Hide extraneous GCN plumbing from returned objects

[adam-sandor]

Fetching a list of files from my bucket produces 271K of JSON output from the server. And I just have 4 files in the bucket.
The data includes things like the complete description of the bucket itself (per every file) or the description of the googlecloud storage library (authors, dependencies, etc). Some of this is probably not loaded from the server but is added by the node api library. Even if that doesn't lead to network traffic it's still unnecessary and inconvenient when I try to find the relevant data. The screenshot below show the data structure with most elements collapsed. None of those elements are relevant for listing files in the bucket. I only want to see the data that is printed by gsutil when doing ls.

https://www.dropbox.com/s/4dgpu7jh8z6lb9h/Screen%20Shot%202017-04-13%20at%2013.29.58.png?dl=0

I might be missing some option or parameter that would limit the output from getFiles.

Environment details

• OS: OSX
• Node.js version: 7.2.0
• npm version: 3.10.9
• google-cloud-node version: 1.0.0

Steps to reproduce

var storage = require('@google-cloud/storage');
var fs = require('fs');

var gcs = storage({
    projectId: 'MYPROJECT'
});

var bucket = gcs.bucket('MYBUCKET');

bucket.getFiles(function(err, files) {
    if (!err) {
        fs.writeFileSync("files.json", JSON.stringify(files))
    } else {
        console.error(err)
    }
});

[stephenplusplus]

There is a lot of stuff on the object-- some is plumbing that enables our API to function. The only thing that's really meant for the user is the metadata property, which is the API's raw record of the resource.

In the context of getFiles(), the API response returns all of the metadata for each file. If we didn't include it on the object, it would require further API calls to download that information. To find the data you need, you should look into to the metadata object. In the case that you just want names, it's a simple map over the response:

bucket.getFiles(function(err, files) {
    if (!err) {
        var fileMetadataObjects = files.map(function(file) {
            return file.metadata
        })
        fs.writeFileSync("files.json", JSON.stringify(fileMetadataObjects))
    } else {
        console.error(err)
    }
})

And of course, if you just want the name, you can use file.metadata.name or any other properties that are relevant for your use.

[stephenplusplus]

Just to address the 271,000 lines of JSON, though, I'm not particularly thrilled about that. From a quick test, the output of one file for me was about 1,200 lines. I'll re-open so we can define those huge objects as non-enumerable, so they're hidden from the output.

[adam-sandor]

Ok so do I understand correctly that only the metadata is actually downloaded from the server?

[stephenplusplus]

Yes, everything else is state and internal data.

[adam-sandor]

Cool, then it was just me panicking :)
Still would be user friendlier to not have that metadata in the request I think.

[lukesneeringer]

This issue was moved to googleapis/nodejs-storage#19