_user_constraints() loads dietary_constraints from user_settings once per
request. All 7 _enrich_item call sites now pass constraints so wine (and
any future alcohol-containing entries) are suppressed for halal/alcohol-free
users at the API response layer.
Adds 10 new secondary use entries and corrects all 8 existing ones.
New: apples/soft, leafy_greens/wilting, tomatoes/soft, cooked_pasta/day-old,
cooked_potatoes/day-old, yogurt/tangy, cream/sour, wine/open,
cooked_beans/day-old, cooked_meat/leftover.
Corrections: milk uses (specific recipes, not 'baking'/'sauces'); dairy uses
expanded; cheese label well-aged→rind-ready with named dishes (minestrone,
ribollita); rice uses (onigiri, arancini, congee); tortillas warning added;
bakery uses and synonyms expanded to named pastries; bananas synonyms
(spotty/brown/black/mushy); rice synonyms (old rice).
New fields on every SECONDARY_WINDOW entry:
- discard_signs: qualitative cues for when the item has gone past its
secondary window (shown in UI alongside uses)
- constraints_exclude: dietary labels that suppress the entry entirely
(wine suppressed for halal/alcohol-free)
ExpirationPredictor.filter_secondary_by_constraints() applies constraint
suppression; _enrich_item() now accepts user_constraints and passes
secondary_discard_signs through to the API response.
Integrates cf-core reranker into the L1/L2 recipe engine. Paid+ tier
gets a BGE cross-encoder pass over the top-20 FTS candidates, scoring
each recipe against the user's full context: pantry state, dietary
constraints, allergies, expiry urgency, style preference, and effort
preference. Free tier keeps the existing overlap sort unchanged.
- New app/services/recipe/reranker.py: build_query, build_candidate_string,
rerank_suggestions with tier gate (_RERANKER_TIERS) and graceful fallback
- rerank_score field added to RecipeSuggestion (None on free tier, float on paid+)
- recipe_engine.py: single call after candidate assembly, before final sort;
hard_day_mode tier grouping preserved as primary sort when reranker active
- Fix pre-existing circular import in app/services/__init__.py (eager import
of ReceiptService triggered store.py → services → receipt_service → store)
- 27 unit tests (mock backend, no model weights) + 2 engine-level tier tests;
325 tests passing, no regressions
- Add max_total_min to RecipeRequest schema and TypeScript interface
- Add _within_time() helper to recipe_engine using parse_time_effort()
with graceful degradation (empty directions or no signals -> pass)
- Wire max_total_min filter into suggest() loop after max_time_min
- Add time_first_layout to allowed settings keys
- Add timeFirstLayout ref to settings store (preserves sensoryPreferences)
- Add maxTotalMin ref to recipes store, wired into _buildRequest()
- Add time bucket selector UI (15/30/45/60/90 min) in RecipesView
Find tab, gated by timeFirstLayout != 'normal'
- Add time-first layout selector section in SettingsView
- Add 5 _within_time unit tests and 2 settings key tests
- Cook/Exit toggle button in recipe detail header (hidden for recipes with no steps)
- Cook mode progress bar between header and body showing step N of M
- Single-step view replaces recipe body; shows Active/Wait badge and passive hint
from #50 time_effort data (null-safe — degrades gracefully without it)
- Prev/Next nav buttons; Next becomes green Done on last step
- ArrowLeft/ArrowRight keyboard navigation (preventDefault to suppress scroll)
- Touch swipe left/right (40px horizontal threshold, 80px vertical abort)
- Done triggers handleCook() then exitCookMode() so success banner appears instantly
- Add app/services/recipe/time_effort.py: parse_time_effort(), TimeEffortProfile,
StepAnalysis dataclasses; two-branch regex for time ranges and single values;
whole-word passive keyword detection; 480 min/step cap; 1825 day global cap
- Add directions to browse_recipes and _browse_by_match SELECT queries in store.py
- Enrich browse and detail endpoints with active_min/passive_min/time_effort fields
- Add StepAnalysis, TimeEffortProfile TS interfaces to api.ts
- RecipeBrowserPanel: split pill badge showing active/passive time
- RecipeDetailPanel: collapsible ingredients summary, effort cards (Active/Hands-off/Total),
equipment chips, annotated step list with Active/Wait badges and passive hints
- 45 new tests (40 unit + 5 API); 215 total passing
Zero-count subcategory buttons show a ＋ badge. Clicking opens a modal:
- Recipe search (debounced, 3-char min) using existing browse API
- Pre-filled domain/category/subcategory from current browse context,
fully correctable via selects populated from loaded domains/categories
- Submit calls POST /recipes/community-tags; 409 on duplicate
- Success message: 'It will appear once a second user confirms'
api.ts: adds submitRecipeTag(), upvoteRecipeTag(), listRecipeTags() to browserAPI.
