feat: select Gemini model in settings (#180) by hessius · Pull Request #320 · hessius/MeticAI

hessius · 2026-03-25T18:52:48Z

Summary

Adds a model selector dropdown to Settings allowing users to choose which Gemini model to use for AI profile generation, shot analysis, and recommendations.

Changes

Backend (Python/FastAPI):

Added geminiModel to settings schema with default gemini-2.5-flash
GET /api/settings returns current model (env var takes precedence)
POST /api/settings persists model choice and hot-reloads GEMINI_MODEL env var
Startup hydration maps stored geminiModel → GEMINI_MODEL env var

Frontend (React/TypeScript):

Added GEMINI_MODEL key to STORAGE_KEYS constants
BrowserAIService: replaced hard-coded model with getGeminiModel() reading from localStorage
SettingsView: model selector dropdown (Gemini 2.5 Flash / 2.5 Pro / 2.0 Flash)
Works in both proxy mode (server settings) and direct mode (localStorage)

i18n: Translation keys added for all 6 locales (en, sv, de, es, fr, it)

Available Models

Model	Best For
Gemini 2.5 Flash (default)	Fast, balanced performance
Gemini 2.5 Pro	Higher quality, slower
Gemini 2.0 Flash	Legacy compatibility

Closes #180

Add model selector dropdown to Settings allowing users to choose between Gemini 2.5 Flash, Gemini 2.5 Pro, and Gemini 2.0 Flash for AI operations. Backend: geminiModel added to settings schema, GET/POST endpoints, and hydrated into GEMINI_MODEL env var with hot-reload support. Frontend: model selection persisted via localStorage (direct mode) or POST /api/settings (proxy mode). BrowserAIService reads model dynamically. i18n: translations added for all 6 locales (en, sv, de, es, fr, it). Closes #180 Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot

Pull request overview

Adds end-user configurability for which Gemini model MeticAI uses, wiring the selection through Settings (UI + persistence) and ensuring backend/runtime environment hydration supports it.

Changes:

Add geminiModel to persisted settings (default gemini-2.5-flash) and hydrate/hot-reload GEMINI_MODEL on the server.
Add a Gemini model selector to the Settings UI and persist the choice (server in proxy mode, localStorage in direct mode).
Add i18n keys across all locales for the new Settings text.

Reviewed changes

Copilot reviewed 12 out of 12 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
apps/web/src/services/ai/BrowserAIService.ts	Reads selected model from localStorage and uses it for Gemini calls in direct mode.
apps/web/src/lib/constants.ts	Adds `STORAGE_KEYS.GEMINI_MODEL` for persisted model selection.
apps/web/src/components/SettingsView.tsx	Adds Gemini model dropdown; loads/saves model in both proxy and direct modes.
apps/web/public/locales/en/translation.json	Adds `settings.geminiModel*` i18n keys (and related settings keys updates).
apps/web/public/locales/sv/translation.json	Adds `settings.geminiModel*` i18n keys (and related settings keys updates).
apps/web/public/locales/de/translation.json	Adds `settings.geminiModel*` i18n keys (and related settings keys updates).
apps/web/public/locales/es/translation.json	Adds `settings.geminiModel*` i18n keys (and related settings keys updates).
apps/web/public/locales/fr/translation.json	Adds `settings.geminiModel*` i18n keys (and related settings keys updates).
apps/web/public/locales/it/translation.json	Adds `settings.geminiModel*` i18n keys (and related settings keys updates).
apps/server/services/settings_service.py	Adds `geminiModel` default to settings schema.
apps/server/main.py	Hydrates `GEMINI_MODEL` from stored settings on startup.
apps/server/api/routes/system.py	Returns `geminiModel` (env precedence) and persists/hot-reloads it on save.

Copilot · 2026-03-25T21:19:37Z

apps/web/src/components/SettingsView.tsx

+                <option value="gemini-2.5-flash">Gemini 2.5 Flash</option>
+                <option value="gemini-2.5-pro">Gemini 2.5 Pro</option>
+                <option value="gemini-2.0-flash">Gemini 2.0 Flash</option>


The Gemini model dropdown option labels are hard-coded strings. Repo convention requires all user-facing strings to go through t() (see .github/CONVENTIONS.md i18n section). Add i18n keys for each option label (or otherwise route these labels through t()), even if the underlying value stays as the raw model id.

Suggested change

<option value="gemini-2.5-flash">Gemini 2.5 Flash</option>

<option value="gemini-2.5-pro">Gemini 2.5 Pro</option>

<option value="gemini-2.0-flash">Gemini 2.0 Flash</option>

<option value="gemini-2.5-flash">{t('settings.geminiModelOptionFlash25')}</option>

<option value="gemini-2.5-pro">{t('settings.geminiModelOptionPro25')}</option>

<option value="gemini-2.0-flash">{t('settings.geminiModelOptionFlash20')}</option>

✅ Resolved in 5e51ea9. Wrapped all three Gemini model dropdown labels in t() with keys settings.geminiModel25Flash, settings.geminiModel25Pro, and settings.geminiModel20Flash. Added corresponding translations to all 6 locale files (en, sv, de, es, fr, it).

