Is it possible to use the Mongoose query builder to just return an aggregation pipeline array and not run the query?

Question

I am working locally - this would never be done on a live website. I built a node/express server that will accept a stringified aggregation pipeline for Mongo, and return the results to the browser:

app.get('/aggregate/:collection', function(req, res) {

    var queryObject = JSON.parse(req.param('q'));

    var recProvider = getProviderForTable(req.params.collection);
    recProvider.getCollection(function(err, collection) {

        collection.aggregate(queryObject, {}, function(err, data) {
            res.json(data);
        });
    });

});

This makes it convenient to do some quick queries in the browser where I am building data visualizations:

$.get(LOCAL_SERVER + '/aggregate/my_records?q=' + JSON.stringify(pipeline));

I almost started to build a query builder like Mongoose has, but just to construct the aggregation pipeline array. I am wondering if I can use Mongoose's query builder in the browser just to make the array, then use that to hit my express server for the data, as above?

I want to build a chainable object so eg...

pipeline()
    .where('value').gt(1000)
    .where('category').in([1, 2, 3])
    .sort('-date')
    .skip(100).limit(10);

...would return the aggregation pipeline:

[
    {
        $match: {
            value: {
                $gt: 1000
            },
            category: {
                $in: [1, 2, 3]
            }
        }
    },
    {
        $sort: {
            date: -1
        }
    },
    {
        $skip: 100,
        $limit: 10
    }
]

Answer 1

Looking through the source, it looks like whenever you're using the aggregate() function, the pipeline is stored in a property called _pipeline.

Using your example, if you write an aggregation pipeline like this...

var myAggregation = Model  // "Model" should actually be the name of your model
    .aggregate()
    .match({ value: { $gt: 1000 } })
    .match({ category: { $in: [1, 2, 3] } })
    .sort('-date')
    .skip(100)
    .limit(10);

Then myAggregation._pipeline will look like this...

[ { '$match': { value: { $gt: 1000 } } },
  { '$match': { category: { $in: [1, 2, 3] } } },
  { '$sort': { date: -1 } },
  { '$skip': 100 },
  { '$limit': 10 } ]

I noticed, however, that you're using the where() function, which actually creates a Query in mongoose instead of an aggregation pipeline. If you're set on using where(), the conditions and options are set a little differently. Instead of myAggregation._pipeline, they will be split into myAggregation._conditions and myAggregation.options. They are not written exactly the same way, but perhaps you can do some transformation into the format you want.