Three new fields: cost (Decimal, not Float -- exact money math),
costCurrency (free text), externalPoNumber (free text PO number from an
external system, distinct from purchaseOrderId which is a real FK into
this apps own PurchaseOrder table -- external exports PO numbers are
messy free text with nothing to link to here).
importAssetItems() now upserts: a row matching an existing assetTag or
serialNumber updates that row instead of erroring on the uniqueness
conflict. Needed to backfill cost/PO onto the 613 already-imported rows
without recreating them, and sets this up for future re-syncs when the
company re-exports their ITAM report. Update only touches fields the row
actually carries a value for (blank cell = no change, not clear) --
warrantyEndDate is the one field NOT auto-recomputed from warrantyMonths
on update (unlike create), so a manually-corrected date survives a
re-sync. Placeholder junk (Not available/Pending/N/A/etc, case
-insensitive) in PO number/cost/currency cells is treated as absent.
Re-ran the full 613-row import after this: 0 created, 613 updated, 0
errors -- backfilled cost onto the 8 rows that had one and PO number onto
the 2 that had a real (non-placeholder) value, verified against the
source file.
Hand-written migration, no Vitest coverage -- same deviation as the rest
of this branch, see CLAUDE.md. tsc/lint clean.
Prep work for importing a real company ITAM export (613 real assets):
- AssetItem.assetTag (nullable, unique) -- company physical tag, was only
on InventoryScan (move feature) before, now on the actual asset record.
- AssetItem.purchaseDate is now nullable (was required) -- some legacy
rows genuinely have no known purchase date. Manual /assets create form
still requires it in the HTML input; this only relaxes the DB/import
path for historical data.
- importAssetItems() accepts a status column now (free-text "In use"/
"In stock"/etc mapped via IMPORT_STATUS_MAP), unlike manual creation
which always defaults in_stock -- bulk-imported inventory is usually
already deployed, not freshly purchased, so forcing in_stock would
misrepresent reality.
Hand-written migrations (asset_tag, purchase_date nullable), no Vitest
coverage -- same deviation as the rest of this branch, see CLAUDE.md.
tsc/lint clean.
Actually imported 613/613 real AssetItem rows from the export with 0
errors after this (not committed here -- thats data, not code; see
CLAUDE.md for the seeding/import process and whats deliberately out of
scope this round: Cost, PO number, employee Assignment).
/assets gets Export CSV (GET /api/assets/export) and Import CSV
(POST /api/assets/import) buttons. Export emits human-readable
manufacturer/model columns instead of raw modelId; import resolves those
back to an AssetModel (case-insensitive match, must already exist -- this
does not auto-create models) so export -> edit in Excel -> re-import is a
real round trip.
Per-row import errors (unmatched model, missing required field, bad date)
are collected and returned alongside the created count rather than
aborting the whole file on the first bad row -- a typo in row 200 should
not cost rows 1-199. warrantyEndDate still auto-fills from the models
warrantyMonths when the CSV omits it, same as manual creation.
Extracted src/lib/csv.ts (splitCsvLine/splitCsvLines/normalizeHeader/
csvField) out of expected-assets.ts so both CSV parsers share one
implementation instead of duplicating it.
Manual smoke test only (tsx script: export shape, valid row + unmatched-
model row + missing-date row all handled correctly), no Vitest coverage,
not through the ba->pm->feature-implementer->test-runner->code-reviewer
pipeline -- same deviation as the rest of this branch, see CLAUDE.md.
tsc/lint clean.
AssetModel gains an optional warrantyMonths (standard vendor warranty
length). createAssetItem() now defaults warrantyEndDate to purchaseDate +
warrantyMonths (month-end clamped, e.g. Dec 31 + 2mo -> Feb 28, not Mar 3)
whenever the caller omits warrantyEndDate -- still overridable per item
for batches with extended/different warranty terms. /asset-models UI gets
an input for the new field; /assets needs no change since it already
omits warrantyEndDate from the payload when left blank.
Hand-written migration + manual smoke test (tsx script, not Vitest),
same deviation as the other 2026-08-2x work in this branch -- see
CLAUDE.md. tsc/lint clean.
Lets the operator upload a CSV of already-known assets (asset tag, model,
serial, old area/assignee) before an office-move walk, via a new
session-authed /inventory-scans/import page. The Telegram bot gains a
"tim trong list" search step at the asset-tag prompt: picking a match
pre-fills tag/serial and offers one-tap "giu nguyen" buttons for the
old area/assignee, then marks that row matched once the scan is created.
