Skip to content

Releases: eirenik0/log-analyzer

0.1.3 (2026-02-19)

19 Feb 19:39
@github-actions github-actions
v0.1.3
84f7353

Choose a tag to compare

View all tags
0.1.3 (2026-02-19) Latest
Latest

Features

  • add source line tracking for log parsing and validation
  • add unified filter module with expression-based log filtering
  • enhance comparison output with table formatting and JSON shorthand support
  • introduce CLI enhancements and output improvements
  • add support for config file via --config flag in CLI
  • improve filtering logic
  • add generate-config command with embedded templates for profile generation
  • ensure unique/unpaired entries are included in JSON and text diff outputs

Improve CLI usability with new global flags:

  • Added -j, --json as shorthand for -F json -c for compact machine-readable output.
  • Added -f, --filter for unified filter expressions (for example: c:core l:ERROR !t:timeout).
  • -f, --filter can also be set with LOG_ANALYZER_FILTER.
  • --json conflicts with explicit -F, --format to avoid ambiguous output settings.

Add configurable profiles and starter templates for custom log formats:

  • Added runtime profile loading via --config <path> or LOG_ANALYZER_CONFIG.
  • Parser markers, pairing markers, and correlation keys are now configurable through profiles.
  • Added profile-aware info insights for unknown components, commands, requests, and session prefixes.
  • Added reusable templates: base, custom-start, service-api, and event-pipeline.

Improve diff output readability and diagnostics:

  • Differences are now classified as added, removed, or modified (+, -, ~).
  • Summary output now includes counts of additions, removals, and modifications.
  • Diff entries now include source line numbers for faster navigation to original logs.
  • JSON diff output now includes change_type, and text diffs are split into text1/text2.

Add profile generation command and built-in template support:

  • Added generate-config (gen-config) to create a TOML profile from a log file.
  • Generated profiles include discovered components, commands, requests, and session prefix hints.
  • Supports --profile-name, --template, and -o, --output.
  • --template now accepts either a file path or built-in template name (base, custom-start, service-api, event-pipeline).
  • Embedded built-in templates into the binary and use embedded base as default for better portability.

Add unified filter expression syntax:

  • Added expression-based filtering via -f, --filter.
  • Supports component, level, text, and direction terms (with short aliases like c:, l:, t:, d:).
  • Supports exclusions with ! (for example: !l:DEBUG).
  • Multiple terms are combined with AND semantics.
  • Matching for level and direction values is case-insensitive.
  • Unknown filter values now produce warnings to catch typos.

Fixes

Fix regressions in compare/diff filtering and output:

  • Repeated shared keys are now paired one-to-one; unmatched occurrences are kept as unique entries.
  • --sort-by time/component/level/type now sorts correctly.
  • --full now prints full payload JSON in comparison output.
  • -o, --output now writes the correct output for compare, diff, llm-diff, process, and perf (text and JSON).
  • Filter logic is now consistent: different filter types are AND-ed, multiple values of the same type are OR-ed.
  • diff output now includes unpaired unique entries in both text and JSON modes.
  • Parser no longer panics when profile config has empty command_payload_markers.

Improve output formatting for summaries:

  • Summary statistics now render as styled, width-aware tables.
  • Table formatting is applied consistently across console and file output.
  • Improves readability of command output for large result sets.
Assets 9
Loading

0.1.2 (2026-01-22)

22 Jan 09:14
@github-actions github-actions
v0.1.2
01eeeef

Choose a tag to compare

View all tags
0.1.2 (2026-01-22)

Features

  • Add filter for connection direction
  • Add advanced filtering, sorting, and CLI enhancements
  • Enhance info command with detailed analysis options
  • Add individual llm log preparation
  • Sanitize by default
  • Improve request parsing for name, ID, and direction detection
  • Add performance analysis command to CLI
  • Add installation scripts and Claude Code skill for log analysis
  • Improve install script for user-friendliness and compatibility
  • Add plugin support for Claude Code and update documentation

Add Claude Code plugin support for cross-project skill installation:

  • Add .claude-plugin/plugin.json manifest to enable plugin distribution
  • Add .claude-plugin/marketplace.json for plugin marketplace discovery
  • Create skills/ symlink to support both plugin and project-level usage
  • Users can install with /plugin marketplace add then /plugin install log-analyzer
  • Update documentation with plugin installation instructions in README.md and CLAUDE.md

Fixes

Improve installation workflow and documentation:

  • Fix scripts/install-skill.sh to use repository directory instead of current working directory
  • Change default install location from /usr/local/bin to $HOME/bin (no sudo required)
  • Add automatic PATH setup instructions for zsh, bash, and fish shells
  • Recommend WSL for Windows users instead of native binary
  • Rewrite README.md to be more compact and user-friendly (~50% smaller)
  • Add Claude Code Integration section with /analyze-logs skill examples
Loading