Node API Issue

What you expected to happen

CALIBRE_HOST env variable should be set before the code reaches

What happened

While try to use the ad-hoc "Test" API using the node.js API approach, CALIBRE_HOST isn't set as an env variable

Steps to Reproduce (for bugs)

const { Test } = require('calibre');
const urlToTest = 'https://www.yahoo.com';

const payload = {
  url: urlToTest,
  location: 'NorthVirginia',
  device: 'iPhone8',
  connection: 'good3G'
};

Test.create(payload)
  .then(r => console.log(r))
  .catch(e => console.error(e));

I keep getting calibre: command not found when trying to run in terminal (mac). Installation and setup with API seemed to go through successfully and pretty straightforward. Am I missing something basic? :)

When I list my sites via CLI, apart from the name that is displayed in the calibre ui I also get site's slug:

calibre site list --json

[
  {
    "name": "FirstSite Name",
    "slug": "firstsite-name",
    "createdAt": "2015-05-19T07:23:39Z"
  },
  {
    "name": "SecondSite (Something)",
    "slug": "secondsite-something",
    "createdAt": "2015-10-05T09:42:53Z"
  }
]

But when I'm fetching list of pages for the site I'm getting only original name of the page and its uuid:

calibre site pages --site=firstsite-name --json

[
  {
    "uuid": "uuid1",
    "name": "Name Of The Page With Spaces",
    "url": "https://legiturl.com/",
    "canonical": true
  },
  {
    "uuid": "uuid2",
    "name": "Weird (N)a me",
    "url": "https://legiturl.com/2",
    "canonical": false
  }
]

If I want to get pulse metrics for particular page I have to guess its slug, because I can't use uuid there.

Shouldn't calibre site pages return also page's slug? 🤔

CLI Installation method

• NPM / NPX
• Binary

Output of calibre --version

1.2.4

Operating system

• Mac
• Windows
• Linux

What you expected to happen

From both the terminal interface (calibre test create [url] --[location]) and from a node script (using the example provided) I expect to run a test and after some time results returned or any errors handled gracefully.

What happened

Steps to Reproduce (for bugs)

1. From the CLI Terminal, run tests on a sample site.

After 3-4 tests the test will run for an extended period of time and this unhandled exception error will return: UnhandledPromiseRejectionWarning: Error: [object Object]

The number of tests before this error is returned has not been consistent. I was able to duplicate in both the example node script provided with the CLI and in the terminal using the CLI commands.

I've been unable to get any more information about the error other than the [object Object] portion.

CLI Installation method

• NPM
• Binary

Operating system

• Mac
• Windows
• Linux

What you expected to happen

I've been running this command:

$ ./node_modules/.bin/calibre test create http://$CIRCLE_BRANCH.$MYURL --location=SaoPaulo --device=iPhone8 --connection=good3G

According to the documentation, specifying iPhone8 and good3G as the device and connection parameters.

I expected a test to be run with these conditions.

What happened

http://$CIRCLE_BRANCH.$MYURL on Chrome Desktop

The --device and --connection parameters have been ignored without any error message.

How to improve

either:

• add a device-list and connection-list subcommand similar to location-list
• document the correct parameters and their possible values
• report an error for unknown devices and connection properties
• all of the above

When using Snapshot.create({ site: site.slug }), a snapshot will be created, however the ref of the snapshot will be set to undefined.

Using this issue to track changes coming into CLI for 2.0.0 release.

There's been a few things we've wanted to do but didn't because we didn't want to risk changing CLI flags etc.

• Add adblocking to single page tests. Add ability to turn it on or off for a test profile (#48)
• Tidy up command line flags so that they're all --dash-separated rather than --camelCase (#48)
• Rename flags used to turn JavaScript on / off (#48)
• Remove GIF artifact download (#50)
• Move get-pulse-metrics to use the top level timeseries query method
• Add / Remove deploy markers (#52)
• Change snapshots to use snapshotsList for pagination (#52)
• Return completed, errored and timeout tests with runtimeError (#53)
• Change the metrics chart for single page tests to swap out FMP for FCP (#54)
• Match Lighthouse grade colours to app (#54)

CLI Installation method

• NPM
• Binary

Operating system

• Mac
• Windows
• Linux

What you expected to happen

When an error occurs, calibre should exit with a non-zero error-code.

What happened

When an error is encountered, calibre exits with a 0 code.

Steps to Reproduce (for bugs)

1. npm install --save calibre
2. npx calibre site create-snapshot --site=INCORRECT_SITE_NAME --json
3. echo $?

When no device or connection limiting is used, the get-pulse-metrics --csv command errors.

At the moment our existing examples for creating a site using the Node.js API do not mention that the scheduler is turned off by default, nor does it show how to turn the scheduler on.

Let's update the examples to include this 👍

At the moment the Node AgentSettings API allows for settings to be updated for a given site, however there is no method to easily retrieve the current settings. They are also not returned via the Site.list function.

