Zone de dépôt : le contour s'allume en doré et la carte s'agrandit
quand on la survole avec un fichier (l'information existait, elle
n'était jamais affichée).
Or lisible : nouveau jeton « or encre » (#8B6F47) pour les textes,
l'or clair reste aux traits, fonds et au mode sombre ; 52 textes dorés
convertis, pastille comprises ; plus aucun texte sous 10 px
(8,5 px → 11 px dans le sélecteur de moteurs).
Tarifs : le choix redevient binaire — les trois abonnements payants en
grandes cartes, le gratuit en ligne discrète au-dessus, le sur mesure
en ligne dorée dessous. Le tableau comparatif complet reste dessous.
Accueil : le mur de six cartes génériques est remplacé par un vrai
avant/après — une page « Cahier des charges — Traitement d'air » à
côté de sa traduction anglaise, structure identique au mot près
(terme technique en gras des deux côtés), sceau « Même mise en page,
mot pour mot » entre les deux, et trois preuves précises (SmartArt
reconstruits, séries de graphiques traduites, tables des matières
régénérées) à la place des promesses vagues.
Batch upload: dropzone and inputs accept multiple files (cap 10/run,
validated + deduped); 2+ files switch the submit to a queue screen —
sequential jobs with per-file progress, failure messages through the
humanizer, per-file download/review links, staggered download-all,
real server-side stop (cancel API) plus between-files abort. The
one-file flow is untouched: files.length === 1 takes the exact
existing path.
Runner: standalone runTranslationJob (submit + poll + abort + cancel)
with injectable poll interval; 5 dedicated vitest cases covering
completed-with-progress, backend failure, submit HTTP error, mid-flight
abort with server cancel, and network-failure cutoff.
Context page close review: /dashboard/context is a clean redirect to
glossaries (no duplicate); the fake 300ms save delay removed (instant
zustand persist); suggestion chips now resolve their prompt bodies
through i18n (EN+FR — no more French prompts for English users); the
translate config column shows a 'Context guidelines active' chip that
links to the editor when a Pro LLM prompt is set.
Dead code: legacy file-uploader.tsx (+webllm.ts, its only consumer)
and the unused PRESETS/applyPreset/clearContext store block removed
(-14KB of glossary strings).
Verified: build exit 0, vitest 14/14 (9 prior + 5 runner), eslint 62
errors (vs 64 at HEAD), 0 missing i18n keys.
Pricing: full editorial redesign — serif card headers with accent pills
replace the colored font-black blocks, tone sweep across toggle/metrics/
features/CTAs, PLAN_COLORS removed; one design system app-wide.
Translate: decorative titles one step down (CTA hierarchy restored);
glossary and image-translation blocks hidden entirely for free users
(progressive disclosure — three controls for free).
Reviews: XLIFF hint line explains the exchange format; backend errors
routed through a friendly mapper (session/not-found/rate-limit/server).
Landing: fabricated hero UI cards (fake 'Context Engine' overlay)
removed — the photo no longer promises screens that don't exist.
Nav: single DashboardNavLinks component shared by sidebar and mobile
drawer (was duplicated markup).
API: GET /api/v1/translations (user job history, paginated; completed
jobs retained 24h) and POST /api/v1/translations/{id}/cancel —
cooperative cancellation with worker checkpoints before dispatch and
before finalisation, reserved quota released immediately. Translate
monitor now offers a real 'Cancel translation' next to 'Back to start';
recent-jobs list reads server history first, localStorage fallback.
DeepL purge (backend): provider module, registry registration, config
attrs/defaults, dispatch branch, admin settings schema + test branch,
legacy availability block, validation rules, plan provider lists,
error-code mappings, MCP enums, translator prompt mention, related
tests updated/removed. Fallback resolver skips unknown providers, so
stale chains containing 'deepl' degrade gracefully.
Verified: backend 110 tests passed; frontend build exit 0, vitest 9/9,
0 missing i18n keys, eslint 63 errors (vs 64 at HEAD).
Pricing: 'deepl' removed from plan provider lists and the comparison
table row. Landing and pricing copy (13 locales) no longer mention
DeepL (65 values cleaned, 39 dead keys deleted: pricing.comparison.deepl,
providerTheme.classic.deepl.*).
Providers: available-provider responses are filtered so a DeepL entry
from the backend can never surface in the engine picker or the services
page; DeepL theme entry and static lib/api.ts list entry removed.
