Strange query bug with Next js, Knex, and SWR - javascript

Using Next API routes and Knex + MySQL, and using React and SWR for fetching, I get a strange bug. If a request fails, my queries start appending , * to the select statement, resulting in SQL syntax errors. For example, the query should use select *, but comes out as select *, * and then select *, *, * and so on. Does anyone have any idea why this might happen?
SWR fetch:
export const swrFetcher = async (...args) => {
const [url, contentType = 'application/json'] = args;
const res = await fetch(url, {
method: 'GET',
headers: {
'Content-Type': contentType,
},
});
if (!res.ok) throw new Error(res.statusText);
const json = await res.json();
return json;
};
const { data, error } = useSWR('/api/user/me', swrFetcher, {
revalidateOnFocus: false,
revalidateOnReconnect: false,
revalidateOnMount: true,
});
knex query:
const User = knex(TABLE_NAMES.user);
export const readById = (id) => User.select('*').where({ id });

You probably need to create the knex instance within the function call and not reuse the same every time, like it's currently happening.
export const readById = (id) => {
const User = knex(TABLE_NAMES.user);
return User.select('*').where({ id });
}

Related

How implement types properly?

I'm new to Typescript and have been doing a refactor a colleague code, I'm currently doing a typecheck and removing all any types. The goal is to make an MSGraph API call and return the a JSON file that translated into BirthdayPerson with a name, birthday date and a ID
I've been trying to a assign a type instead of any in the following code, but whether I assign number, string or any other type a different error will show up.
Perhaps I'm not tackling the solution correctly:
graph.ts
* #param accessToken
* #param endpoint url to call from MS Graph
*/
async function callMsGraph(accessToken: string, endpoint: string) {
const headers = new Headers();
const bearer = `Bearer ${accessToken}`;
headers.append('Authorization', bearer);
const options = {
method: 'GET',
headers,
};
try {
return fetch(endpoint, options);
} catch (error) {
console.error(error);
throw error;
}
}
export const callMsGraphWithoutPagination = async (
accessToken: string,
url: string,
dataToReturn: any[] = []
): Promise<any[]> => {
try {
const data = await callMsGraph(accessToken, url);
const dataJson = await data.json();
const newData = dataToReturn.concat(dataJson.value);
if (dataJson['#odata.nextLink']) {
const NEXT_URL = dataJson['#odata.nextLink'].split('/v1.0')[1];
return await callMsGraphWithoutPagination(
accessToken,
process.env.REACT_APP_GRAPH_URL + NEXT_URL,
newData
);
}
return dataToReturn.concat(dataJson.value);
} catch (error) {
/* eslint-disable no-console */
console.error(error);
/* eslint-enable no-console */
throw error;
}
};
export default callMsGraph;
useUsers.tsx
export const useUsers = () => {
const token = useToken();
const [users, setUsers] = React.useState<BirthdayPerson[]>([]);
React.useEffect(() => {
if (token) {
callMsGraphWithoutPagination(token, graphConfig.usersEndpoint).then(async (data: any) => {
const processedData: any[] = await Promise.all(
data.map(async (element: any) => {
const user = await callMsGraph(token, graphConfig.userBirthdayEndpoint(element.id));
const userJson = await user.json();
const image = await callMsGraph(token, graphConfig.userPhotoEndpoint(element.id));
const blob = await image.blob();
const returnElement: BirthdayPerson = {
displayName: element.displayName,
birthday: userJson.value,
id: element.id,
};
if (blob !== null) {
window.URL = window.URL || window.webkitURL;
returnElement.picture = window.URL.createObjectURL(blob);
}
return returnElement;
})
);
setUsers([].concat(...processedData));
});
}
}, [token]);
return users;
};
helpers.ts
interface IUpdateData {
slideCount: number;
}
const sortAndFilterBirthdays = (people: BirthdayPerson[], daysToGet: number) =>
people
.sort((firstEl, secondEl) => sortDate(firstEl.birthday, secondEl.birthday))
.filter(({ birthday }) => filterByAmountOfDays({ date: birthday, daysAfter: daysToGet }));
const getBirthdays: any = (people: BirthdayPerson[], daysToGet: number) => {
const validBirthdays = people.filter((element: any) => {
const year = moment(element.birthday).year();
return year !== 0;
});
const result = sortAndFilterBirthdays(validBirthdays, daysToGet);
// if it's okay
if (result.length > 1 && daysToGet <= 30) {
return result;
}
// if not okay, filters by future dates, concats with 'next year' dates, returns 2 dates
const fallbackResult = validBirthdays
.sort((firstEl, secondEl) => sortDate(firstEl.birthday, secondEl.birthday))
.filter((person: BirthdayPerson) => {
const currentYear = moment().year();
const date = moment(person.birthday, DATE_FORMAT).set('years', currentYear);
return moment().diff(date, 'days') <= 0;
});
return fallbackResult.concat(validBirthdays).splice(0, 2);
};
Any help or indication would be great!
From all the changes I've done another object will complain that Type 'x' is not assignable to type 'string'
Firstly you need to somehow type responses from API, because right now, as you have seen, call to .json() on Response object returns unknown, which make sense because no one knows what response the server returns. You may know what response it is expected to return, but not what it actually does.
Ideally therefore you need a parser that will verify that the response has correct structure and throws an error otherwise. There are libraries such as superstruct, yup, joi and others which you can use for this. Of course this is a lot of work and will need refactoring. Or if you don't care enough you can just cast the response object to appropriate type, but then if server returns something unexpected and the application cannot handle it, it's your fault.
Example with response parsing using superstruct
import {string, number, create, Infer} from 'superstruct'
// I assume `BirthdayUser` is just a string, but you can type more complex objects as well
const BirthdayUser = string()
// This is so that you don't have to list fields twice: once
// for the parser and once for typescript
type BirthdayUser = Infer<typeof BirthdayUser>
// Then use the parser
const response = await callMsGraph(acessToken, url)
const userJson = await response.json()
// user variable has inferred appropriate type, and if the response doesn't
// comply with the definition of `BirthdayUser` an error will be thrown
// Also I assume MS graph doesn't just return plain value but wraps it in an object with `value` field, so writing it here
const user = create(userJson, object({ value: BirthdayUser }))
Example of "I don't care enough" solution
type BirthdayUser = string
const response = await callMsGraph(accessToken, url)
// Same thing with wrapping into object with `value` field
const userJson = (await response.json()) as {value: BirthdayUser}
This is a bit awkward, because API call returns Response object and not the actual data. It might be easier to work with if you move parsing and casting logic inside of callMsGraph. It's not obligatory of course, but I still provide an example because after that you need to type callMsGraphWithoutPagination and it will use the same idea
import {object, create, Struct} from 'superstruct'
async function callMsGraphParsed<T>(
accessToken: string,
url: string,
// Since we need information about object structure at runtime, just making function
// generic is not enough, you need to pass the parser structure as an argument
struct: Struct<T>
) {
// ...
const response = await fetch(...)
const json = await response.json()
// Same, verifies that response complies to provided structure, if it
// does returns type object (of type `T`), otherwise throws an error
return create(json, object({ value: struct }))
}
async function callMsGraphLazy<T>(accessToken: string, url: string) {
// ...
const response = await fetch(...)
const json = await response.json()
return json as {value: T}
}
However I only call .json() here, if you want to use this solution, you will then need either a different function or another argument if you also want it to call .blob() for some API calls.
Now you type callMsGraphWithoutPagination using in the same way:
export const callMsGraphWithoutPaginationParsed = async <T>(
accessToken: string,
url: string,
dataToReturn: T[] = [],
struct: Struct<T>,
): Promise<T[]> => {
// If you rewrote `callMsGraph` to use parsing
const dataJson = await callMsGraph(accessToken, url, struct);
const newData = dataToReturn.concat(dataJson.value);
// ...
}
export const callMsGraphWithoutPaginationLazy= async <T>(
accessToken: string,
url: string,
dataToReturn: T[] = [],
): Promise<T[]> => {
// If you left `callMsGraph` as is
const data = await callMsGraph(accessToken, url);
const dataJson = (await data.json()) as {value: T}
const newData = dataToReturn.concat(dataJson.value);
// ...
}
Then use it
// Not sure if you are requesting `BirthdayUser` array here or some other entity, so change it to whatever you expect to receive
callMsGraphWithoutPagination<BirthdayUser>(token, graphConfig.usersEndpoint).then(async (data) => {
// "data" is inferred to have type BirthdayUser[]
data.map(element => {
// "element" is inferred to have type BirthdayUser
})
})
Also everywhere I wrote "I assume" and "Not sure" is missing info that you should probably have provided in the question. It didn't turn out to be relevant for me, but it could have. Good luck!

Strange behaviour of params.append with axios

export const getCharactersAsync = createAsyncThunk('getCharactersAsync', async (data) => {
const response = await axios.get('users', { params: { limit: data.limit } });
return response.data;
});
this code block allows me to control limit attribute.
export const getCharactersAsync = createAsyncThunk('getCharactersAsync', async (data) => {
const params = new FormData();
// const params = new URLSearchParams();
params.append('limit', data.limit);
const response = await axios.get('users', params);
console.log(response);
return response.data;
});
However I cannot control limit with using params.append. I tried URLSearchParams instead of FormData but still cannot manipulate limit attribute of the response. Why they differ from each other?
EDIT: This question has missleading information. I should have mention that i am using react-native. I found that react native doesn't fully support everything the web supports. So i need to install package called react-native-url-polyfill.Here is a github issues link
https://github.com/facebook/react-native/issues/23922#issuecomment-648096619
docs
params are the URL parameters to be sent with the request. Must be a plain object or a URLSearchParams object
It can't be FormData
Solution
You wanted to use { params }, not params
export const getCharactersAsync = createAsyncThunk('getCharactersAsync', async (data) => {
const params = new URLSearchParams();
params.append('limit', data.limit);
const response = await axios.get('users', { params });
console.log(response);
return response.data;
});

Next.js: filter on rest api url doesn't work when the route is a dynamic api route

I am trying to implement filter on API before fetching.
I have a dynamic API route([productType]), which works fine. But when I try to filter the API, it doesn't work.
So the syntax for filtering API looks like this ?$filter=name eq 'Milk' and adding this to the url doesn't work.
As in docs says, I have a tried to change the filename like this [...productType] or [[...productType]] but none of them helps. It returns the api data without filtering.
Here is my code in API dynamic route[productType]. pages/api/[productType]
import axios, { AxiosRequestConfig, Method } from 'axios'
import { pick } from 'lodash'
import { NextApiRequest, NextApiResponse } from 'next'
const baseUrl ='https://myUrl...'
const getFromAPIServer = async (
req: NextApiRequest,
res: NextApiResponse,
) => {
const { productType } = req.query
const param = req.query.params //here I tried to add the parameter for filter, which doesn't work
const url = `${baseUrl}/${productType}?${param}?`
const method = req.method as Method
const { body } = req
const allowedHeaders = ['If-Match']
const headers = pick(req.headers, allowedHeaders) as Record<string, string>
try {
const request: AxiosRequestConfig = {
method,
url,
headers: {
...headers,
Authorization: `Basic ${process.env.TOKEN}`,
'Content-Type': 'application/json',
},
}
if (method !== 'GET' && body) {
request.data = body
}
const { data } = await axios.request(request)
res.status(200).json(data)
} catch (error) {
if (!axios.isAxiosError(error)) {
throw error
}
// Return AxiosError to the client
res.status(error?.response?.status || 500).json(error?.response?.data)
}
}
export default getFromAPIServer
And in components I have the fetching function where I try to fetch the api with filter
export const getApiRequest = (url: string) => {
const { data } = useQuery(['key', url], async () => {
const response = await fetch(url, {
method: 'get',
})
if (!response.ok) throw new Error(response.statusText)
return await response.json()
})
return { data }
}
const { data } = getApiRequest(`${productTypeUrl}?$filter=name eq 'Milk'`)
Forgot to mention: In Postman filtering, the API works fine. It works also fine when I try to fetch with the filter url in components, not importing from api folder.
Update
Doing the filtering in api folder works as below
const url = `${baseUrl}/${productType}?$filter=name eq 'Milk'?`
However, this is not what I want. I want to have filter as params on API folder, like productType so that I can use it on client side.
Any help will be appreciated

SvelteKit: How to call mongodb without using endpoints?

(1/9/2023) Update : SvelteKit now supports server only load functions and Form actions to send requests to the server.
I want to call my database, but I don't want it be able to get accessed by end users by them going to the API endpoint that I set up. I was wondering how I would be able to just call my database from a file in the lib folder and just returning the data there. When I try it I get the error global not defined:
lib/db.js:
import dotenv from "dotenv";
dotenv.config();
import { MongoClient } from "mongodb";
const uri = process.env["MONGODB_URI"];
const options = {
useUnifiedTopology: true,
useNewUrlParser: true,
};
let client;
let clientPromise;
if (!uri) {
throw new Error("Please add your Mongo URI to .env.local");
}
if (process.env["NODE_ENV"] === "development") {
if (!global._mongoClientPromise) {
client = new MongoClient(uri, options);
global._mongoClientPromise = client.connect();
}
clientPromise = global._mongoClientPromise;
} else {
client = new MongoClient(uri, options);
clientPromise = client.connect();
}
export default clientPromise;
routes/items/index.js:
import clientPromise from "$lib/db";
export async function get() {
const client = await clientPromise;
const db = client.db();
const data = await db.collection("items").find({}).toArray();
const items = data.map(({ name }) => ({ name }));
if (items) {
return {
body: {
items,
},
};
}
}
My attempt:
lib/stores/items.js
import clientPromise from "$lib/db";
import { writable } from "svelte/store";
export const items= writable([]);
const fetchItems = async () => {
const client = await clientPromise;
const db = client.db();
const data = await db.collection("items").find({}).toArray();
const items = data.map(({ name }) => ({ name }));
substances.set(items);
};
fetchItems();
Trying the above code in various places always yields a global not defined error in the client.
I found one question from someone with the same problem, but I couldn't figure out how to create a helper file.
Protecting API is done on back-end side. Usually it either server (like NodeJS) or tools Nginx/Apache (proxy, etc.). You're basically looking for Content-Security-Policy topic, which is vaporous but not related to SvelteKit.
Btw, calling DB directly from the Front-end wouldn't be secure and is not possible.
To get data from any database, you should create enpoint
For user authentication, you can create handle hook:
export async function handle({ request, resolve }) {
let user = await authenticate(request)
request.locals.user = user
request.locals.isAuthenticated = !!user
if (request.path.startsWith('/api')) {
if (!user) {
return {
status: 401,
body: JSON.stringify({
error: {
message: 'Unauthorized'
}
})
}
}
const response = await resolve(request)
return response
}

How to migrate request-promise to axios or fetch

I want to code a app with React-Native which loads JSON-files from a website with cookie-authentication.
For testing I tried it in a normal JS-file without React-native and with request-promise.
const fs = require("fs");
const request = require("request-promise").defaults({ jar: true });
async function main() {
var incodeHeader = "";
var incodeToken = "";
try {
const loginResult = await request.post("https://somepage/login.php", {
form: {
client: "XXX",
login: "username",
password: "password",
},
});
} catch (err) {
console.log(err);
}
incodeHeader = getIncodeHeader();
incodeToken = getIncodeToken();
const data = await request.post("https://somepage/load.json", {
headers: {
[incodeHeader]: incodeToken,
},
form: {
max: "10",
},
});
fs.writeFileSync("data.json", data);
}
main();
This worked well, so I wanted to use this method in my App, but I couldn't find a way to use request-promise in React-Native so I decided to use axios.
const axios = require("axios");
const qs = require("qs");
axios.defaults.withCredentials = true;
async function main() {
const data = {
client: "XXX",
login: "username",
password: "password",
};
await axios
.post("https://somepage/login.php", qs.stringify(data))
.catch((err) => console.log(err));
const incodeHeader = getIncodeHeader();
const incodeToken = getIncodetoken();
await axios
.get(
"https://somepage/load.json",
{ data: { max: "5" } },
{
headers: {
[incodeHeader]: incodeToken,
},
}
)
.then((respone) => console.log(respone))
.catch((err) => console.log(err));
}
main();
But in this code not even the login works and I really don't know why. Can somebody tell me how to do this right, or can tell me another solution which works in React-Native?
First, I don't know why you're stringifying the request body in the first request, axios already handle this, you can pass just the data object, maybe it's the solution for your problem.
Second (just a tip). Create a helper object to make http requests and do not instance axios directly, so then, you can change the http request handler in an easy way instead changing it on each file, one day you probably will need to do this if you want to keep your app updated.
Third, don't mix await and then, choose:
try {
const result = await action();
// ...
} catch (err) {
// ...
}
or
action()
.then((result) => {
// ...
})
.catch((err) => {
// ...
});
change await axios.get to await axios.post

Categories

Resources