Re-exports§
pub use analysis::SessionDigest;pub use analysis::analyze_and_select_sessions;pub use diagnostics::DiagnoseResult;pub use diagnostics::FailureExample;pub use diagnostics::FailureType;pub use diagnostics::categorize_parse_error;pub use domain::EventFilters;pub use domain::SessionState;pub use domain::TokenLimit;pub use domain::TokenLimits;pub use domain::filter_events;pub use session::SessionAnalysisExt;pub use session::SessionSummary;pub use session::assemble_session;pub use state_updates::StateUpdates;pub use state_updates::extract_state_updates;
Modules§
Structs§
- Agent
Session - Complete agent conversation session assembled from normalized events.
- Agent
Step - Single LLM inference cycle with optional tool executions.
- Agent
Turn - Single user-initiated interaction cycle within a session.
- Cache
Creation Tokens - Tokens used to create new cache entries
- Cache
Read Tokens - Tokens read from existing cache entries (these still consume context)
- Context
Limit - Context window limit (maximum tokens allowed)
- Context
Window Usage - Complete snapshot of token usage for a single turn.
- Fresh
Input Tokens - Fresh input tokens (new content, not from cache)
- Message
Block - Agent text response message.
- Model
Spec - Output
Tokens - Output tokens generated by the model
- Reasoning
Block - Agent reasoning/thinking block.
- Session
Stats - Aggregated statistics for an entire session.
- Token
Count - Total token count (always non-negative)
- Tool
Call Block - Tool invocation request with timing information.
- Tool
Execution - Paired tool call and result with execution metrics.
- Tool
Result Block - Tool execution result with timing information.
- Turn
Metrics - Computed context window metrics for turn visualization.
- Turn
Stats - Aggregated statistics for a single turn.
- User
Message - User input message that initiates a turn.