We'll need to add an AgentSettings.get function to make this more ergonomic.

This issue is to track the process of having support for managing budgets.

At the moment we do not have the capacity to mange (view, create, update, delete) metric budgets for a site. Add that.

I've successfully run + output the 7-days and (30 days) script to JSON. Fiddling around with it, it would be good to be able to adjust variables:
$sites
$pages
$duration
$metrics retrieved

For example:

    // Iterate through each page, returning a promise for the metrics being fetched
    return Promise.all(pages.map(async page => {
      const pulse = await SnapshotMetrics.pulse({
        site: site.slug,
        page: page.uuid,
        durationInDays: 30,
        metrics: ['consistently-interactive']
      })

      // Update the allPulseMetrics hash with metrics for this page
      allPulseMetrics[site.slug][page.uuid] = {
        page: page,
        pulse: pulse
      }

When I modified metrics: ['consistently-interactive'] to metrics: ['consistently-interactive','meaningful-paint'], I got back (node:84764) UnhandledPromiseRejectionWarning: [object Array] and when I tried other variations, I got back other similar errors. I also did not know which variables to put into metrics, because they are not named 1:1 with the desktop UI labels.

There are times that we will need to update our test profiles for cookies to provide required authentication information for example. Would it be possible to add this to the CLI?

Thanks :-)

Using this issue to track changes coming into CLI for 2.2.0 release.

• Add Integrations to the Node.JS API (#79)
• Add Metric Budgets to the Node.JS API (#78, #81)
• Add Agent Settings to the Node.JS API (#83)

Command:

calibre site get-snapshot-metrics --site=calibre --snapshot=1000000

Returns:

 TypeError: e.map is not a function

Command:

calibre site get-snapshot-metrics --site=calibre --snapshot=1000000 --json

Returns:

TypeError: Cannot read property 'tests' of null
    at Object.main [as handler] (/Users/micdijkstra/.asdf/installs/nodejs/11.6.0/.npm/lib/node_modules/calibre/src/cli/site/get-snapshot-metrics.js:68:27)
    at process.internalTickCallback (internal/process/next_tick.js:77:7)

It should return a more helpful error message.

CLI Installation method

• NPM / NPX
• Binary

Output of calibre --version

1.0.8

Operating system

• Mac
• Windows
• Linux

What you expected to happen

Following instructions in docs to be able to run the sample JS files which use the API.

What happened

Every example script I run results in the same error:

Steps to Reproduce (for bugs)

$ git clone [email protected]:calibreapp/cli.git calibre-cli
$ cd calibre-cli/examples/nodejs/
$ npm install
npm WARN [email protected] No repository field.

added 160 packages in 7.08s
$ CALIBRE_API_TOKEN=blah ./list-agent-locations.js
module.js:549
    throw err;
    ^

Error: Cannot find module './src/api/test_profile'
    at Function.Module._resolveFilename (module.js:547:15)
    at Function.Module._load (module.js:474:25)
    at Module.require (module.js:596:17)
    at require (internal/module.js:11:18)
    at Object.<anonymous> (/Users/henry/git/calibre-cli/examples/nodejs/node_modules/calibre/index.js:16:21)
    at Module._compile (module.js:652:30)
    at Object.Module._extensions..js (module.js:663:10)
    at Module.load (module.js:565:32)
    at tryModuleLoad (module.js:505:12)
    at Function.Module._load (module.js:497:3)

$ node --version
v8.11.3

CLI Installation method

• NPM / NPX
• Binary

Output of calibre --version

root@ip-*****:/home/delops# calibre --version
Traceback (most recent call last):
File "/usr/bin/calibre", line 19, in
from calibre.gui2.main import main
File "/usr/lib/calibre/calibre/gui2/init.py", line 21, in
from calibre.utils.date import UNDEFINED_DATE
File "/usr/lib/calibre/calibre/utils/date.py", line 13, in
from dateutil.tz import tzlocal, tzutc, EPOCHORDINAL
File "/usr/local/lib/python2.7/dist-packages/dateutil/tz/init.py", line 2, in
from .tz import *
File "/usr/local/lib/python2.7/dist-packages/dateutil/tz/tz.py", line 21, in
from six.moves import _thread

Operating system

• Mac
• Windows
• Linux

What you expected to happen

Clean installation of calibre CLI without any issue

What happened

I performed "npm install calibre -g ", but got the below error. I am not able to install calibre CLI in my Ubuntu machine.

