feat(orders): make an ingested order legible in the transactions view
ci / lint-test (push) Failing after 43s

Four things the view could not tell you, all from reading the rows (user,
2026-07-27).

**Which platform.** The parser has always known — it has to, to read the
template — and then discarded it. "Order - Burger Corner" gives no way to know
whether to open DoorDash or Uber Eats for the detail, and restaurants exist on
both. Now stored on expense_metadata and named in the description:
"Order - Burger Corner (Uber Eats)". Migration 0021 recovers it for the 101
backfilled rows from the order_reference shape — DoorDash receipts carry no id
of their own so ingestion synthesises `msg:<message-id>`, Uber carries a real
trip UUID, which makes the discriminator exact.

**Bank said "Manual".** That label is derived, not stored, and "Manual" reads
as "hand-entered, still awaiting a card line to match". A gift-card order has
no card line coming, ever. It now reads "Gift Card", and — the part that
actually mattered — credits joins cash in needsCardMatch(), so these stop
sitting in the pending-reconciliation queue. All 81 were queued against a match
that could not exist.

**Uber line items were never parsed.** 67 of 101 orders had none. Uber itemises
groceries but not restaurant orders, so some of that is genuine; the rest was
simply unread. Its markup is better than DoorDash's — every cell carries a
data-testid with the item's uuid, so qty/title/amount bind by id rather than by
column position. Sold-out items (0.00) are kept: they are why a total is lower
than what was ordered.

**Uber prints pick-up and delivery addresses on every receipt** and they were
thrown away. Captured as `route` [{label, time, address}], de-duplicated
because the template renders the whole block twice for narrow screens. Wording
is kept as printed ("Pick-up" on some receipts, "Pickup" on others) rather than
normalised, so a template change stays visible. This is the same block a *trip*
receipt uses for start and destination — rides are not ingested today, but the
reader will not need changing when they are.

Also stores source_email_subject/from, which order ingestion had left null on
columns that already existed.

