v0.4.0

What’s changed

Compare two papers

• Select exactly two library PDFs and turn Compare on.
• The turn binds CompareResearch only (not SearchResearch / FetchResearch).
• Retrieves top chunks per paper (COMPARE_TOP_N, default 4), then TypeSafe Jev 1.13 labels each pair: corroboratory, contradictory, or neutral.
• Keeps pairs above COMPARE_MIN_CONF (default 0.70).
• Side panel shows compared pairs (choice, confidence ring, both passages). Click a passage to highlight the PDF. Ask pins both chunks and turns Compare off for the follow-up.
• Needs TYPESAFE_API_KEY. If Jev fails, the turn stops with Compare is unavailable

Persistence

• New migration adds pair_id, choice, and side on sources.
• History reloads pairs; refresh still shows the pair panel (toggle is still in-memory, restored from loaded pairs when possible).
• New chats take the active default from llm_models (gpt-6-luna on this DB).

Docs

• README: compare feature, TypeSafe Jev, TYPESAFE_API_KEY.
• Screenshot: screenshots/compare.png.

Full Changelog: v0.3.1...v0.4.0

v0.3.1

What’s new

Citation counts. OpenAlex cited_by_count shows on paper search results and in the paper panel.

OpenAlex links. Paper chips open from a PDF URL or a landing-page URL. A missing or unverified PDF no longer hides the chip.

Dedup. The same paper from OpenAlex and arXiv is merged once. Citation count and PDF URL are kept when one source has them and the other does not.

Upgrade

Restart the backend once so the SQLite migration adds cited_by_count. No new .env keys. Counts show on the next search; you do not need to re-upload PDFs.

git fetch
git checkout v0.3.1

Or stay on `main` after `git pull`.

## Stay on the previous version

```bash
git fetch
git checkout v0.3.0

Full Changelog: v0.3.0...v0.3.1

v0.3.0

What’s new

OCR for scanned PDFs. Off by default. Set DO_OCR=true in backend/.env to OCR image-only pages on ingest. Leave it false for normal text PDFs (faster).

Figures and diagrams. Ingest describes plots and pipelines so chat and summaries can cite them. Click a citation to highlight the figure on the page.

Upgrade

Copy DO_OCR=false into backend/.env if you don’t have it yet. Re-upload PDFs if you want figure search or OCR; old vectors won’t have those.

git fetch
git checkout v0.3.0

Or stay on main after git pull.

Stay on the previous version

git fetch
git checkout v0.2.0

What's Changed

• fix: preserve conversations when server deletion fails by @kalelooz in #1

New Contributors

• @kalelooz made their first contribution in #1

Full Changelog: v0.2.0...v0.3.0

v0.2.0

What’s new

Full / Focused highlights. Click a citation to jump the PDF preview. Focused (default) paints a tighter highlight on the cited passage. Full highlights the whole retrieved block. Toggle in the PDF header.

Retrieval still uses a small span; the model still sees the surrounding passage.

Upgrade

Re-upload your PDFs after pulling this release. Old vectors don’t have the new highlight boxes, so Focused will look empty or wrong until you ingest again.

git fetch
git checkout v0.2.0

Or stay on main after git pull.

Stay on the previous version

git fetch
git checkout v0.1.0

Full Changelog: v0.1.0...v0.2.0

v0.1.0

What’s in this release

The version currently on main: chat over your PDFs, on-page highlights, reader mode, summaries, OpenAlex search, and citation export.

Stay on this version

Already cloned:

git fetch
git checkout v0.1.0

New clone:

git clone https://github.com/Sreehari05055/thesys-core.git
cd thesys-core
git checkout v0.1.0

Next

A later release will add Full / Focused PDF highlights. After that update you will need to re-upload PDFs.

Full Changelog: https://github.com/Sreehari05055/thesys-core/commits/v0.1.0