v0.1.5

v0.1.5 - 2026-04-23

Added

• kawat: Public facade crate
  • kawat::extract(): Single-call HTML → plain text extraction
  • fetch_url() / fetch_url_async(): URL fetching helpers
• Golden test suite: 20 HTML fixtures compared against Python trafilatura 2.0
  • Similarity scoring via Jaccard + word-level containment
  • Average match: 91.2% (threshold: 70%)
• CI/CD enhancements
  • cargo-about license certification (about.toml)
  • Security audit workflow with cargo-audit + RUSTSEC advisory tracking

Changed

• cascade.rs: Removed dead with_metadata branch (single default path)
• containment(): Switched from substring to HashSet word-level matching

Fixed

• Word-level containment false positives ("a" matching inside "article")
• rustls-webpki vulnerability (RUSTSEC-2026-0104) via cargo update

v0.1.4

v0.1.4

Added

• kawat-core: Core extraction orchestrator (0.1.4 milestone)
  • cascade::run(): Full extraction pipeline with steps 1, 5, 6, 8a
  • ExtractorOptions: Configuration struct mirroring trafilatura Extractor
  • Document: Extracted document model with metadata and body text
  • Size validation with min_extracted_size threshold
• kawat-output: TXT output formatter
  • to_txt(): Plain text output with metadata header
  • to_txt_body_only(): Body text without headers
  • OutputFormat enum with #[non_exhaustive] for forward compatibility
  • OutputError custom error type using thiserror
• CI/CD: GitHub Actions workflow with cargo-deny
  • License audit with explicit allow list
  • Security advisory checking
  • Dependency banning for duplicates

Changed

• Document::to_formatted_string() now dispatches to format-specific handlers
• TXT output gates metadata header on with_metadata option (matches trafilatura behavior)
• Pre-commit hooks include cargo-deny for local license checking

Fixed

• Deprecated deny key removed from cargo-deny config
• Added missing licenses: Unicode-3.0, Zlib

Release v0.1.3

Changes

Installation

From crates.io (manual publishing)

cargo install kawat

From source

cargo install --git https://github.com/SHA888/kawat.git --tag v0.1.3

Downloads

Check the assets below for pre-compiled binaries.

Note

Crates.io publishing is done manually to ensure quality control.

What's Changed

• cargo(deps): update dom_smoothie requirement from 0.5 to 0.17 by @dependabot[bot] in #13
• ci: bump codecov/codecov-action from 5 to 6 by @dependabot[bot] in #10
• ci: bump actions/upload-artifact from 4 to 7 by @dependabot[bot] in #11
• ci: bump peter-evans/create-pull-request from 6 to 8 by @dependabot[bot] in #12

Full Changelog: v0.1.2...v0.1.3

v0.1.2

Added

• kawat-html: Complete HTML tree cleaning and tag normalization pipeline
  • tree_cleaning(): Remove 44 MANUALLY_CLEANED tags and strip 20 MANUALLY_STRIPPED tags
  • convert_tags(): Normalize HTML tags to internal catalog (h1-h6→head, b/strong/em/i→hi, a→ref, ul/ol→list, li→item, br→lb, blockquote→quote, del/s→del, code/pre→code)
  • convert_link(): Resolve relative URLs against base_url using standards-compliant URL resolution
  • _is_code_block(): Distinguish between inline code and code blocks
  • handle_textnode() + process_node(): Text extraction and normalization for all element types
  • link_density_test() and link_density_test_tables(): Link density filtering for content extraction
  • delete_by_link_density(): Remove high-density link elements with backtracking
• kawat-extract: Custom KawatTree structure for lightweight HTML processing
  • KawatNode and KawatTree structs with full traversal and manipulation methods
  • HTML parsing with proper text/tail distinction
  • Integration with kawat-html transformations
  • 23 comprehensive unit tests

Changed

• Improved HTML processing pipeline with immutable-first design
• Enhanced error handling with proper Result types throughout

Fixed

• Lifetime syntax errors in tree.rs (explicit '_ lifetime parameters)
• Test failures in convert_link and textnode modules
• Inline code formatting (missing closing backtick)

Testing

• 34 kawat-html unit tests (all passing)
• 23 kawat-extract unit tests (all passing)
• 57 total tests across all crates (all passing)
• Pre-commit hooks: Rust Format, Clippy, Cargo Audit (all passing)

v0.1.1

v0.1.1 - 2026-03-24

Added

• Comprehensive XPath evaluation engine with CSS selector fallback
• Benchmark suite for XPath performance testing
• Extensive test coverage (15+ tests) for all BODY_XPATH expressions
• Support for complex XPath patterns including (article)[1]

Changed

• Updated dependencies: lru (0.12→0.16), quick-xml (0.37→0.39), scraper (0.22→0.26), criterion (0.5→0.8), reqwest (0.12→0.13)
• Improved TLS configuration in reqwest (rustls-tls → rustls)
• Enhanced error handling in XPath evaluation

Fixed

• XPath fallback test failures for (article)[1] pattern
• CI/CD pipeline issues with deny.toml configuration
• Codecov upload failures for protected branches
• Security workflow permission handling

v0.1.0

• Initial workspace import of kawat crates (core, extract, html, xpath, metadata, output, CLI, etc.).
• Added date parsing regex fallback and chrono clock feature for htmldate-rs.
• Implemented FromStr for OutputFormat and hardened CLI format parsing.
• Resolved clippy warnings (unused imports/vars, format args) across crates.
• Added .gitignore and .pre-commit-config.yaml (fmt, clippy, cargo-audit) and ensured hooks pass.