1842 verbose linkStuff '/usr/local/lib/node_modules/calibre/node_modules' ]
1843 info linkStuff [email protected]
1844 silly gunzTarPerm extractEntry index.js
1845 silly gunzTarPerm extractEntry readme.md
1846 silly addNameRange number 2 { name: 'samverschueren/stream-to-observable',
1846 silly addNameRange range: '>=0.3.0-0 <0.4.0-0',
1846 silly addNameRange hasData: true }
1847 silly addNameRange versions [ undefined, [] ]
1848 error TypeError: Cannot read property 'latest' of undefined
1848 error at next (/usr/share/npm/lib/cache.js:687:35)
1848 error at /usr/share/npm/lib/cache.js:675:5
1848 error at saved (/usr/share/npm/node_modules/npm-registry-client/lib/get.js:142:7)
1848 error at /usr/lib/nodejs/graceful-fs/polyfills.js:133:7
1848 error at Object.oncomplete (fs.js:107:15)
1849 error If you need help, you may report this log at:
1849 error http://github.com/isaacs/npm/issues
1849 error or email it to:
1849 error [email protected]
1850 error System Linux 3.13.0-43-generic
1851 error command "/usr/bin/nodejs" "/usr/bin/npm" "install" "calibre" "-g"
1852 error cwd /root
1853 error node -v v0.10.25
1854 error npm -v 1.3.10
1855 error type non_object_property_load
1856 verbose exit [ 1, true ]

Steps to Reproduce (for bugs)

npm install calibre -g

Thanks in advance,
Rahul

CLI Installation method

• NPM
• Binary

Operating system

• Mac
• Windows
• Linux

What you expected to happen

I expected npx calibre site create-snapshot --site=SITENAME to create a snapshot without an error.

What happened

When attempting to create a snapshot, I received an error: Field 'createSnapshot' doesn't accept argument 'ref'

I was attempting to pass in a reference id with --ref but this error is occurring whether or not I specify a ref.

Steps to Reproduce (for bugs)

1. npm install --save calibre
2. npx calibre site create-snapshot --site=SITENAME

Dear calibre maintainers,
Thank you for your contribution to the open-source community.

This issue was automatically created to inform you a new version (5.0.0) of calibre was published without a matching tag in this repo.

As part of our efforts to fight software supply chain attacks, we would like to verify this release is known and intended, and not a result of an unauthorized activity.

If you find this behavior legitimate, kindly close and ignore this issue. Read more

• Update version
• Update minimum node version
• Update CHANGELOG
• Update docs
  • Create site with agent settings

• HAR files
• Screenshots
• GIFs
• Videos
• Lighthouse report JSON

Can we enable calling the api from the client-side?

My first approach was to try to use the node api as graphql-request works for both node and js. This didn't work as it expected a CALIBRE_API_TOKEN. The entire api takes input as environment variables.

If the node api could be refactored to allow passing those values it would make for a cleaner and more flexible api. If that's not feasible, can the graphql api be tweaked such that it works for cross origin requests?
I tried calling the graphql api directly but currently when I call the api from the client-side, the network request fails as newer browsers makes a pre-flight OPTIONS request which throws a 404.

CLI Installation method

• NPM / NPX
• Binary

Output of calibre --version

1.0.9

Operating system

• Mac
• Windows
• Linux

What you expected to happen

I expect that calibre test create --cookie-jar= to allow multiple cookies be passed to it.

What happened

Only the first cookie is used.

CLI Installation method

• NPM / NPX
• Binary

Output of calibre --version

Operating system

• Mac
• Windows
• Linux

What you expected to happen

Expecting Speed Index to be included in the metrics list

What happened

N/A

Steps to Reproduce

calibre site get-pulse-metrics --site=site-name --page=page-slug --metrics

CLI Installation method

• NPM / NPX
• Binary

Output of calibre --version

Operating system

• Mac
• Windows
• Linux

What you expected to happen

npx calibre site create-snapshot --site abc

What happened

it crashes

Steps to Reproduce (for bugs)

1. Use node 8, and latest calibre version: 1.2.0
2. npx calibre site create-snapshot --site abc
3. crash boom

CLI Installation method

• NPM / NPX
• Binary

Output of calibre --version

1.1.1

Operating system

• Mac
• Windows
• Linux

What you expected to happen

I'm using a script based on the get-all-pulse-metrics.js which has been working fine. Today the script failed with a GraphQL error.

What happened

The following GraphQL error is breaking the SnapshotMetrics.pulse call

[ { message: 'Field \'hasCompletedSnapshots\' doesn\'t exist on type \'Site\'',
    locations: [ [Object] ],
    fields:
     [ 'query GetPulsePageData',
       'organisation',
       'site',
       'hasCompletedSnapshots' ] } ]

Removing hasCompletedSnapshots field from the query (

Steps to Reproduce (for bugs)

Run the get-all-pulse-metrics.js example with valid API token

If a snapshot has failed, the response (json) returned will look something like:

"tests": [
  {
    "page": {
    "name": "Name",
    "url": "https://domain.tld/path"
  },
  "testProfile": {
    "id": "1"
  },
  "measurements": []
}

The view errors and does not exit cleanly (terminal view only, does not occur when the --json flag is in use)