impress-2020/api/allWakaValues.js
Matchu 61e7a38b33 Normalize names for Waka, and log mismatches
By printing out this logging data, of item names we have that Waka doesn't, vs item names Waka has that we don't, I was able to solve a lot of them with new code in `normalizeItemName`!

Here's the mismatches that are left at time of writing:

```
[Item: Y, Waka: N] No Waka value for NC DTI item "goldenlulumedallion" (43034)
[Item: Y, Waka: N] No Waka value for NC DTI item "faelliebirthdaybagsurprise" (51350)
[Item: Y, Waka: N] No Waka value for NC DTI item "neopets11thbirthdaycommemorativemysterycapsule" (53448)
[Item: Y, Waka: N] No Waka value for NC DTI item "dyeworksblue:iscawig-blue" (70896)
[Item: Y, Waka: N] No Waka value for NC DTI item "mysteriousdoorwithlocks" (75601)
[Item: Y, Waka: N] No Waka value for NC DTI item "grapefruitnecklace" (77779)
[Item: Y, Waka: N] No Waka value for NC DTI item "discofeverbackground" (79250)
[Item: Y, Waka: N] No Waka value for NC DTI item "featherflaredshoes" (80047)
[Item: Y, Waka: N] No Waka value for NC DTI item "dyeworksredradioactivemutantmarkings" (80441)
[Item: N, Waka: Y] No NC DTI data for Waka item "7thbirthdaycakeslice#1"
[Item: N, Waka: Y] No NC DTI data for Waka item "7thbirthdaycakeslice#2"
[Item: N, Waka: Y] No NC DTI data for Waka item "7thbirthdaycakeslice#3"
[Item: N, Waka: Y] No NC DTI data for Waka item "8thbirthdayrainbowcupcake"
[Item: N, Waka: Y] No NC DTI data for Waka item "8thbirthdaysparklercupcake"
[Item: N, Waka: Y] No NC DTI data for Waka item "8thbirthdaytiedwithabowcupcake"
[Item: N, Waka: Y] No NC DTI data for Waka item "babyspringdress"
[Item: N, Waka: Y] No NC DTI data for Waka item "butterflydress(fromfaeriefestivalevent)"
[Item: N, Waka: Y] No NC DTI data for Waka item "discofever"
[Item: N, Waka: Y] No NC DTI data for Waka item "dyeworksblue:iscawig"
[Item: N, Waka: Y] No NC DTI data for Waka item "dyeworksred:radioactivemutantmarkings"
[Item: N, Waka: Y] No NC DTI data for Waka item "featherflairedshoes"
[Item: N, Waka: Y] No NC DTI data for Waka item "festivebooktree"
[Item: N, Waka: Y] No NC DTI data for Waka item "floralblackcardigan"
[Item: N, Waka: Y] No NC DTI data for Waka item "grapefruitneckace"
[Item: N, Waka: Y] No NC DTI data for Waka item "mysteriousdoorwithlocksbackground"
[Item: N, Waka: Y] No NC DTI data for Waka item "valiantchampionwings"
[Item: N, Waka: Y] No NC DTI data for Waka item "waxcrayonwig"
[Item: N, Waka: Y] No NC DTI data for Waka item "youngsophiesdress"
```
2021-04-07 17:25:58 -07:00