CSS: tag-cta pill on subcat buttons, modal-backdrop + modal-box with theme vars.
TODO: wire real community pseudonym (currently hardcoded 'anon').
Refs kiwi#118.
browse_counts_cache.py: after FTS counts, _merge_community_tag_counts() queries
accepted tags (upvotes>=2) grouped by (domain,category,subcategory) and adds
distinct recipe_id counts to the cached keyword-set totals. Skips silently
when community Postgres is unavailable.
store.py: fetch_recipes_by_ids() fetches corpus recipes by explicit ID list,
used by the FTS fallback when a subcategory returns zero FTS results.
recipes.py (browse endpoint): when FTS total==0 for a subcategory, queries
community store for accepted tag IDs and serves those recipes directly.
Sets community_tagged=True in the response so the UI can surface context.
Refs kiwi#118.
GET /recipes/community-tags/{recipe_id} — all tags for a recipe
POST /recipes/community-tags — submit tag (requires pseudonym)
POST /recipes/community-tags/{id}/upvote — vote on a tag
Validates (domain, category, subcategory) against DOMAINS taxonomy before
accepting. Returns 409 on duplicate submission or double-vote. Fails soft
(503) when community Postgres is unavailable so the browse path is unaffected.
Refs kiwi#118.
food.com rarely tags BBQ in its own taxonomy fields, so BBQ recipes were
previously untagged. Added content-derived signals (brisket, pulled pork,
dry rub, regional styles) so infer_recipe_tags.py correctly tags them
as cuisine:BBQ. Companion to the browser_domains.py BBQ keyword expansion.
Switches recipe generation service type from 'cf-text' to 'vllm' so the
coordinator can route to quantized small models (Qwen2.5-3B, Phi-4-mini)
rather than the full text backend. Passes CF_APP_NAME for per-product
VRAM/request analytics in the coordinator dashboard.
- llm_recipe.py: _SERVICE_TYPE = 'vllm'; _MODEL_CANDIDATES list; passes
model_candidates and pipeline= to CFOrchClient.allocate()
- compose.cloud.yml: CF_APP_NAME=kiwi env var for coordinator attribution
Shopping links previously hardcoded to US storefronts. Users in other regions
got broken Amazon Fresh and Instacart links. Now locale is stored as a user
setting and passed to GroceryLinkBuilder at request time.
- locale_config.py: per-locale Amazon domain/dept config (already existed)
- grocery_links.py: GroceryLinkBuilder accepts locale=; routes Instacart to .ca
for Canada, uses amazon_domain per locale, Instacart/Walmart US/CA only
- settings.py: adds 'shopping_locale' to allowed settings keys
- shopping.py: reads locale from user's stored setting on all list/add/update paths
- SettingsView.vue: Shopping Region selector (NA, Europe, APAC, LATAM)
- stores/settings.ts: shoppingLocale reactive state, saves via settings API
Users often have ingredients they want to avoid today (out of stock, not feeling it)
that aren't true allergies. The new 'Not today' filter lets them exclude specific
ingredients per session without permanently modifying their allergy list.
- recipe.py schema: exclude_ingredients field (list[str], default [])
- recipe_engine.py: filters corpus results when any ingredient is in exclude_set
- llm_recipe.py: injects exclusions into both prompt templates so LLM-generated
recipes respect the constraint at generation time
- RecipesView.vue: tag-chip UI with Enter/comma input, removes on × click
- stores/recipes.ts: excludeIngredients reactive list (not persisted to localStorage)
Adds 'Best match' sort button to the recipe browser. When selected, recipes are
ordered by the fraction of their ingredients that are in the user's pantry.
- store.py: _browse_by_match() pushes match_pct computation into SQL via json_each()
so ORDER BY can sort the full result set before LIMIT/OFFSET pagination
- recipes.py: extends sort pattern validation to accept 'match'; falls back to
default when no pantry_items provided
- RecipeBrowserPanel.vue: adds 'Best match' button (disabled when pantry empty);
watcher auto-engages match sort when pantry goes from empty to non-empty
Multiple concurrent users browsing the 3.2M recipe corpus would cause FTS5 page
cache contention and slow per-request queries. Solution: pre-compute counts for
all category/subcategory keyword sets into a small SQLite cache.
