InventoryCubit subscribed to watchInventory() and watchDrafts()
separately — two overlapping StreamGroups (both watching attachments)
re-emitted in a tight loop that starved the event loop and hung widget
tests' pumpAndSettle for 10+ minutes (home/inventory/quick-add all
timed out). Replace both with a single watchInventoryView() stream that
loads the catalogued list and the draft tray from one StreamGroup.
Full widget suite no longer hangs: 165 pass. One unrelated failure
remains in the v8 variety edit sheet (Save button pushed off-screen by
new fields).
Lower the bulk-digitization cliff with two more routes on top of the
already-landed CSV import and "save and add another":
- Photo-first drafts (capture now, catalogue later): burst-capture
photos (camera or multi-gallery) into unnamed draft varieties, shown
in a "to catalogue" tray, hidden from the main list until named.
Adds Variety.isDraft (schema), addDraftVariety/watchDrafts/nameDraft,
the triage sheet and the inventory banner.
- On-device OCR label suggestion (Tesseract, offline, no Google): a
"Suggest name from photo" button in the naming dialog behind a
LabelTextExtractor interface (Tesseract on Android/iOS, no-op
elsewhere). Reads the largest print via hOCR bounding boxes, drops
boilerplate/low-confidence noise, preprocesses (grayscale, contrast,
upscale) and sweeps rotations (0-315 deg) so tilted packets still
read. Bundles tessdata_fast eng+spa; validated on-device against real
packets. The photo is written to a temp file deleted immediately in a
finally block (the plugin needs a path) - a bounded, documented
exception to no-plaintext-at-rest.
This commit also carries the co-developed schema evolution v5 to v8 that
shares these files (organic flag, species viability years, crop
calendar, lot provenance/abundance/preservation format, condition
checks) plus their exports/migrations and i18n.
Tests: CSV/draft/OCR unit + widget + migration green in isolation.
Note: the full widget suite currently hangs (>10 min) - under investigation.
Import a spreadsheet list (the same 21-column schema the CSV export
already produces) as new varieties/lots. Tolerant reader: columns
matched by header name in any order, only variety_label required,
unknown enums fall back to a safe default, rows sharing variety_id
collapse into one variety with several lots. Species re-linked by
scientific name. Additive by design (fresh ids + HLC), so it never
overwrites existing rows; JSON stays the canonical id-reconciling import.
Adds a hand-rolled RFC 4180 parser (no new dependency), an importCsv
path in VarietyRepository/ExportImportService, an 'Import from CSV'
tile in the backup section, and en/es strings.
Note: i18n .g.dart also carries pre-existing About/intro strings from
the working tree (regenerated by slang).
Interchange export/import for Phase 1 (data-model §7):
- JSON: canonical, versioned envelope (formatVersion 1) with all sync
metadata verbatim, photos embedded as base64, tombstones excluded.
Species are re-resolved on import by scientific name (catalog ids are
per-install). Reader tolerates unknown fields/enum values (§5.2) and
rejects newer format versions with a clear error.
- CSV: export-only spreadsheet flatten, one row per lot, RFC 4180
escaping, never any photo bytes.
- Import merges by UUIDv7 id in one transaction: insert-if-unknown
preserving original stamps, last-writer-wins by packed HLC for known
mutable rows, append-only movements; afterwards the local clock
receiveEvent()s the newest imported stamp so it never runs behind.
- Settings gains a Backup section (export CSV/JSON, import JSON with
confirmation); file dialogs behind a FileService interface backed by
file_picker (MIT).
- Tests: codec round-trip and tolerance, reconciler LWW, repository
export→import round-trip (fresh DB, idempotent re-import, newer-local
wins, clock monotonicity, species re-resolution), backup widget flows.
Add filter chips above the inventory list, one per category in use and
one per lot form actually held; results apply search ∧ category ∧ form.
An active filter that empties the list shows a distinct 'no matches'
state and a clear-filters button. VarietyListItem now carries the set of
lot forms it holds (one aggregate query), and the stream re-emits on lot
changes.
Accessibility: category headers use the text theme so they honour the
system font scale; the list avatar is excluded from semantics (the tile
title already announces the name); the edit action uses the green action
colour (≥3:1 on the canvas, up from 2.62:1). Add category/form/clear
filter tests and a tap-target-size guideline test.
Expand LotType from {seed, plant} to six append-only forms (seed,
seedling, plant, tree, bulb, cutting) and add an optional Presentation
attribute (pot/tray/plug/bareRoot/rootBall) kept separate from quantity
so 'how many' and 'in what packaging' never conflate.
Add attachments.sortOrder so the cover photo is the lowest-ordered one,
enabling reordering from the full-screen viewer. Expand QuantityKind
with the new plant-aware forms.
