v1.0.0

v1.0.0 — First Stable Release

CorpusMind

DOI: 10.5281/zenodo.21226650

CorpusMind Lens

DOI: 10.5281/zenodo.21673083

---

Full release notes: https://github.com/waleedmandour/CorpusMind/releases/tag/v1.0.0

Downloads:

• CorpusMind (main app): macOS, Windows, Linux, Web PWA
• CorpusMind Lens: macOS, Windows, Linux
• User Guide PDF (15 pages, covers both apps)

Verification: 205 tests pass, ruff clean, typecheck + build clean, cargo check passes for both Tauri shells.

v0.1.27

What's Changed

• feat(lens): vision-message providers (step 1) + image-set manager UI (step 2) by @waleedmandour in #1
• feat(lens): POST /images/{img_id}/describe route (build step 3) by @waleedmandour in #2
• feat(lens): vision-LM sibling paths for 8 discourse routes (build step 4) by @waleedmandour in #3
• feat(lens): consent-gate filter for all vision-LM output (build step 5) by @waleedmandour in #4
• feat(lens): alignment inspector UI + LLM alignment mode (build step 6) by @waleedmandour in #5
• feat(lens): batch view — recurring themes + OCR frequency (build step 7) by @waleedmandour in #6
• feat(lens): new Tauri shell for CorpusMind Lens (build step 8 — final step) by @waleedmandour in #7

New Contributors

• @waleedmandour made their first contribution in #1

Full Changelog: v0.1.26...v0.1.27

v0.1.26

Full Changelog: v0.1.25...v0.1.26

v0.1.25

Full Changelog: v0.1.24...v0.1.25

CorpusMind v0.1.24

Remove arTenTen + progress bar in status bar

Full Changelog: v0.1.23...v0.1.24

CorpusMind v0.1.23 — Full Corpus Availability Fix

Fix: full corpora now show as Downloadable

Full Changelog: v0.1.22...v0.1.23

CorpusMind v0.1.20 — Full Reference Corpora (BNC Baby, BAWE, Leipzig)

Phase 2: Full Reference Corpora

New Full Corpus Downloads

• BNC Baby (4M words, BNC User Licence) — 4 subcorpora: academic, fiction, news, spoken
• BAWE (6.5M words, CC BY-NC-SA 3.0) — academic English with discipline metadata
• Leipzig English News 10K (CC BY 4.0) — full sentences with NLP pipeline
• Leipzig Arabic News 10K (CC BY 4.0) — Arabic news with NLP pipeline

New Endpoint

POST /reference-corpora/{name}/download-full — downloads archive, extracts text, creates Corpus row, ingests through NLP pipeline, tags documents with genre metadata

UI Improvements

• Full Corpus vs Frequency List badge in BundledReferences
• Auto-detection of format (full_corpus calls download-full endpoint)

Verified

• ruff: 0 errors
• pytest: 24/24 passed
• typecheck: clean
• build: clean

Full Changelog: v0.1.19...v0.1.20

CorpusMind v0.1.19 — ACL Fix + Subcorpus + Metadata Tagging

New Features

ACL Fix

• Replaced window.confirm() with custom ConfirmDialog (fixes Tauri 2 ACL error)

Subcorpus

• Saved filters over document metadata (genre, register, year)
• Create/list/delete subcorpora via API
• Subcorpus creation form in document list
• Subcorpus chips with delete buttons

Document Metadata Tagging

• PATCH /corpora/{cid}/documents/{did}/meta endpoint
• Genre, register, year fields per document
• Inline editing in document table (✎ Tag button)

Professional Document Table

• Genre and Register columns
• Inline metadata editing
• Subcorpus button

Verified

• ruff: 0 errors
• pytest: 24/24 passed
• typecheck: clean
• build: clean

Full Changelog: v0.1.18...v0.1.19

CorpusMind v0.1.18 — 6 Bug Fixes

Fixed

1. python-multipart + cryptography declared in pyproject.toml
2. ruff check passes (0 errors, was 40+)
3. export_queue crash fixed (session not passed to producers)
4. 9 analysis panels export in chosen format (was always JSON)
5. Document-delete/recompile stats now persist (SQLAlchemy JSON fix)
6. Arabic i18n wiring started (10 views have lang hook)

Verified

• ruff: 0 errors
• pytest: 24/24 passed
• typecheck: clean
• build: clean

Full Changelog: v0.1.17...v0.1.18

CorpusMind v0.1.17 — Reference Corpora + Document Management + Dark Mode

What's New

Reference Corpora

• 3 new downloadable reference corpora: Leipzig English News, Quranic Arabic, CAMeL Arabic
• Select downloaded reference for keyness analysis across the platform

Document Management

• Delete individual files from a corpus
• Recompile corpus (re-run NLP pipeline)
• Professional table layout for document list

UX Improvements

• Auto-load last active corpus on startup
• Dark mode button contrast fixes

Verified

• 24/24 engine tests pass
• Web typecheck + build clean
• cargo check clean

Full Changelog: v0.1.16...v0.1.17