- browse_counts_cache.py: refresh(), load_into_memory(), is_stale() helpers
- config.py: BROWSE_COUNTS_PATH setting (default DATA_DIR/browse_counts.db)
- main.py: warms in-memory cache on startup; runs nightly refresh task every 24h
- infer_recipe_tags.py: auto-refreshes cache after a successful tag run so the
app picks up updated FTS counts without a restart
Top-level category keywords were too narrow, missing common food.com corpus terms
like 'barbecue', 'smoky', 'charcoal'. Subcategory terms also expanded to cover
broader corpus vocabulary so FTS counts register hits across more recipes.
Documents Ollama/vLLM local backends, cf-text via cf-orch allocation,
and BYOK cloud options (Anthropic, OpenAI). cf_text leads fallback_order
for meal planning and expiry prediction paths.
Paid+ local users with circuitforge_orch installed now get the coordinator-
aware scheduler automatically — no env var needed. The coordinator's
allocation queue already prefers the local GPU first, so latency stays low.
Priority: USE_ORCH_SCHEDULER env override > CLOUD_MODE > cf-orch importable.
Free-tier local users without cf-orch installed get LocalScheduler as before.
USE_ORCH_SCHEDULER=false can force LocalScheduler even when cf-orch is present.
Switches to OrchestratedScheduler in cloud mode so concurrent recipe_llm
jobs fan out across all registered cf-orch GPU nodes instead of serializing
on one. Under load this eliminates poll timeouts from queue backup.
USE_ORCH_SCHEDULER env var gives explicit control independent of CLOUD_MODE:
unset follow CLOUD_MODE (cloud=orch, local=local)
true OrchestratedScheduler always (e.g. multi-GPU local rig)
false LocalScheduler always (e.g. cloud single-GPU dev instance)
ImportError fallback: if circuitforge_orch is not installed and orch is
requested, logs a warning and falls back to LocalScheduler gracefully.
Frontend now uses the async job queue for level 3/4 requests instead
of a 120s blocking POST. Submits with ?async=true, gets job_id, then
polls every 2.5s up to 90s. Button label reflects live server state:
'Queued...' while waiting, 'Generating...' while the model runs.
- api.ts: RecipeJobStatus interface + suggestAsync/pollJob methods
- store: jobStatus ref (null|queued|running|done|failed); suggest()
branches on level >= 3 to _suggestAsync(); CLOUD_MODE sync fallback
detected via 'suggestions' key on the response
- RecipesView: button spinner text uses jobStatus; aria-live
announcements updated for each phase (queued/running/finding)
- compose.override.yml: cf-orch agent --node-id renamed kiwi -> sif
for the upcoming Sif hardware node
Adds the recipe_jobs table and background task pipeline for level 3/4
recipe generation. POST ?async=true returns 202 with job_id; clients
poll GET /recipes/jobs/{job_id} until status=done.
Key fix: _enqueue_recipe_job now calls scheduler.enqueue() after
insert_task() to wake the in-memory work queue immediately. Without
this, tasks sat in 'queued' until the scheduler's 30s idle cycle or
an API restart triggered _load_queued_tasks().
- Migration 034: recipe_jobs table (job_id, user_id, status, request,
result, error) with indexes on job_id and user_id/created_at
- Store: create/get/update_running/complete/fail recipe job methods
- runner.py: recipe_llm task type + _run_recipe_llm handler; MUST
call fail_recipe_job() before re-raising so status stays consistent
- CLOUD_MODE guard: falls back to sync generation (scheduler only
polls shared settings DB, not per-user DBs)
- L4 wildcard is covered by the same req.level in (3, 4) dispatch
When the coordinator returns 429 (all nodes at max_concurrent limit), the previous
code fell back to LLMRouter which is also overloaded at high concurrency. This
caused the request to hang for ~60s before nginx returned a 504.
Now: detect 429/max_concurrent in the RuntimeError message and return "" immediately
so the caller gets an empty RecipeResult (graceful degradation) rather than a timeout.
cf-orch-agent in compose.override.yml was crash-looping (exit 127) because
the circuitforge_orch package wasn't installed in the kiwi conda env.
Same COPY + editable-install pattern already used for circuitforge-core.
Adds minimal sort/search to the recipe browser for cognitive access diversity —
linear scanners, alphabet browsers, and keyword diggers each get a different
way in without duplicating the full search tab.
