Skip to content

Function.md

Latest commit

 

History

History
53 lines (47 loc) · 3.46 KB

Function.md

File metadata and controls

53 lines (47 loc) · 3.46 KB

This document describes LoggerKit core capabilities and provides a post-refactor functional validation checklist.

List Rendering

  • Pagination: thread-safe offset/limit queries
  • Real-time stats: background aggregation for level, function, file, context, thread, and session
  • Smart cache: prefer aggregated stats and search preview data to reduce repeated queries

Search

  • Data source: load up to 10,000 records from database for preview cache, independent from currently loaded list items
  • Multi-field search: message, file name, function, context, and thread; keep at least one active field
  • Realtime debounce: 100ms debounce for responsive interactions
  • Background execution: asynchronous search runs off the main thread
  • Categorized results: return grouped results in 5 dimensions with de-duplicated counts
  • Smart limits: message dimension returns top 5; other dimensions return all matches

Filtering

  • 7-dimension filtering:
    • Level (selectedLevels): exact match, multi-select, default all selected (verbose/debug/info/warning/error)
    • Function (selectedFunctions): exact match, multi-select, IN predicate
    • File name (selectedFileNames): exact match, multi-select, IN predicate
    • Context (selectedContexts): exact match, multi-select, IN predicate
    • Thread (selectedThreads): exact match, multi-select, IN predicate
    • Message keywords (selectedMessageKeywords): fuzzy match via CONTAINS[cd], multi-keyword OR logic
    • Session (selectedSessionIds): exact match, multi-select, IN predicate; empty means all sessions, and at least one session must be selected when filtering by session
  • Set operations: add/remove/toggle filter values
  • Active count: show number of active filters
  • Quick reset: one-click clear all filters; levels reset to default all-selected
  • Debounced callback: trigger onFilterChanged with 100ms debounce to avoid frequent queries

Export

  • Streaming export: batched query + append write, 1,000 records per batch, peak memory < 10MB, supports million-scale exports
  • LOG format:
    • Without thread: {datetime} [{first8SessionId}] [{level}] - ({function} at {file}:{line}) - {message}
    • With thread: {datetime} [{first8SessionId}] [{level}] <{context}> {thread} - ({function} at {file}:{line}) - {message}
    • Datetime format: from DateFormatters.displayFormatter
    • Encoding: UTF-8, extension: .log
  • Smart filename: logs_{session}_{date}.log (for example logs_all_20251217.log or logs_abc12345_20251217.log)
  • Data source strategy: choose optimal source based on loaded data volume and active filters
  • Progress callback: report written count and total count in real time

Deletion

  • Two-level deletion strategy:
    • Delete all: remove all logs in one action
    • Delete by session: remove selected sessions (single/multi-select)
  • Single/multi-session deletion: support deleteLogs(forSession:) and deleteLogs(forSessions:)
  • Confirmation flow: double confirmation to prevent accidental deletion
  • Background execution: run on background context to keep UI responsive

Performance Optimizations

  • Predicate reuse: centralize query predicate building to remove duplication
  • Query optimization: reduce stats queries from 9 to 2 (grouped query + popular function query)
  • Thread safety: all database operations use background context
  • Task cancellation: allow cancellation for running search/load tasks