Admin: DeepL config card, type fields, defaults and fallback-chain
mentions removed; stats/chart/status maps and mock data cleaned.
Backend adapter untouched — DeepL is invisible app-wide via the UI
filter; removing the Python adapter itself is a separate step if wanted.
Verified: build exit 0, vitest 9/9, eslint clean on touched files,
zero 'deepl' occurrences outside the two intentional UI filters.
Nav: teams/settings/services back out of the nav until product-ready (user
decision); teams page was already half-finished (UUID member display is a
backend limitation).
Payment: explicit confirmation dialog (plan, amount, billing period, Stripe
note) before any redirect; ?plan= URL now pre-opens the dialog instead of
triggering a silent checkout.
Contrast AA sweep (71 replacements): functional micro-labels raised to
>=60-65% opacity and >=10px across sidebar, header, pricing, landing;
decorative all-caps 'interface' label removed.
Design unification: new PageHeader component (accent pill + serif
base/accent title) applied to settings, services, reviews, teams; pricing
header returned to the editorial voice (serif + accent pill); dead
GlossaryCard deleted.
i18n residuals: suggestion chips, 'Standard' provider label, notification
close, model-combobox strings extracted (+broken bg-surface/border-border-
subtle tokens fixed); ~45 new keys EN+FR; 13 dead keys removed; zero
missing keys verified.
Reviews: icon-only row actions now carry visible text; 'Approve all' is
two-step armed-confirm.
Accelerators: Ctrl/Cmd+Enter submits; arrow-key navigation in the language
combobox; recent-jobs history cap 8->20 with per-job download; source=target
config rejected; empty 'Master Quality' badge removed.
Cookie consent: reopenable via footer link, emoji replaced with drawn icon.
Verified: build exit 0, vitest 9/9, eslint 64 errors = previous level,
0 missing i18n keys.
Foundations:
- TranslationSegment model + migration f7e8d9c0b1a2 (segments, workspaces,
workspace_members, glossaries.workspace_id)
- SegmentRecorder injected into all 4 translators: unique (source,
translation) pairs captured per job and persisted (best-effort)
- set_segment_overrides: human-reviewed translations applied verbatim on
rebuild — top priority over TM and provider, zero API calls
Review API (routes/review_routes.py):
- GET /translations/{id}/segments (owner or job token)
- PATCH /segments/{id} edit/approve — feeds the per-user TM so approved
translations are reused in later jobs
- POST /translations/{id}/rebuild — rebuild document with reviewed text
- GET/POST /translations/{id}/xliff — XLIFF 1.2 export/import (edited
segments export their reviewed text)
Review editor (frontend /dashboard/reviews/[jobId]):
- side-by-side source/translation table, inline edit, approve (single or
all), rebuild & download (auth blob), XLIFF export/import, 13 locales
- 'Relire et corriger' link on the translation-complete screen
Team workspaces (routes/workspace_routes.py + /dashboard/teams):
- Workspace/WorkspaceMember models, roles owner/admin/member
- create (Business plan), list with seat usage, invite by email with
seat-limit enforcement (Business=5, Enterprise unlimited), removal
- shared glossaries: workspace members can use a glossary shared to their
workspace (access check extended)
Tests: 1184 passed / 0 failed (11 new: recorder, overrides, docx
capture->rebuild e2e, XLIFF structure/escaping, seats, workspace CRUD,
shared glossary access)
Multiple translation bugs in Word/Excel/PPTX that caused chart elements
to be left untranslated or charts to render as empty series.
Backend
-------
* providers (deepseek/openai/minimax): tighten system prompt so the LLM
actually translates chart titles/axis labels/legend/category labels,
month abbreviations, and clarifies what counts as a 'real' proper noun
(people/place/company/product names) vs. technical labels. Old rule
'keep proper nouns unchanged' was being read too broadly by the model
and caused chart text to be skipped.
* excel_translator.py:
- Sheet reference rewrite: when a sheet is renamed, the chart XML's
c:f refs (e.g. 'Sales 2024'!$D$2:$D$61) are now rewritten to the
new name with proper apostrophe escaping (Chiffre d'affaires ->
'Chiffre d''affaires') and auto-quoting when the new name contains
spaces or special chars. Without this, the chart points at a sheet
that no longer exists and renders 0/empty series.
- Sheet name offset bug: sheet_name_offset was computed after chart
text was appended to text_elements, causing sheet names to receive
chart text translations. Now captured BEFORE sheet names are added.