- browse_recipes: q (LIKE title filter) + sort (default/alpha/alpha_desc)
- API endpoint: q/sort query params with validation
- Frontend: debounced search input (350ms) + sort pills (Default/A→Z/Z→A)
- Search and sort reset on domain/category change
- _all path supports q+sort; keyword-FTS path adds AND filter on top
Adds a two-level browse tree (domain → category → subcategory) to the
recipe browser, plus an "All" unfiltered option at the top of every
domain.
browser_domains.py:
- Category values now support list[str] (flat) or dict with "keywords"
and "subcategories" keys — backward compatible with all existing flat
categories
- Added subcategories to: Italian (Sicilian, Neapolitan, Tuscan, Roman,
Venetian, Ligurian), Mexican (Oaxacan, Yucatecan, Veracruz, Street
Food, Mole), Asian (Korean, Japanese, Chinese, Thai, Vietnamese,
Filipino, Indonesian), Indian (North, South, Bengali, Gujarati),
Mediterranean (Greek, Turkish, Moroccan, Lebanese, Israeli), American
(Southern, Cajun/Creole, BBQ, Tex-Mex, New England), European
(French, Spanish, German, British/Irish, Scandinavian), Latin American
(Peruvian, Brazilian, Colombian, Cuban, Caribbean), Dinner, Lunch,
Breakfast, Snack, Dessert, Chicken, Beef, Pork, Fish, Vegetables
- New helpers: category_has_subcategories, get_subcategory_names,
get_keywords_for_subcategory
store.py:
- get_browser_categories now accepts has_subcategories_by_category and
includes has_subcategories: bool in each result row
- New get_browser_subcategories method for subcategory count queries
recipes.py endpoints:
- GET /browse/{domain}/{category}/subcategories — returns subcategory
list with recipe counts (registered before /{subcategory} to avoid
path collision)
- GET /browse/{domain}/{category} gains optional ?subcategory=X param
to narrow results within a category
- GET /browse/{domain}/{category}/_all — unfiltered paginated browse
(landed in previous commit)
api.ts: BrowserCategory adds has_subcategories; new BrowserSubcategory
type; listSubcategories() call; browse() gains subcategory param
RecipeBrowserPanel.vue:
- Category pills show a › indicator when subcategories exist
- Selecting such a category fetches subcategories in the background
(non-blocking — recipes load immediately at the category level)
- Subcategory row appears below the category list with an
"All [Category]" pill + one pill per subcategory with count
- Active subcategory highlighted; clicking "All [Category]" resets
to the full category view
Secondary-state items (stale bread, overripe bananas, day-old rice, etc.)
are now surfaced to the recipe engine so relevant recipes get matched even
when the ingredient is phrased differently in the corpus (e.g. "day-old
rice" vs. "rice").
Backend:
- Add rice and tortillas entries to SECONDARY_WINDOW in expiration_predictor
- Add secondary_pantry_items: dict[str, str] field to RecipeRequest schema
(maps product_name → secondary_state label, e.g. {"Bread": "stale"})
- Add _SECONDARY_STATE_SYNONYMS lookup in recipe_engine — keyed by
(category, state_label), returns corpus-matching ingredient phrases
- Update _expand_pantry_set() to accept secondary_pantry_items and inject
synonym terms into the expanded pantry set used for FTS matching
Frontend:
- Add secondary_pantry_items to RecipeRequest interface in api.ts
- Add secondaryPantryItems param to _buildRequest / suggest / loadMore
in the recipes store
- Add secondaryPantryItems computed to RecipesView — reads secondary_state
from inventory items (expired but still in secondary window) and builds
the product_name → state_label map
- Pass secondaryPantryItems into handleSuggest and handleLoadMore
Closes#83
Add E2E_TEST_USER_ID setting (opt-in via env); session bootstrap logs
at DEBUG instead of INFO for the known test user so test runs don't
inflate session counts. Still visible with DEBUG=true.
- App.vue: lazy-mount pattern (v-if + v-show) so non-active tabs only mount on
first visit, eliminating concurrent onMounted calls across all components (#98)
- nginx.cloud.conf: add /kiwi/api/ location to proxy API calls on direct-port
access (localhost:8515); was serving SPA HTML → causing M.map/filter/find
TypeError chain on load (#98)
- nginx.cloud.conf: $host → $http_host so 307 redirects preserve port number (#107)
- RecipeBrowserPanel: show friendly "corpus not loaded" notice and skip auto-select
when all category counts are 0, instead of rendering confusing empty buttons (#106)
- Defensive Array.isArray guards in inventory store, mealPlan store, ReceiptsView
Cloud mode: attach shared read-only corpus DB (RECIPE_DB_PATH env var)
as "corpus" schema so per-user SQLite DBs can access 3.19M recipes.
