node package manager
Stop wasting time. Easily manage code sharing in your team. Create a free org »



This is a Node module and command line tool for creating structured data from tabular input using a declarative "meta-structure" in JSON or YAML, which looks something like this:

    '{y}': '{z}'

In a meta-structure, any object key with { and } markers becomes a nesting point for your data. Everything else (such as the top-level values key) is preserved as-is. In the case above, the tabular input would be assumed to have columns named x, y, and z. The other implicit assumption is that there is only one unique value of z for each permutation of x and y. So, if you had some CSV data like this:


Then nestly would combine the above meta-structure and data to create the following output in YAML:

    'a': '2'
    'b': '3'
    'a': '1'
    'b': '5'

Command line interface

The nestly command line tool works like this:

nestly [--config file | filename] data-filename [-o | output-filename]
  --cf          The format of the config file: "json" or "yaml"
                                                       [default: "json"]
  --if          The format of the input data: "csv", "tsv", "json", or
  --of          The format of the output data: "json" or "yaml"
                                                       [default: "json"]
  -c, --config  The path to your nesting configuration file
  -i, --in      The path of your input data file
  -o, --out     The name of the ouput file
  -h, --help    Show this help screen

The config should be a JSON or YAML file that encodes the "meta-structure" described above. If you don't provide the --if (input format) and --of (output format) options, then the formats are inferred from the filenames. (You should provide the format options if you're piping to or from stdio.) The above output would be generated with:

nestly --config structure.yml xyz.csv -o nested.yml

What problem does this solve?

I made it to ease the incorporation of data into Jekyll projects, where tabular formats can be tricky to work with. For instance, let's say you have some data in a spreadsheet like this:

City Year Population
San Diego 2012 1,337,000
San Diego 2013 1,356,000
San Francisco 2012 827,420
San Francisco 2013 837,442
San Jose 2012 982,579
San Jose 2013 998,537

If I wanted to get the population for a particular city and year in a template, I would need to do something funky like this:

{% assign row = | where: 'City', city | where: 'Year': year | first %}
{{ row.Population }}

But if we generated data with a structure like this:

    '{Year}': '{Population}'

Then getting the population for a city becomes:

{{[city].population[year] }}