Migrations are additive (v1→v5 all covered by migration_test); enum
values are stored by name and appended only, keeping CRDT sync safe.
Add Wikipedia/forum-style URLs to a variety:
- Repository: addExternalLink/removeExternalLink (soft delete); watchVariety
loads links and re-emits on external_links changes; ExternalLinkItem model
(title preferred over URL for display).
- Cubit: addExternalLink/removeExternalLink.
- Detail: a "Links" section with tappable link chips (open via url_launcher,
http(s):// prepended if missing), delete per chip, and an "Add link" dialog
(URL + optional title).
Tests: repo add/remove + widget add/remove. 52 green; Linux runs.
Photos:
- Repository addPhoto/removePhoto; VarietyDetail.photos (list); loads all photos.
- Detail: a 140×140 photo carousel with page dots, delete-current, and add-photo
(gallery); empty state shows an add-photo button.
- watchInventory now re-emits on attachment + species changes (StreamGroup), so
a photo added on the detail screen refreshes the inventory avatar immediately
(fixes avatars not updating).
Also carries the concurrent harvest-date work: lots gain an optional harvest
month (schemaVersion 3 + from2To3 migration, drift_schema_v3 exported), a
month/year harvest picker, and localized month names.
48 tests green (add/remove photos, avatar reactivity, v1→v3 migration); Linux runs.
Replace the free-text harvest year with a themed month/year picker:
a tappable field opens a green-styled dialog with month and year grids
(navigable, paginated years), month optional. Lots now read e.g.
"September 2021" or just "Year 2021".
- schema v3: add nullable Lot.harvest_month (1..12) + migration + tests
- repo/cubit propagate harvestMonth; localized month names (en/es) via slang
The "also known as" names were read-only; now they can be added and removed:
- Repository: addVernacularName + removeVernacularName (soft delete);
VarietyDetail.vernacularNames carries ids (new VernacularName model).
- Cubit: addVernacularName / removeVernacularName.
- Detail: deletable name chips + an "Add name" chip opening a small dialog;
the section is always visible so names can be added when empty.
Tests: repo add/remove + widget add/remove. 44 green.
Lots could only be created; now they can be edited and removed:
- Repository: updateLot (LWW) + softDeleteLot (tombstone).
- Cubit: updateLot / deleteLot.
- The add-lot sheet is now a shared add/edit sheet — tapping a lot opens it
prefilled (type, year, quantity), with a trash action to delete. The reactive
detail view refreshes on save/delete.
Tests: repo updateLot/softDeleteLot + widget edit-lot and delete-lot. 42 green.
Rework quantities the way seeds are actually described — a number of an informal
unit — and let a lot hold seeds or plants (plantel).
Quantity model (commons_core):
- Quantity is now { kind, count?, label } — "3 cobs", "2 packets", "a few".
Count is optional; uncountable vibes (a few, some, pinch) carry no number.
- QuantityKind reorganised into an ascending scale: vibes → containers
(teaspoon/spoon/cup/jar/sack) → seed & fruit forms → plant units, each tagged
countable/not. Stored by name; count reuses the existing numeric column.
Lot type (schema v2):
- Lots gain a `type` (seed | plant); germination stays meaningful for seeds.
schemaVersion 2 with a from1To2 migration (adds the column), drift_schema_v2
exported, migration test covers v1→v2.
UI:
- New QuantityPicker: a horizontal scale of large seed-glyph icons + a number
stepper (shown only for countable units), plus a seed/plant SegmentedButton.
Used in quick-add and add-lot. Lot lines render "3 cobs" / "a few".
- i18n: unit singular/plural (ES/EN, Weblate-friendly), lot-type labels.
Build: slang moved to its CLI (disabled in build_runner to avoid a multi-file
collision); CI runs `dart run slang` before build_runner. 38 tests green,
Linux migrates the existing v1 DB and runs.
Visual pass toward docs/mockups (06 inventory, 07 item):
- Bundle the seedks icon font (chars a–k → seed pictograms: jar, packet, spoon…)
and a SeedGlyph/QuantityKindIcon helper.
- Green app bar + pale-green canvas theme (buildTaneTheme).
- Inventory list: rounded-square photo thumbnail or green initial avatar, bold
title, scientific-name subtitle, trailing edit pencil, styled category
headers, rounded search field, seed-glyph empty state.
- Quantity selectors (quick-add + add-lot) show the seed pictogram + word;
detail lot lines use the kind's glyph.
- Detail header restyled to mockup 07: category link + scientific name on the
left, photo on the right.
watchInventory now also returns the linked species' scientific name (subtitle).
37 app tests green; seedks font verified in the Linux asset bundle.