All corpus table references now use self._cp prefix ("corpus." in cloud,
"" in local). FTS5 pseudo-column kept unqualified per SQLite spec.
compose.cloud.yml: bind-mount /Library/Assets/kiwi/kiwi.db read-only.
Also fix batch of audit issues:
- #101: OCR approval used source="receipt_ocr" for inventory_items — use "receipt"
- #89/#100: Shopping confirm-purchase used source="shopping_list" — use "manual"
- #103: Frontend inventory filter sent ?status= but API expects ?item_status=
- #104: InventoryItemUpdate schema missing purchase_date field; store.py allowed set also missing it
- #105: Guest cookie Secure flag tied to CLOUD_MODE instead of X-Forwarded-Proto; broke HTTP direct-port access
- MealPlanGrid: edit-mode toggle (visible when >1 meal type) reveals
per-row ↑/↓ reorder and ✕ remove controls; label column expands to
auto width via CSS class swap
- mealPlan store: removeMealType() and reorderMealTypes() — both send
the full updated array to the existing PATCH /meal-plans/{id} endpoint
- MealPlanView: wires remove-meal-type and reorder-meal-types emits;
shares mealTypeAdding loading flag to disable controls during save
- Guard: remove disabled when only one type remains (always keep ≥1)
Slot click now opens an inline editor panel:
- Pick from saved recipes via dropdown (pre-loaded on mount)
- Or type a custom label
- Clear slot button when a slot is already filled
- Save/Cancel with loading state
Add meal type opens a chip picker showing the types not yet active
(breakfast / lunch / snack minus whatever is already on the plan).
Selecting one calls the new PATCH /meal-plans/{plan_id} endpoint.
Backend:
- PATCH /meal-plans/{plan_id} with UpdatePlanRequest(meal_types)
- store.update_meal_plan_types() UPDATE ... RETURNING *
- 409 on IntegrityError in create_plan (already in place)
- Add autoSelectPlan() to the store: after loadPlans() resolves, set
activePlan to the current week's plan (or most recent) without a second
API round-trip -- list already returns full PlanSummary with slots
- Call autoSelectPlan(mondayOfCurrentWeek()) in onMounted so the grid
populates immediately without the user touching the dropdown
- Make onNewPlan idempotent: if a 409 comes back, activate the existing
plan for that week instead of surfacing an error to the user
- Fix sqlite3.OperationalError: the recipes table uses `title` not `name`;
get_plan_slots JOIN was crashing every list_plans call with a 500,
making the + New week button appear broken (plans were being created
silently but the selector refresh always failed)
- Add migration 032 to add UNIQUE INDEX on meal_plans(week_start)
to prevent duplicate plans accumulating while the button was broken
- Raise HTTP 409 on IntegrityError in create_plan so duplicates produce
a clear error instead of a 500
- Fix mondayOfCurrentWeek to build the date string from local date parts
instead of toISOString(), which converts through UTC and can produce the
wrong calendar day for UTC+ timezones
- Add planCreating/planError state to MealPlanView so button shows
"Creating..." during the request and displays errors inline
- Refactor _lookup_in_database to accept a shared httpx.AsyncClient so
all three Open*Facts database attempts reuse one TLS connection instead
of opening a new one per call; restores pre-fallback scan speed
- Increase recipe suggest timeout to 120s (was 30s) to survive cf-orch
model cold-start on first request of a session
- Include product brand in barcode scan success message so the user can
clearly see what was found (e.g. "Added: Cheerios (General Mills) to pantry")
#55 — Complexity rating on recipe cards:
- Derived from direction text via _classify_method_complexity()
- Badge displayed on every card: easy (green), moderate (amber), involved (red)
- Filterable via complexity filter chips in the results bar
#58 — Cooking time + difficulty as filter domains:
- estimated_time_min derived from step count + complexity
- Time hint (~Nm) shown on every card
- complexity_filter and max_time_min fields in RecipeRequest
- Both applied in the engine before suggestions are built
#53 — Surprise Me: picks a random suggestion from the filtered pool,
avoids repeating the last pick. Shown in a spotlight card.
#57 — Just Pick One: surfaces the top-matched suggestion in the same
spotlight card. One tap to commit to cooking it.
Closes#55, #58, #53, #57
Adds GET /export/json that bundles inventory and saved recipes into a
single timestamped JSON file for data portability. The export envelope
includes schema version and export timestamp so future import logic can
handle version differences.
Frontend: new primary-styled JSON download button in the Export card with
a short description of what is included.
Closes#62
