normalize-session-analysis 0.3.1

Session analysis metrics for AI coding agent logs
Documentation
1
2
3
4
5
# normalize-session-analysis

Session analytics computation for AI coding agent logs — computes metrics from parsed `Session` data produced by `normalize-chat-sessions`.

Key types: `SessionAnalysisReport` (the top-level report, renamed from `SessionAnalysis` to follow the `*Report` convention), `ToolStats` (call counts, error rates, and total output characters per tool), `LargestToolResult` (top-10 individual tool results by character count, with turn index and preview), `TokenStats` (input/output/cache token totals), `ModelPricing` (cost estimation per model; now has per-version sonnet constants: `SONNET_4_5`, `SONNET_3_7`, `SONNET_3_5`, `SONNET_3`). Implements `OutputFormatter` for text and pretty output. Consumes `Session`/`Turn`/`Message`/`ContentBlock` from `normalize-chat-sessions`; outputs via `normalize-output`. Intentionally separate from parsing — what metrics matter is consumer-specific.