Follows the same-day decision to turn off local OCR (qwen2.5vl via Ollama)
for this feature -- see CLAUDE.md and the ITAM/Ollama vault notes for the
accuracy/speed numbers behind that call.
Deviates from repo convention same as the InventoryScan work above: no
Vitest coverage (manual smoke test only, via simulated webhook calls),
hand-written migration, not through the ba->pm->feature-implementer->
test-runner->code-reviewer pipeline. tsc/lint are clean.
List + inline edit (model within category, area, assignee, specs, asset tag, serial number) + delete. No create form -- creation only happens via the bot.
Co-Authored-By: Claude Sonnet 5 <[email protected]>
GET /api/areas, GET /api/inventory-scans, PATCH+DELETE /api/inventory-scans/[id], GET /api/inventory-scans/[id]/photo (proxies Telegram file_id on demand, derives Content-Type from file_path extension since Telegram serves application/octet-stream regardless of actual type).
Co-Authored-By: Claude Sonnet 5 <[email protected]>
Auth via BOT_INGEST_SECRET bearer header (src/lib/bot-auth.ts), not Auth.js session -- caller is the inventory-capture Telegram bot service, not a browser.
Co-Authored-By: Claude Sonnet 5 <[email protected]>
Per-row employee <Select> + Transfer button alongside the existing Return
button, shown for active (returned_at null) rows. PATCHes
/api/assignments/[id] with {employeeId}, which the route already dispatches
to transferAssignment() vs returnAssignment() based on body shape.
Mirrors assignments/page.tsx's per-row close mini-form pattern; targetStatus
select reuses the closed AssetItemStatus set client-side, same precedent as
licenses page's licenseType select.
createMaintenanceRecord()'s AssetItem status guard pinned status:{not:"retired"}
instead of status:"in_stock", so two concurrent opens (or an open racing a
concurrent createAssignment()) could both pass the guard, producing two open
MaintenanceRecords or an orphaned active Assignment. Guard now pins "in_stock"
exactly, mirroring createAssignment()'s own atomic guard.
closeMaintenanceRecord()'s targetStatus accepted any of the 4 AssetItemStatus
values, including "assigned" — which sets AssetItem.status without creating a
corresponding Assignment row, desyncing Assignment (source of truth for who
holds what) from AssetItem.status with no way back. Restricted to
in_stock|retired, the only two outcomes a closed repair can actually resolve
to.
createMaintenanceRecord() opens a record and flips AssetItem.status to
"repair" inside one transaction, rejecting on missing/retired AssetItem,
an active Assignment, or an already-open record for the same item.
closeMaintenanceRecord() sets endDate/resolutionNotes and applies a
caller-supplied targetStatus (never defaulted) to the AssetItem, guarding
double-close and endDate>=startDate. Mirrors src/lib/assignments.ts's
Zod-validate + $transaction + WHERE-guarded updateMany race-closing pattern.
The only way to move an Assignment to a new employee was two separate calls
(PATCH return + POST create), which is not atomic and left a race window
where a third party could grab the now-in_stock AssetItem in between, plus
no rollback if the second call failed. Adds transferAssignment() which
closes the old Assignment and creates the new one inside a single
prisma.$transaction(), per docs/asset-license-management.md's "Assignment
(chốt 2026-08-02)" spec. Does not touch AssetItem.status or
License.seatsUsed — the resource never becomes free mid-transfer. Wired
into PATCH /api/assignments/[id]: an employeeId key in the body dispatches
to transfer, a returnedAt key to return, both together is rejected as
mutually exclusive.
POST/GET /api/assignments mirror the existing entity route pattern
(auth-gated, thin handler delegating to src/lib/assignments.ts). PATCH
/api/assignments/[id] is the first dynamic-segment route in the codebase,
per the stakeholder-confirmed convention (docs/asset-license-management.md)
of update routes being PATCH /api/<entity>/[id] rather than an
action-based sub-route.
createAssignment() enforces the XOR (assetItemId/licenseId) at the Zod
layer and the AssetItem.status/License.seatsUsed business rules inside a
single prisma.$transaction(), using atomic WHERE-guarded updateMany() calls
(not plain read-then-write) to close the seat-overbooking and
double-assignment race conditions flagged for License by code-reviewer.
returnAssignment() reverses those side-effects the same way and closes the
double-return race with the same pattern.
