Skip to content

Releases: ddv1982/csv-data-anonymizer

CSV Anonymizer v1.0.22

19 Jun 08:00

Choose a tag to compare

  • Rewrite the README into a shorter user-facing guide with clearer install, workflow, privacy-boundary, and development sections.
  • Add concise Local LLM usage documentation for the optional Ollama and Gemma 3 4B Smart replacement flow, including setup, preview, fallback, and data locality notes.

CSV Anonymizer v1.0.21

19 Jun 06:43

Choose a tag to compare

  • Add optional Local AI smart replacement for selected columns through Ollama and Gemma 3 4B, including status checks, model download controls, and validated replacement maps with rule-based fallbacks.
  • Rename the end-user deterministic setting to Repeatable replacements and clarify that the seed is private and useful for matching replacements across files.
  • Extend privacy reporting with smart replacement columns, generated smart values, and fallback counts while preserving the existing local-first anonymization warnings.
  • Keep model weights and local runtime binaries out of source and release metadata, with checks that block tracked .gguf, model-cache, Ollama-cache, and llama-server artifacts.

CSV Anonymizer v1.0.20

19 Jun 05:53

Choose a tag to compare

  • Add run-scoped pseudonym mapping so repeated source values reuse the same replacement and distinct readable names avoid duplicate assignments while pool capacity remains.
  • Expand readable first/last-name pools, preserve first/last/full-name consistency through shared semantic domains, and report avoided candidate collisions plus pool-exhaustion fallbacks.
  • Add a Tokenize strategy for opaque tok_... values and upgrade deterministic pseudonyms to keyed HMAC-SHA256 with seed-sensitivity notes in the privacy report.
  • Improve the privacy report UI and serialized report fields with unique pseudonym counts, repeated-source reuses, opaque-token counts, avoided collisions, and exhausted pseudonym pools.

CSV Anonymizer v1.0.19

19 Jun 00:10

Choose a tag to compare

  • Improve name pseudonymization so generic name columns with single-token values use name-like replacements instead of generic alphanumeric strings.
  • Make full-name replacements compose consistent first/last token pseudonyms and avoid preserving original full-name tokens where the local replacement pool allows it.
  • Add a representative people-name CSV fixture and regression tests for name detection, preview output, token consistency, and replacement quality.

CSV Anonymizer v1.0.18

18 Jun 23:21

Choose a tag to compare

  • Improve anonymization quality with numeric-shape-preserving values, expanded sensitive type detection, type-specific phone/name strategies, per-column type and strategy controls, preview warnings, and a privacy report that distinguishes masking/pseudonymization from stronger anonymization guarantees.
  • Fix Linux Debian package icon readiness by bundling standard hicolor PNG sizes and validating installed desktop Icon= resolution during release checks.

CSV Anonymizer v1.0.17

18 Jun 20:53

Choose a tag to compare

  • Add esbuild as an explicit frontend dev dependency so clean CI and release npm ci installs can run the Vite 8 production build.
  • Supersede v1.0.16 without moving the failed tag, keeping the same background job, file access, cleanup, and CI hardening changes.

CSV Anonymizer v1.0.15

18 Jun 18:31

Choose a tag to compare

  • Make CSV opening feel immediate by sampling columns first and moving the exact row count into a follow-up background command.
  • Move CSV analyze, preview, row-count, and anonymize work onto Tauri blocking tasks so synchronous file I/O does not tie up the async command runtime.
  • Fix remembered file-picker directories so they seed the native dialogs without being shown as fake input or output file paths.
  • Add clearer opening/loading button states and regression coverage for the sampled analysis path.

CSV Anonymizer v1.0.14

18 Jun 14:29

Choose a tag to compare

  • Supersede v1.0.13 with RPM metadata validation that can extract Tauri RPM payloads through bsdtar when rpm2cpio cannot.
  • Install libarchive-tools in Linux packaging workflows so Debian/RPM metadata validation keeps inspecting package contents.
  • Keep the v1.0.13 Linux package identity fix for csv-anonymizer.desktop.

CSV Anonymizer v1.0.11

18 Jun 08:52

Choose a tag to compare

  • Polish the native Rust desktop interface used by the Linux AppImage, Debian/RPM packages, tarball, and macOS app.
  • Add app-level egui theme styling, structured sections, status chips, clearer column risk badges, and a stronger primary anonymization action.
  • Keep the release packaging unchanged while making the shipped native UI look complete without relying on web CSS assets.

CSV Anonymizer v1.0.10

18 Jun 08:05

Choose a tag to compare

  • Rewrite the desktop app as a native Rust eframe/egui application with a shared Rust anonymization core.
  • Replace Bun/Electrobun/Vue app builds with Rust CI, Rust smoke checks, and Rust release artifacts under dist/rust.
  • Add Rust macOS .app/.dmg packaging and Linux .deb, .rpm, AppImage, portable tarball, and signed APT repository packaging.
  • Persist native app settings, remembered file-picker directories, deterministic seed, overwrite behavior, sample counts, and output suffix.
  • Keep Cmd+Q close handling in the native app and add output-folder reveal after successful anonymization.