Provider>, model: String) -> Self

Create with default configuration.

pub fn for_simple_conversation( provider: Box<dyn Provider>, model: String, ) -> Self

Create for simple conversations.

pub fn for_complex_technical(provider: Box<dyn Provider>, model: String) -> Self

Create for complex technical discussions.

pub fn without_ai_focus(provider: Box<dyn Provider>, model: String) -> Self

Create without AI focus tracking (rule-based only).

pub fn with_hardcode_config(self, config: HardcodeConfig) -> Self

Set custom hardcode config.

pub fn config(&self) -> &ProcessorConfig

Get configuration reference.

pub fn config_mut(&mut self) -> &mut ProcessorConfig

Get mutable configuration reference.

pub fn pattern_registry(&self) -> &PatternRegistry

Get pattern registry reference.

pub fn pattern_registry_mut(&mut self) -> &mut PatternRegistry

Get mutable pattern registry reference.

pub async fn process( &mut self, messages: Vec<Message>, session_id: Option<&str>, project_path: Option<&str>, ) -> Result<ProcessedResult>

Process long context with the integrated workflow.

Workflow

1. Get existing focuses from FocusManager
2. Extract all information with focus selection (one AI call)
3. Update FocusManager based on AI’s focus_decision
4. Segment messages by coherence
5. Compress segments with focus priority
6. Inject focus message
7. Learn patterns

Arguments

• messages - Messages to process.
• session_id - Optional session ID.
• project_path - Optional project path.

Returns

Processed messages and extraction result.

pub async fn process_with_provider( &mut self, messages: Vec<Message>, provider: Option<&dyn Provider>, session_id: Option<&str>, project_path: Option<&str>, ) -> Result<ProcessedResult>

Process with a provider reference (for progressive compressor).

This variant allows using a provider for AI-based compression stages.

pub fn quick_process( &mut self, messages: Vec<Message>, ) -> Result<ProcessedResult>

Quick process without AI extraction (for testing or fallback).

This skips extraction step and uses only rule-based processing.