Verified against the captured corpus: route on all 6 Uber fixtures, 5/5 items
on the GLOMARK grocery receipt including the sold-out one. Production data
updated by smarthome:docker/scripts/order-presentation-2026-07-27.sql
(81 descriptions, `backfill` tag, re-run clean). `route` and Uber line items
are parsed from here on only — recovering them for already-ingested orders
means re-reading the mail, which I7 idempotency refuses by design.
This commit is contained in:
2026-07-27 10:51:30 +10:00
parent ae0c34fce7
commit df4b875b82
7 changed files with 354 additions and 29 deletions
+95 -1
View File
@@ -20,6 +20,18 @@ export interface LineItem {
options?: string[];
}
/**
* A stop on the receipt's map: pick-up, delivery, or (for a trip) the ride's
* start and end. Uber prints these for every order under `Order details`.
*/
export interface RoutePoint {
/** "Pick-up" / "Delivery" — whatever the receipt itself calls it. */
label: string;
/** Local time as printed, e.g. "1:20 pm". No date; the receipt gives none. */
time: string | null;
address: string;
}
export interface PaymentBreakdown {
credits_amount: number | null;
card_amount: number | null;
@@ -47,6 +59,8 @@ export interface ParsedOrder {
payment: PaymentBreakdown;
totals: OrderTotals;
line_items: LineItem[];
/** Uber only. Empty for DoorDash, whose receipts carry no addresses. */
route: RoutePoint[];
is_family: boolean;
flags: string[];
}
@@ -116,6 +130,81 @@ function tdPairValue(html: string, label: string): number | null {
return m ? money(m[1]) : null;
}
/**
* Uber itemises only *grocery* orders — a restaurant receipt states a total and
* nothing else, which is why 67 of the 101 backfilled orders have no items.
* When it does itemise, the markup is far better than DoorDash's: every cell
* carries a `data-testid` naming its role and the item's own uuid, so quantity,
* title and amount can be bound to each other by id rather than by position.
*/
function parseUberLineItems(html: string): LineItem[] {
const items: LineItem[] = [];
const titleRe =
/data-testid="shoppingCart_item_title_([0-9a-f-]+)"[^>]*>([\s\S]*?)<\/td>/gi;
for (const m of html.matchAll(titleRe)) {
const [, id, rawTitle] = m;
const description = collapse(decodeEntities(stripTags(rawTitle)));
if (!description) continue;
const qtyM = html.match(
new RegExp(`data-testid="shoppingCart_item_quantity_${id}"[^>]*>\\s*(\\d+)\\s*<`, "i")
);
const amtM = html.match(
new RegExp(
`data-testid="shoppingCart_item_amount_${id}"[^>]*>([\\s\\S]*?)<\\/td>`,
"i"
)
);
const amtText = amtM ? collapse(decodeEntities(stripTags(amtM[1]))) : "";
const amtNum = amtText.match(/(-?[\d,]+\.\d{2})/);
items.push({
qty: qtyM ? parseInt(qtyM[1], 10) : 1,
description,
// A sold-out item prints 0.00 and is genuinely part of the order — it
// explains a total that does not match what was asked for. Keep it.
amount: amtNum ? money(amtNum[1]) : 0,
});
}
return items;
}
/**
* Uber's `Order details` block, anchored on `data-testid="address_point_N_*"`.
*
* The template repeats the whole block twice (once hidden for narrow screens),
* so the same stop appears more than once and has to be de-duplicated. This is
* the same markup a *trip* receipt uses for its start and destination — rides
* are not ingested today, but the reader will not need changing when they are.
*/
function parseUberRoute(html: string): RoutePoint[] {
const seen = new Set<string>();
const points: RoutePoint[] = [];
const labelRe = /data-testid="address_point_(\d+)_time"[^>]*>([\s\S]*?)<\/td>/gi;
for (const m of html.matchAll(labelRe)) {
const [, idx, rawLabel] = m;
const addrM = html.match(
new RegExp(`data-testid="address_point_${idx}_address"[^>]*>([\\s\\S]*?)<\\/td>`, "i")
);
if (!addrM) continue;
const address = collapse(decodeEntities(stripTags(addrM[1])));
// "1:20 pm - Pick-up" — time and label share one cell.
const combined = collapse(decodeEntities(stripTags(rawLabel)));
const split = combined.match(/^(.*?)\s+-\s+(.*)$/);
const time = split ? split[1] : null;
const label = split ? split[2] : combined;
const key = `${label}|${time}|${address}`;
if (!address || seen.has(key)) continue;
seen.add(key);
points.push({ label, time, address });
}
return points;
}
function parseDoorDashLineItems(html: string): LineItem[] {
// <td width="10%">1x</td><td width="75%"><b>Name</b> (Cat)<br><font>• Opt</font>…</td><td width="15%">$22.10</td>
const re =
@@ -397,11 +486,15 @@ export function parseOrderHTML(html: string, meta: MessageMeta): ParsedOrder {
// ---- line items ----------------------------------------------------------
// Uber Eats receipts carry no itemisation (verified across 29 real mails).
const line_items =
platform === "doordash" ? parseDoorDashLineItems(clean) : [];
platform === "doordash" ? parseDoorDashLineItems(clean) : parseUberLineItems(clean);
if (platform === "doordash" && line_items.length === 0) {
flags.push("no_line_items_parsed");
}
// Uber prints addresses on every receipt; DoorDash prints none at all, so an
// empty route there is expected rather than a parse failure.
const route = platform === "doordash" ? [] : parseUberRoute(clean);
const currency =
explicitCurrency ||
(/\b(NZD|USD|LKR|CHF|EUR|GBP|SGD|INR)\b/.test(text)
@@ -417,6 +510,7 @@ export function parseOrderHTML(html: string, meta: MessageMeta): ParsedOrder {
payment,
totals,
line_items,
route,
is_family,
flags,
};