AssetItem.model was a free-text string with no referential integrity.
Now a required FK to AssetModel, mirroring the Request.employeeId and
License.vendorId retrofit precedent (no backfill needed, no real
AssetItem data existed).
Zod schema + validate function, Prisma-backed CRUD following the
{ok,value}|{ok,error} pattern from the start. createAssetCategoryAttributeDef()
catches P2002 for the composite (categoryId, key) unique constraint (mirrors
createAssetItem()'s serialNumber handling) and P2003 for the categoryId FK
(mirrors createPurchaseOrder()).
Nullable FK from License to PurchaseOrder for supply-side traceability
(which PO a license came in through), mirroring AssetItem.purchaseOrderId's
just-merged pattern exactly. Separate commit from the earlier
License.vendorId retrofit per the project's 1-commit-per-business-change
convention.
License.vendor (free-text string) is now vendorId, a required FK to
Vendor, per docs/asset-license-management.md's "## Vendor (chốt
2026-08-02)" spec. Mirrors the Request.employeeId string->FK precedent:
Zod validates UUID format only, Prisma's own FK constraint (P2003) is
caught at insert time and translated into a clean 400 instead of a raw
500. No License data existed yet, so this is a pure column-type
migration, no backfill needed.
Also converts createLicense() to the {ok,value}|{ok,error} result
pattern used elsewhere (createRequest, createAssetItem) instead of
returning a bare LicenseRecord promise - closes the gap flagged in
CLAUDE.md now that License has its first FK-violation error case to
handle.
- createRequestSchema now requires employee_id to be a UUID-format
string (was a free-text non-empty check), matching the new FK to
Employee in schema.prisma.
- createRequest() now returns a CreateRequestResult discriminated
union and catches Prisma P2003 (FK violation), returning a friendly
ok:false error instead of a raw 500 when employee_id does not
reference an existing Employee. POST /api/requests branches on
created.ok and returns 400 in that case.
- tests/requests.test.ts: every test that used to POST a fake string
employee_id (emp-1, emp-2) now creates a real Employee first via
createTestEmployee() and uses its id, since employee_id is now a
real FK. Added coverage for a syntactically invalid UUID and for a
well-formed but non-existent UUID (expects the friendly 400 from
the P2003 catch, not a raw 500) - the exact case this retrofit was
meant to protect against.
POST/GET /api/employees, same thin-route pattern as
src/app/api/requests/route.ts including the auth() 401 check
(middleware.ts excludes /api from page-route gating). POST also
surfaces createEmployee()'s FK-violation translation (nonexistent
managerId) as a 400, not a raw 500.
Mirrors src/lib/requests.ts's shape: Zod schema, ValidationResult
discriminated union, CRUD via Prisma, toRecord() camelCase->snake_case
mapping. managerId is only format-checked as a UUID here; a
nonexistent managerId is caught via Prisma's P2003 FK-violation error
in createEmployee() and translated into a friendly {ok:false,error}
instead of leaking a raw Prisma error.
GET and POST both call auth() first and return 401 { error:
Unauthorized } when there is no session, before doing any
validation or DB work. This is the API-side half of route-gating -
middleware.ts handles page routes but deliberately excludes /api,
so API routes verify the session themselves.
Flat permission model only (per CLAUDE.md): this checks logged-in
or not, nothing more granular. No role/permission logic added.
middleware.ts gates page routes by default (redirect to /login when
unauthenticated), except /login itself and Next.js internal asset
paths - config.matcher excludes /api entirely, /_next/static,
/_next/image, and favicon.ico. Default-deny: everything requires login
unless explicitly excluded, including the home page, rather than an
opt-in allowlist new routes could forget to join.
Builds its own lightweight NextAuth instance from auth.config.ts
(no Credentials provider) instead of importing auth from auth.ts, to
stay Edge-Runtime compatible - see auth.config.ts.
API routes are deliberately NOT gated here (config.matcher excludes
/api). Auth.js guidance is to verify the session as close to the data
fetching as possible rather than rely on middleware alone; API route
auth is added directly in each route handler instead (next commit).
This project pins Next.js 14.2.35, which predates the Next.js 16
middleware.ts -> proxy.ts rename, so this stays middleware.ts with a
default export, not proxy.ts.
toanhoang
toanhoang