Release 0.3.0

CHANGELOG.md

@@ -11,12 +11,13 @@ documentation, and CI work are left out.
 
 ## [Unreleased]
 
+## [0.3.0] — 2026-08-20
+
 ### Added
 
 - Optional sign-in at the organisation's OpenID Connect provider, for
-  deployments without an authenticating proxy in front: `OIDC_ENABLED` plus
-  the client credentials and `APP_PUBLIC_URL` gate every API route on a
-  signed session cookie. No accounts, no roles — see
+  deployments that have no authenticating proxy in front. Set `OIDC_ENABLED`
+  plus the client credentials and `APP_PUBLIC_URL` — see
   [Single sign-on](docs/operations/sso.md).
 
 ## [0.2.1] — 2026-08-19
@@ -40,26 +41,19 @@ documentation, and CI work are left out.
 
 ### Added
 
-- `VISION_OCR_DIALECT` selects the vision-OCR model family (`unlimited_ocr`,
-  `chandra`, `plain`) and supplies the prompt, token and body defaults — see
-  [OCR engines](docs/operations/ocr-engines.md).
-
-- `VISION_OCR_PROFILES` offers several OCR models at once, one pickable per
-  document in the advanced settings — see
-  [OCR engines](docs/operations/ocr-engines.md).
-
-- `compose.chandra.yml` runs a datalab chandra GPU OCR sidecar, as
-  `compose.unlimited-ocr.yml` does for Unlimited-OCR — see
+- Vision OCR gained `VISION_OCR_DIALECT` (model family `unlimited_ocr`,
+  `chandra` or `plain`, each supplying its own prompt, token and body
+  defaults), `VISION_OCR_PROFILES` (several models at once, one pickable per
+  document), and `compose.chandra.yml` for a chandra GPU sidecar next to the
+  existing `compose.unlimited-ocr.yml` — see
   [OCR engines](docs/operations/ocr-engines.md).
 
 ### Changed
 
-- `.env.example` is now a short setup worksheet that ships with the LLM
-  detector enabled: fill in `OPENAI_API_BASE`/`LLM_MODEL` before first start
-  (or set `DETECTORS=rules` for a look without an LLM) — see
-  [Configuration](docs/operations/configuration.md).
-
-- `LLM_REQUEST_TIMEOUT_SECONDS` now defaults to `600` instead of `120` — see
+- `.env.example` is now a short setup worksheet shipping with the LLM detector
+  enabled: fill in `OPENAI_API_BASE`/`LLM_MODEL` before first start, or set
+  `DETECTORS=rules` for a look without an LLM.
+- `LLM_REQUEST_TIMEOUT_SECONDS` defaults to `600` instead of `120`. Both:
   [Configuration](docs/operations/configuration.md).
 
 ## [0.1.3] — 2026-08-07
@@ -153,6 +147,7 @@ First tagged version.
   configurable deployment banner sits above the header.
 - Documentation site (MkDocs Material) under `docs/`.
 
+[0.3.0]: https://github.com/KatherLab/deidentifier/compare/v0.2.1...v0.3.0
 [0.2.1]: https://github.com/KatherLab/deidentifier/compare/v0.2.0...v0.2.1
 [0.2.0]: https://github.com/KatherLab/deidentifier/compare/v0.1.3...v0.2.0
 [0.1.3]: https://github.com/KatherLab/deidentifier/compare/v0.1.2...v0.1.3

CITATION.cff

@@ -45,8 +45,8 @@ keywords:
   - "privacy-preserving artificial intelligence"
 
 license: "AGPL-3.0-or-later"
-version: "0.2.1"
-date-released: "2026-08-19"
+version: "0.3.0"
+date-released: "2026-08-20"
 
 references:
   - type: article

THIRD_PARTY_NOTICES.md

@@ -1,6 +1,6 @@
 # Third-Party Notices
 
-The Medical Document Anonymizer (version `0.2.1`) is distributed
+The Medical Document Anonymizer (version `0.3.0`) is distributed
 under the `AGPL-3.0-or-later` license. It bundles the third-party
 open-source components listed below, each governed by its own license.
 Full license texts are available in each project's repository (linked

backend/src/core/config.py

@@ -41,7 +41,7 @@ class Settings(BaseSettings):
 
     # Application
     APP_ENV: str = "development"
-    APP_VERSION: str = "0.2.1"
+    APP_VERSION: str = "0.3.0"
     APP_MAX_UPLOAD_MB: int = Field(default=20, ge=1)
     APP_MAX_TEXT_CHARS: int = Field(default=500_000, ge=1)
     APP_ALLOW_INSECURE_CONTENT_LOGGING: bool = False

docs/operations/deployment.md

@@ -20,8 +20,8 @@ has been published to `ghcr.io`, a deployment host can skip the build
 toolchain entirely:
 
 ```bash
-DEIDENTIFIER_IMAGE_TAG=v0.2.1 docker compose pull
-DEIDENTIFIER_IMAGE_TAG=v0.2.1 docker compose up -d
+DEIDENTIFIER_IMAGE_TAG=v0.3.0 docker compose pull
+DEIDENTIFIER_IMAGE_TAG=v0.3.0 docker compose up -d
 ```
 
 You still need `compose.yml` and your `.env` on that host.

package.json

@@ -1,6 +1,6 @@
 {
   "name": "deidentifier",
-  "version": "0.2.1",
+  "version": "0.3.0",
   "private": true,
   "description": "Deidentifier frontend — anonymize German clinical documents locally.",
   "license": "AGPL-3.0-or-later",

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "deidentifier"
-version = "0.2.1"
+version = "0.3.0"
 description = "Locally deployable anonymizer for German clinical documents"
 requires-python = ">=3.13,<3.15"
 # Every dependency is pinned exactly, like the frontend's package.json: this is