139 lines
4.4 KiB
JavaScript

const beeline = require("honeycomb-beeline")({
writeKey: process.env["HONEYCOMB_WRITE_KEY"],
dataset:
process.env["NODE_ENV"] === "production"
? "Dress to Impress (2020)"
: "Dress to Impress (2020, dev)",
serviceName: "impress-2020-gql-server",
});
import fetch from "node-fetch";
import connectToDb from "../src/server/db";
async function handle(req, res) {
const allNcItemNamesAndIdsPromise = loadAllNcItemNamesAndIds();
let itemValuesByName;
try {
itemValuesByName = await loadWakaValuesByName();
} catch (e) {
console.error(e);
res.setHeader("Content-Type", "text/plain");
res.status(500).send("Error loading Waka data from Google Sheets API");
return;
}
// Restructure the value data to use IDs as keys, instead of names.
const allNcItemNamesAndIds = await allNcItemNamesAndIdsPromise;
const itemValues = {};
for (const { name, id } of allNcItemNamesAndIds) {
if (name in itemValuesByName) {
itemValues[id] = itemValuesByName[name];
} else {
console.warn(
`[Item: Y, Waka: N] No Waka value for NC DTI item ${JSON.stringify(
name
)} (${id})`
);
}
}
const allNcItemNames = new Set(allNcItemNamesAndIds.map(({ name }) => name));
for (const name of Object.keys(itemValuesByName)) {
if (!allNcItemNames.has(name)) {
console.warn(
`[Item: N, Waka: Y] No NC DTI data for Waka item ${JSON.stringify(
name
)}`
);
}
}
// Cache for 1 minute, and immediately serve stale data for a day after.
// This should keep it fast and responsive, and stay well within our API key
// limits. (This will cause the client to send more requests than necessary,
// but the CDN cache should generally respond quickly with a small 304 Not
// Modified, unless the data really did change.)
res.setHeader(
"Cache-Control",
"public, max-age=3600, stale-while-revalidate=86400"
);
return res.send(itemValues);
}
async function loadAllNcItemNamesAndIds() {
const db = await connectToDb();
const [rows] = await db.query(`
SELECT items.id, item_translations.name FROM items
INNER JOIN item_translations ON item_translations.item_id = items.id
WHERE
(items.rarity_index IN (0, 500) OR is_manually_nc = 1)
AND item_translations.locale = "en"
`);
return rows.map(({ id, name }) => ({ id, name: normalizeItemName(name) }));
}
async function loadWakaValuesByName() {
if (!process.env["GOOGLE_API_KEY"]) {
throw new Error(`GOOGLE_API_KEY environment variable must be provided`);
}
const res = await fetch(
`https://sheets.googleapis.com/v4/spreadsheets/` +
`1DRMrniTSZP0sgZK6OAFFYqpmbT6xY_Ve_i480zghOX0/values/NC%20Values` +
`?fields=values&key=${encodeURIComponent(process.env["GOOGLE_API_KEY"])}`
);
const json = await res.json();
if (!res.ok) {
if (json.error) {
const { code, status, message } = json.error;
throw new Error(
`Google Sheets API returned error ${code} ${status}: ${message}`
);
} else {
throw new Error(
`Google Sheets API returned unexpected error: ${res.status} ${res.statusText}`
);
}
}
// Get the rows from the JSON response - skipping the first-row headers.
const rows = json.values.slice(1);
// Reformat the rows as a map from item name to value. We offer the item data
// as an object with a single field `value` for extensibility, but we omit
// the spreadsheet columns that we don't use on DTI, like Notes.
//
// NOTE: The Sheets API only returns the first non-empty cells of the row.
// So, when there's no value specified, it only returns one cell.
// That's why we set `""` as the default `value`.
const itemValuesByName = {};
for (const [itemName, value = ""] of rows) {
const normalizedItemName = normalizeItemName(itemName);
itemValuesByName[normalizedItemName] = { value };
}
return itemValuesByName;
}
function normalizeItemName(name) {
return (
name
// Remove all spaces, they're a common source of inconsistency
.replace(/\s+/g, "")
// Lower case, because capitalization is another common source
.toLowerCase()
// Remove diacritics: https://stackoverflow.com/a/37511463/107415
// Waka has some stray ones in item names, not sure why!
.normalize("NFD")
.replace(/[\u0300-\u036f]/g, "")
);
}
export default async (req, res) => {
beeline.withTrace({ name: "allWakaValues" }, () => handle(req, res));
};