How to read large JSON files that have different data types in it

I am creating an npm package that will read json files and then validate the content against some predefined json-schemas everything was working fine when I was testing against small size (~1MB) files but when I started reading 50MB and larger sizes things started to fail, like when I reach the code to pares the file Allocation failed - JavaScript heap out of memory is being thrown so I tried to increase the size of node memory node --max-old-space-size=4096 but now the parsing is taking forever (waited almost one hour and nothing happened).

So here is how the json file may look like:

[{
    "numberVerification": [
        {
            "longNumber": 281474976710655
        }
    ]
},
{
    "metaData": [
        {
            "name": "nodes",
            "elementCount": 155,
            "idCounter": 155,
            "version": "1.0",
        },
        {
            "name": "edges",
            "elementCount": 312,
            "idCounter": 312,
            "version": "1.0",
        },
        {
            "name": "networkAttributes",
            "elementCount": 14,
            "idCounter": 14,
            "version": "1.0",
        },
        {
            "name": "nodeAttributes",
            "elementCount": 330,
            "idCounter": 330,
            "version": "1.0",
        },
        {
            "name": "edgeAttributes",
            "elementCount": 3120,
            "idCounter": 3120,
            "version": "1.0",
        },
        {
            "name": "cartesianLayout",
            "elementCount": 155,
            "idCounter": 156
        },
    ]
},
{
    "nodes": [
        {
            "@id": 0,
            "n": "TYK2",
            "r": "uniprot:P29597"
        },
        {
            "@id": 1,
            "n": "ISGF3 complex",
            "r": "signor:SIGNOR-C124"
        },
        {...}
    ]
},
{
    "edges": [
        {
            "@id": 0,
            "s": 0,
            "t": 1,
            "i": "up-regulates activity"
        },
        {
            "@id": 1,
            "s": 2,
            "t": 1,
            "i": "up-regulates activity"
        },
        {...}
    ]
},
{
    "nodeAttributes": [
        {
            "po": 0,
            "n": "type",
            "v": "protein"
        },
        {
            "po": 0,
            "n": "location",
            "v": "cytoplasm"
        },
        {...}
    ]
},
{
    "edgeAttributes": [
        {
            "po": 0,
            "n": "citation",
            "v": [
                "pubmed:15120645"
            ],
            "d": "list_of_string"
        },
        {
            "po": 0,
            "n": "mechanism",
            "v": "phosphorylation"
        },
        {...}
    ]
},
{
    "cartesianLayout": [
    {
        "node": 0,
        "x": 97.73626669665249,
        "y": -114.99468800778627
    },
    {
        "node": 1,
        "x": 307.72737757573987,
        "y": 4.091777979752425
    },
    {...}
]
},
{
    "status": [
        {
            "error": "",
            "success": true
        }
    ]
}]

As you can see each object of the main array is of different type.

I was using fs.readFileSync to read json files and then I parse the whole file using JSON.parse(), but then I find that in order to read big json files I need to use streams but now how I am supposed to validate against json-schema? also I am doing some additional custom validation on the data, like checking if the @id property is unique, validate that the node (in cartesian layout array), t & s (in edges array) properties are pointing on a real @id in the nodes array, and after that I want to make some further statistics on the data.

So is there a way to read the large json files and parse it, and if it is applicable, how can I keep the validation running? and the statistic part do I need to use any external providers?

The json files are generated from a 3rd party so I can't really do anything with the structure of the data and how it is presented inside of them.

If anyone want me to provide any additional things let me know, I appreciate the help.