The list showed only initials even when a variety had a photo (quick-add stores
one as an encrypted BLOB). Now the avatar shows the first photo, falling back to
the initial — matching the "foto/inicial" list spec.
- VarietyListItem carries the first photo bytes; watchInventory loads them in a
single query per emission.
- List tile renders a MemoryImage avatar when present.
Tests: repository (photo bytes surfaced, null when none) and a widget test
(avatar replaces the initial). Full suite: 37 passing, 0 skipped.
Record germination tests on a lot and surface the result.
- GerminationEntry model with a derived rate (germinated / sample). VarietyLot
carries its tests (most-recent first) and exposes latestGerminationRate.
- VarietyRepository.addGerminationTest; watchVariety now also re-emits on
germination-test changes, so the detail refreshes live.
- Detail UI: each lot shows a germination % badge and a grass action that opens
a sheet listing past tests and adding a new one (germinated + sample size).
i18n strings (ES/EN).
Tests: derived-rate + reactivity at the repository, and a widget test for the
record-test → badge flow. Full suite: 35 passing, 0 skipped.
Add a small curated catalog of Iberian horticultural species and let a variety
be linked to it from the edit sheet.
- assets/catalog/species.json: 14 species with botanical family and ES/EN
common names (wikidata_qid/gbif_key deferred to the varilla enrichment).
- SpeciesRepository: idempotent seedBundled (is_bundled rows, keyed by
scientific name) + search by scientific/common name with a locale-best label.
Seeded on startup from DI.
- VarietyRepository.linkSpecies: sets species_id and prefills category from the
species' family when empty (never overwrites an existing category).
VarietyDetail now carries the scientific name.
- Edit sheet gains a live species-search field; the detail view shows the
scientific name (italic). i18n strings added (ES/EN).
Tests: catalog parse, idempotent seeding, search by scientific/common name,
linkSpecies prefill semantics, and a widget test for the autocomplete → link →
scientific-name-shown flow. Full suite: 32 passing, 0 skipped.
Tapping an inventory item now opens its detail (the /variety/:id route was a
placeholder). Read + edit one variety end-to-end:
- Detail view: photo, category, "also known as" names, notes and lots, driven
by a reactive VarietyDetailCubit.
- Edit core fields (label/category/notes, LWW) via a bottom sheet.
- Add a lot (harvest year + quantity) — the list refreshes reactively.
- Soft-delete (tombstone) with a confirm dialog.
Repository:
- watchVariety(id): merges Drift table-change streams (variety, lots, names,
attachments) via StreamGroup and reloads the full detail on any change, so a
lot added to a *different* table still refreshes the view.
- updateVariety (LWW), addLot, softDeleteVariety.
i18n: detail/edit/addLot strings (ES/EN); switched slang to {param} braces
interpolation (translator- and Weblate-friendly).
Tests: repository (reactive watchVariety, update, soft-delete) and widget tests
(render, edit updates title, add-lot appears, delete shows not-found). Full
suite green: 23 passing, 1 skipped (SQLCipher-only encryption guard).
Stand up the Tanemaki monorepo and the first end-to-end vertical slice of
Block 1 (offline, encrypted inventory): add a seed → see it in a categorized,
searchable list → it persists → reopen and it's still there.
Architecture (state management like G1nkgo, adapted to Tane's reality):
- flutter_bloc (Cubit-first), but the encrypted Drift DB is the single source
of truth; cubits stream from repositories (no hydrated_bloc/Hive, which would
write plaintext at rest).
- get_it composition root; go_router; slang i18n (ES/EN, Weblate-friendly JSON).
Workspace & core:
- pub workspace: packages/commons_core (pure Dart) + apps/app_seeds (Flutter).
- commons_core primitives: UUIDv7 IdGen, Hybrid Logical Clock, Quantity value
type (+ plant-aware QuantityKind), IdentityService root-seed stub.
Data & security:
- Drift schemaVersion=1 with all 10 Block-1 tables + common CRDT columns
(HLC updated_at, last_author, tombstones); Movement append-only.
- SQLCipher via an injectable executor that refuses to open a plaintext DB;
DB key + root seed in the OS keystore (separate secrets).
- Exported drift_schema_v1.json + migration scaffold.
Tests (near-TDD; nothing merges without tests):
- commons_core units (24), Drift migration test, "no plaintext at rest"
security guard (runs where SQLCipher is present, skips otherwise),
repository, widget, full quick-add flow, file-reopen persistence, and a
no-hardcoded-strings i18n guard. GitLab CI: format + analyze + test + coverage.
Follow-on Block-1 stories (item detail/edit, species catalog, germination UI)
and all of Block 2 remain out of scope.
