CSVConverter

A utility for converting CSV files into an array of JSON objects

Instructions

• Create the file config.json to specify CSV and JSON parameters
• Execute npm run convert
• You can optionally run npm run convert --config=path/to/config

Parameters

Parameter	Type	Description	Required
inputFile	string	Points to the input CSV file. File path should be relative to project directory	yes
outputFile	string	The resulting JSON file to be generated after processing. File path should be relative to the project directory	yes
alignment	"H"/"V"	Specifies which axis the CSV contents should be read along. For CSVs where each row will be an object, this should be "V"	yes
delimiter	char	The character separating data fields. Will most likely just be a ","	yes
headersInFirstLine	boolean	Specifies whether to consider the first line (along the specified alignment headers or data. If headers has a value, this is ignored. These are assessed after ignored rows/columns	no
headers	array	A user specified list of headers to use. This should match the number of items in a line based on the alignment and on any ignored rows/columns. This is an array of strings that will be used as JSON keys	no
trim	boolean	Specify whether to automatically remove extra empty resultant JSON objects	no
skipRows	number	Specifies a number of rows to ignore from the left. These are applied after determining the alignment	no
skipColumns	number	Specifies a number of columns to ignore from the left. These are applied after determining the alignment	no

Type Conversion

It's possible to provide information on what types values should be converted to in the final JSON by passing in an object instead of a string for the corresponding header in the headers parameter array. headers is able to take a combination of string or object in case no type conversion is needed for a particular header (will use string as the default type).

A header object can include:

{
    "name": <header/key name>,
    "type": <"string" | "number" | "array" | "bool" | "custom">,
    "delimiter": <array delimiter>, // "array" only
    "true": <value to check for true>, // "bool" only
    "transform": <transform function>, // "custom" only
    "span": <how many fields to encompass> // "custom" only
    "default": <default value if cell is empty>
}

An example of the headers parameter for a CSV for a Video Cloud video:

[
    "name",
    "short_description",
    "description",
    "related_link_url",
    "related_link_text",
    {
        "name": "duration",
        "type": "number",
        "default": 0
    },
    {
        "name": "tags",
        "type": "array",
        "delimiter": ","
    },
    {
        "name": "status",
        "type": "bool",
        "true": "true"
    }
    ...
]

Custom Types

It's possible to define a custom type by specifying custom as the type and providing the function: transform(value, index, values) => custom_value

Ex: Converting video status to human readable values

{
    "name": "status",
    "type": "custom",
    "transform": (value) => {
        return value === "true" ? "Active" : "Inactive";
    }
}

Ex: Creating link tags from the related_link_url and and related_link_text fields

 {
    "name": "related", // in place of `related_link_url`
    "type": "custom",
    "span": 2,
    "transform": (value, index, values) => {
        let url = value;
        let text = values[index + 1];
        return `<a href="${url}">${text}</a>`;
    }
}