* New tests:
- test_excel_chart_sheet_refs.py (10 unit tests, synthetic inputs)
- test_chart_translation_prompt.py (3 contract tests on the prompt)
Frontend
--------
* DashboardSidebar / translate/page: hide the Memento promo section
for paying users (tier != 'free').
* constants.ts: temporarily comment out the 'CLES API' nav item.
Update constants.test.ts to match the new state.
All fixes are generic - no file-specific hardcoding, edge cases covered
(empty mapping, missing bang, apostrophe escaping, partial renames,
multi-series).
The product name is 'Memento' (not 'Momento'). The 'Momento' typo was
present in 12 i18n locale files (en, fr, es, de, it, pt, nl, ja, ko,
zh, ru, ar, fa) and in code comments.
Also added a 'memento.url' i18n key with the canonical link
https://memento-note.com/ and wrapped the Memento promo cards in
both DashboardSidebar.tsx and translate/page.tsx with an <a> tag
pointing to that URL. Previously the cards were a non-clickable
<div>, so users had no way to reach the Memento product page.
Note: 'momento' is a legitimate word in Italian, Spanish and
Portuguese meaning 'moment' (e.g. 'a qualsiasi momento'). Those
occurrences in pricing.json and services.json were NOT changed —
they are correct translations of UI strings about 'at any moment'.
- Add template_id column to Glossary model (nullable, indexed)
- Backend: return 409 Conflict if user already imported a template
- Frontend: disable preset cards already imported, show 'Imported' badge
- Fix duplicated text in GlossarySelector source warning (hardcoded FR text removed)
- Complete i18n migration for glossaries page and GlossarySelector
- Add glossaries.presets.alreadyImported key in all 13 locales
47 new i18n keys added across all 13 locales (en, fr, es, de, pt, it,
nl, ru, ja, ko, zh, ar, fa). English and French are fully translated,
remaining locales use French as placeholder.
Files migrated:
- EditGlossaryDialog.tsx (18 strings)
- DeleteGlossaryDialog.tsx (7 strings)
- ProUpgradePrompt.tsx (10 strings)
- WebhookSnippet.tsx (4 strings)
- TranslationModeToggle.tsx (8 strings)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Two issues found in screenshot review:
1. Glossary preview always showed English 'target' field (foie → liver)
instead of the language-specific translation. Now looks up
term.translations[targetLang] first, falls back to term.target.
When target is Persian, preview shows 'foie → کبد' (Persian).
2. Previous migration b7c8d9e0f1a2 was already applied on server before
the rename was added. New migration c8d9e0f1a2b3 handles the rename
separately: glossaries with target_language='multi' get their name
changed from 'Anglais' to 'Multilingue'.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Full audit found 18 issues across backend, frontend, and data files.
Root cause: target_language had no single source of truth — 'en' was
hardcoded as default in 6+ places while templates are actually multilingual.
Fixes applied:
- routes/glossary_routes.py: list_glossaries() fallback 'en' → 'multi'
- routes/glossary_routes.py: import reads target_lang from index.json
(source of truth) instead of template file
- data/glossaries/*.json: all 8 template files target_lang 'en' → 'multi'
- services/glossary_service.py: get_glossary_terms() now returns
target_language field (was missing entirely)
- schemas/glossary_schemas.py: all defaults 'en' → 'multi'
(GlossaryCreate, GlossaryResponse, GlossaryListItem)
- useTranslationConfig.ts: only reset glossary on sourceLang change,
not targetLang (multilingual glossaries work with any target)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Templates enriched by enrich_glossary_templates.py already contain
translations for de, es, it, pt, nl, ru, ja, ko, zh, ar, fa (including
Persian). But they were labeled FR→EN, causing incorrect filtering and
warnings when translating to other languages.