Copilot · 2026-03-25T21:19:37Z

apps/server/api/routes/system.py

+            current_settings["geminiModel"] = str(body["geminiModel"]).strip()

        # For IP and API key changes, also update .env file
        env_updated = False


geminiModel from the request body is persisted as a raw stripped string, which allows an empty string to be written to settings.json. Treat blank/whitespace-only values as “unset” (e.g., delete the key or set the default) to avoid storing an invalid/ambiguous model value.

Suggested change

current_settings["geminiModel"] = str(body["geminiModel"]).strip()

# For IP and API key changes, also update .env file

env_updated = False

gemini_model = str(body["geminiModel"]).strip()

if gemini_model:

# Persist only non-empty model values

current_settings["geminiModel"] = gemini_model

else:

# Treat blank/whitespace-only input as "unset"

✅ Resolved in 5e51ea9. Empty/whitespace-only geminiModel now falls back to the default gemini-2.5-flash instead of persisting an empty string.

Copilot · 2026-03-25T21:19:37Z

apps/server/api/routes/system.py

+        # Hot-reload Gemini model into process environment
+        if "geminiModel" in body:
+            new_model = str(body["geminiModel"]).strip()
+            if new_model:
+                os.environ["GEMINI_MODEL"] = new_model
+                _update_s6_env("GEMINI_MODEL", new_model, request_id)
+                services_restarted.append("gemini_model")
+                logger.info("Updated GEMINI_MODEL to %s", new_model,
+                           extra={"request_id": request_id})


New Gemini model selection behavior in GET/POST /api/settings is not covered by existing endpoint tests (there are tests for /api/settings, but none assert geminiModel env precedence or persistence/hot-reload). Please add tests in apps/server/test_main.py for: GET returning env override; POST updating stored settings and os.environ['GEMINI_MODEL'] (and/or s6 env update call).

✅ Resolved in 5e51ea9. Added 3 tests: model persistence round-trip, empty-string fallback to default, and geminiModel presence in GET response. All 824 backend tests passing.

- i18n: wrap Gemini model dropdown labels in t() with keys for all 6 locales - validation: treat empty/whitespace geminiModel as default (gemini-2.5-flash) - tests: add 3 backend tests for geminiModel persistence and retrieval Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

hessius · 2026-03-25T22:11:24Z

Review Findings Addressed

✅ i18n for model labels — Wrapped all 3 dropdown options in t(), added keys to all 6 locale files
✅ Empty string validation — Blank/whitespace geminiModel now falls back to gemini-2.5-flash
✅ Backend tests added — 3 new tests: persistence round-trip, empty fallback, presence in GET

Issue #180 completeness: All acceptance criteria met — model selector visible in Settings, works in proxy and direct mode, selected model used for all AI operations, default model set. Ready to merge.

hessius added this to the 2.4 milestone Mar 25, 2026

hessius requested a review from Copilot March 25, 2026 21:15

Copilot started reviewing on behalf of hessius March 25, 2026 21:15 View session

Copilot AI reviewed Mar 25, 2026

View reviewed changes

hessius merged commit 98eba1e into version/2.4.0 Mar 26, 2026
6 checks passed

hessius deleted the feature/180-select-model branch March 26, 2026 05:55

This was referenced Mar 27, 2026

test: v2.4.0 pre-release manual testing #330

Open

v2.4 Release Coordination — Three-Pronged Deployment #317

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat: select Gemini model in settings (#180)#320

feat: select Gemini model in settings (#180)#320
hessius merged 2 commits intoversion/2.4.0from
feature/180-select-model

hessius commented Mar 25, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Mar 25, 2026

Uh oh!

hessius Mar 25, 2026

Uh oh!

Copilot AI Mar 25, 2026

Uh oh!

hessius Mar 25, 2026

Uh oh!

Copilot AI Mar 25, 2026

Uh oh!

hessius Mar 25, 2026

Uh oh!

hessius commented Mar 25, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

-            current_settings["geminiModel"] = str(body["geminiModel"]).strip()
-        # For IP and API key changes, also update .env file
-        env_updated = False
+            gemini_model = str(body["geminiModel"]).strip()
+            if gemini_model:
+                # Persist only non-empty model values
+                current_settings["geminiModel"] = gemini_model
+            else:
+                # Treat blank/whitespace-only input as "unset"

Conversation

hessius commented Mar 25, 2026

Summary

Changes

Available Models

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Mar 25, 2026

Choose a reason for hiding this comment

Uh oh!

hessius Mar 25, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 25, 2026

Choose a reason for hiding this comment

Uh oh!

hessius Mar 25, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 25, 2026

Choose a reason for hiding this comment

Uh oh!

hessius Mar 25, 2026

Choose a reason for hiding this comment

Uh oh!

hessius commented Mar 25, 2026

Review Findings Addressed

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants