@storyous/common-utils

Common JS utils

Migration guide

4.4.2

Prometheus middleware getHttpRequestMetricsMiddleware is now deprecated, you should remove it, same metrics can be aggregated from getRequestDurationMetricsMiddleware alone

5.2.0

Added loggly adapter. To use loggly you have to set silent:false and loggly token via LOGGLY_TOKEN env or directly in the config.

7.1.0

• Added default timeouts for getMongoCachedJSONFetcher (2000ms for fetch execution, 20000ms for background request)
• parallel call of the fetcher will not cause parallel requests to database nor the URL

8.0

logger.module returns pure Winston child instance, rename logger calls: log.e -> log.error log.w -> log.warn log.i -> log.info

9.0

getMongoCacheFetcher is now async mongoCachedFetcher - supports remote ETag - expose ifNoneMatch - returns result object instead of direct file content

10.0

concurrentTask accepts options object as the second parameter allowing configure: noLaterThan, startAttemptsDelay

11.0

apiTestUtil becomes testUtils. New usage:

config/testing.config.json

const testUtils = require('@storyous/common-utils/src/testUtils');

module.exports = {
   mongodbUrl: testUtils.uniqueDatabase(process.env.MONGODB_URI) // this will generate timestamp-postfixed database name 
        || 'mongodb://127.0.0.1:27018/myProjectTesting',
   // ...the rest of the config
};

test/api.js

const testUtils = require('@storyous/common-utils/src/testUtils');
const mocha = require('mocha');
const app = require('../app'); // this has to be a function providing Koa function

testUtils.init({ app, mocha });

module.exports = testUtils;

12.0

• Added mongoClient module (expects mongodbUrl property in config). Preconfigured native mongodb driver's client.
• Added collection getter returning native mongodb driver's collection. Usage: collection('myOrders').
• Removed db module - use collection & mongoClient instead.
  • collection changed to getCollection in version 14

13.0

Error handler is direct function. Usage:

const { errorHandler } = require('@storyous/common-utils');

// ...

app.use(errorHandler);

14.0

collection('collectionName')

changed to

getCollection('collectionName')

15.0

To have human-readable logs and errors, add

    logging: {
        console: {
            prettyOutput: true,
        },
    },

to development.config.js

Do NOT use anywhere else

15.3

Default mongoLocker() function introduced. Use a prefix for the key when you want to use it in multiple places in the app. Example:

// tokenStorage.js
await mongoLocker('token-renewal-process', async () => {
    // some async stuff
    // ...
    return 'myToken';
});

// payments.js - completely idempendent part of application
const transactionResult = await mongoLocker(`payment-transaction-${merchantId}`, async () => {
    // some async stuff
    // ...
    return true;
});

15.6.0

Loggly network errors does not cause exit of all app anymore

15.7.0

mongoLocker support 'expireIn' (millis) option which can be used to customize default 2 minutes acquisition. The mongoLocker now also handles expired document waiting to be deleted by a MongoDB background job.

16.0.0

Secrets now encrypt strings differently. Use encryptLegacy() function to achieve old behaviour. decrypt() function is compatible with legacy secrets automatically. encrypt() method now generates secrets compatible with implementation on Admin (PHP), and the updated decrypt() function.

16.3.0

storyousAuthorizedFetch - fetch for calling storyous services. It can obtain, cache and automatically refresh access token against the login service

// tokenStorage.js
const serviceResponse = await storyousAuthorizedFetch('https://api.storyous.com/delivery/somePath', {
    // mandatory parameters
    loginUrl: 'https://login.storyous.com',
    clientId: 'abc',
    clientSecret: 'def',
    
    // general fetch options
    method: 'POST',
    // ...
});

16.4.0

Add option to squash multiple logs into one based on URL app.use(log.basicLogMiddleware({ squashByUrls: ['/public/sodexo/restaurants'] }));

16.8.0

Run migrations now have option safeMigration that is true by default.

For new projects you have it to set it to false to allow complete rewriting of array with migrations. Remove safeMigration (or set it as true) when migrations are successful in all envs runMigrations(`${__dirname}/migrations`, {safeMigration: false});

MongoCachedFetcher

Usage:

const collection = mongodb.collection('myCachedFiles');

const fetcher = await getMongoCachedJSOFetcher(collection, /* optional */ {
    url: 'https://my.files.com/file1',
    cacheLifetime: 60 * 1000, // 60 seconds
    fetchOptions: { headers: { Authorization: 'myToken' } }, // options for remote fetch
    transform: async (content, key) => content, // allows decorate the fetched content just before its storage
    ensureIndexes: true, // it allow's more optimal cache manipulation
    logError: (err) => console.error(err)
});

/* optional parameters */
const parameters = {
    url: 'https://my.files.com/file2', // url of json content
    key: 'file2', // key, under which will be the content cached, url is used by default
    metaOnly: false, // boolean, if truthy the content is not returned, useful for finding cache freshness
    ifNoneMatch: 'someOldEtag', // saying, we want to get content only if the current etag is not equal to the value
};

const {
    content, // file content, null in case of etagMatch=true
    isCacheFresh, // boolean saying the content is not after its lifetime
    etag, // entity tag (version). If not null, it can be used in future fetcher calls as ifNoneMatch parameter
    etagMatch // boolean, truthy if isNoneMatch parameter provided and corresponds with latest cached etag value
} = await fetcher(parameters /* optional */ );

etag & ifNoneMatch

MongoCachedFetcher automatically stores ETag of remote resource if it is present in response from remote source. The stored etag is then used for consequent cache-refresh http call to optimise traffic - no data are transferred when the data didn't change. This functionality assumes the remote source of JSON data supports If-None-Match request header and ETag response header.

On top of that, the fetcher accepts optional ifNoneMatch parameter. If it is used, and its value matches currently stored (refreshed) etag value, result object will not contain content and the etagMatch will be true.