Commit graph

11 commits

Author SHA1 Message Date
Daniel
4d92488f0c fix: translated answers keep their source chips; the patient take home can be translated
Translation
- Stop scrubbing markdown before sending it to LibreTranslate. The scrub
  deleted ordered-list numbering ("1. Give amoxicillin" -> "Give amoxicillin"),
  flattened tables into ambiguous whitespace and ate underscores inside
  identifiers. Raw markdown now goes to the translator unchanged.
- Render the translation through the same markdown pipeline as the original
  bubble, with the message's own sources, so [n] markers come back as the usual
  clickable .assistant-cite chips instead of escaped literal text. Headings,
  lists and tables survive with them.
- When the translator drops citation markers, surface the affected sources in a
  recovery block rather than letting the evidence disappear.
- Image cards are live nodes: they are now re-attached on every path out of a
  translation (success, failure and Show original), so a failed translation no
  longer silently removes a generating image from the message.

Patient take home
- Add a language selector to the take-home modal, reusing the existing
  /translate endpoint and offering only what the local LibreTranslate reports.
- Copy, Export and Email carry what the caregiver is actually reading; the
  original stays canonical behind "Original".

Conversation budget
- The admin field no longer prefills with the environment value, which turned
  the next Save into an accidental override and made the documented "leave
  empty to use the environment" path unreachable. The effective limit is shown
  as a placeholder instead.
- Report source 'default' honestly instead of naming an unset env var.
- The load-failure notice now lands on the <p> instead of an <input>'s
  textContent, where it rendered nothing.
- One validator for the budget everywhere: conversationLimit() replaces a
  parseInt that accepted "120000abc".

Other
- /assistant is addressed by its URL, not by ped_last_tab, so "/" no longer
  reopens the assistant; the URL follows tab changes and Back leaves it.
- Remove the dead DeepL path (it referenced an undefined DEEPL_BASES) and stop
  offering admins a provider the server silently ignores.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BkfrkQwA4YGrGw9LZSpeAq
2026-09-09 18:40:23 +02:00
Daniel
9a42b96ec9 feat: admin controls the conversation budget — saved override wins over CLINICAL_ASSISTANT_CONVERSATION_CHARS, validated 1000-1000000, empty falls back to env
All checks were successful
Forgejo Android APK / Root app tests (push) Successful in 58s
Forgejo Android APK / Build signed APK (push) Successful in 2m7s
2026-09-09 17:49:57 +02:00
Daniel
598731e170 fix: conversation budget is env-only — admin panel shows CLINICAL_ASSISTANT_CONVERSATION_CHARS read-only and never saves the key; server ignores saved overrides
All checks were successful
Forgejo Android APK / Root app tests (push) Successful in 48s
Forgejo Android APK / Build signed APK (push) Successful in 2m17s
2026-09-09 17:32:15 +02:00
Daniel
5e6de10d0a fix: delegated model-selection persistence, admin-settable conversation budget (DB overrides env), unbounded image poll 2026-09-09 03:20:21 +02:00
Daniel
b21080f2a6 feat: admin allowlists for chat and image models (users see the model pill once configured) 2026-09-09 02:58:30 +02:00
Daniel
8b072496e2 feat: Open WebUI-style assistant workspace — 3-column layout, markdown/math/code/tables, autosave with images, translation (LibreTranslate+DeepL), citation modal, Learning Hub moved in, handoff removed 2026-09-08 18:52:35 +02:00
Daniel
13fca7887d feat: unify Admin model/prompt UI — single prompt dropdown, two image-model dropdowns, searchable/scrollable users, remove classification/custom-add, consistent font, concise labels 2026-09-07 18:18:48 +02:00
Daniel
bb82295173 fix: single image-model dropdown applied to both workflows; concise budget labels 2026-09-07 17:58:28 +02:00
Daniel
6be2d1375a feat: integrate durable image jobs/private assets into current core 2026-09-07 16:53:42 +02:00
Daniel
c09038bfec fix: compact prompt administration with robust admin loading and concise styling 2026-09-07 16:11:49 +02:00
Daniel
cfaf8e957b feat: ship reviewed prompt history, conversation limits and account protections 2026-09-07 04:01:01 +02:00