Changes:
- index.json: set target_lang='multi' for all 8 templates
- GlossarySelector: treat target_language='multi' as compatible with
any target language (no false warnings, auto-select works)
- GlossarySelector: display '🌐 MULTILINGUE' badge instead of EN flag
- glossary_routes: default target_language to 'multi' instead of 'en'
- Migration: detect existing multilingual glossaries in DB (5+ keys in
translations JSON) and set their target_language to 'multi'
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
When activating the glossary toggle, only auto-select a glossary if it
matches BOTH source and target language. If no matching glossary exists
(e.g. translating to Persian but only have FR→EN glossaries), leave the
selector empty so the user picks manually instead of forcing an
incompatible glossary.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- GlossarySelector: filteredGlossaries now sorts matching target_language
first so FR→FA appears before FR→EN when translating to Persian
- GlossarySelector: toggle auto-select prefers glossary matching
targetLang, falls back to first available
- useTranslationConfig: reset glossaryId when targetLang changes (was
only resetting on sourceLang change)
- glossary_routes: set target_language from template data on import
(was missing, defaulted to None/en)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- US1: Replace exposed provider names (DeepSeek V3 → IA Express, MiniMax → IA Avancée)
in legacy_routes.py, remove 52 DeepSeek references from i18n pricing/landing keys,
add generic modelName key across all 13 languages
- US2: Fix glossary selector filtering (fallback to all glossaries when source lang
filter returns empty), show templates by default, fix 3 broken presets
(hvac/construction/automotive → hr/scientific/ecommerce)
- US3: Replace 15 hardcoded French strings with i18n t() calls, increase font sizes
from 8-9px to 11px, improve text contrast (opacity /20→/40, /25→/45, /30→/50)
- US4: Add missing provider labels in admin ProviderStatus (deepseek, minimax, zai,
google_cloud, openrouter, openrouter_premium)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- Restructure right sidebar as flex-col with scrollable config and
fixed translate button (never scrolls away)
- Replace expanded GlossarySelector list with compact dropdown (~80px
instead of ~400px), collapsible templates section, click-outside close
- Add visible Classic/IA mode badge after provider selector with hint
about glossary availability for Pro users in classic mode
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The backend expects template_id as a Query parameter, but the frontend
was sending it as a JSON body — causing validation error on every click.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
- Round check icon on selected glossary (unmissable)
- Error banner when import fails (no more silent spinner)
- Click template that already exists = select it (no re-import)
- Single fetch for glossaries + templates (faster load)
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
- Enriched 8 glossary templates with 18,191 translations across 11 languages
using LLM batch generation + back-translation validation (99.98% confirmed)
- Rewrote GlossarySelector as inline section with template creation
- Fixed sidebar duplicate (single Glossaries link with proOnly flag)
- Added glossaryId reset when sourceLang changes
- Always show GlossarySelector (locked with Pro badge for free users)
- Added source_language flag on glossary cards
- Redirected /dashboard/context to /dashboard/glossaries
- Updated import endpoint to read translations from templates
- Added enrichment script (scripts/enrich_glossary_templates.py)
- Added 6 i18n keys across all 13 locales
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
- Add system prompt textarea and professional presets (HVAC, IT, Legal, etc.) to Glossaries page
- Remove Context from sidebar navigation (constants.ts)
- Make GlossarySelector always visible for Pro users (not just LLM mode)
- Send system prompt from Zustand store to backend via custom_prompt
- Add 24 new i18n keys across all 13 locales for glossaries page
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
- TermEditor: rewritten with expandable multilingual translation grid
(13 languages), editorial styling, source/target + translations JSON
- GlossarySelector: new component in translate page config panel,
fetches user glossaries, shows flag + term count, Pro+LLM only
- useTranslationConfig: added glossaryId state
- useTranslationSubmit: sends glossary_id to backend
- Context page: removed textarea glossary, presets now create API
glossaries via template import, added link to Glossaries page
- i18n: added 12 keys × 13 locales for glossary/translate/context
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
- Restructured docker-compose for Nginx Proxy Manager (no custom nginx)
- Added domain wordly.art configuration
- Added Prometheus + Grafana monitoring stack with pre-configured dashboards
- Added PostgreSQL backup script to NAS (daily/weekly/monthly rotation)
- Added alert rules for backend, system, and Docker metrics
- Updated deployment guide for NPM + IONOS DNS homelab setup
- Added marketing plan document
- PDF translator and watermark support
- Enhanced middleware, routes, and translator modules
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Major changes across backend, frontend, infrastructure:
- Provider system with model selection (Google, DeepL, OpenAI, Ollama, Google Cloud)
- Admin panel: user management, pricing, settings
- Glossary system with CSV import/export
- Subscription and tier quota management
- Security hardening (rate limiting, API key auth, path traversal fixes)
- Docker compose for dev, prod, and IONOS deployment
- Alembic migrations for new tables
- Frontend: dashboard, pricing page, landing page, i18n (en/fr)
- Test suite and verification scripts
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
