File: dataschema/js/dataschema-text.js
/**
* Provides a DataSchema implementation which can be used to work with
* delimited text data.
*
* @module dataschema
* @submodule dataschema-text
*/
/**
Provides a DataSchema implementation which can be used to work with
delimited text data.
See the `apply` method for usage.
@class DataSchema.Text
@extends DataSchema.Base
@static
**/
var Lang = Y.Lang,
isString = Lang.isString,
isUndef = Lang.isUndefined,
SchemaText = {
////////////////////////////////////////////////////////////////////////
//
// DataSchema.Text static methods
//
////////////////////////////////////////////////////////////////////////
/**
Applies a schema to a string of delimited data, returning a normalized
object with results in the `results` property. The `meta` property of
the response object is present for consistency, but is assigned an
empty object. If the input data is absent or not a string, an `error`
property will be added.
Use _schema.resultDelimiter_ and _schema.fieldDelimiter_ to instruct
`apply` how to split up the string into an array of data arrays for
processing.
Use _schema.resultFields_ to specify the keys in the generated result
objects in `response.results`. The key:value pairs will be assigned
in the order of the _schema.resultFields_ array, assuming the values
in the data records are defined in the same order.
_schema.resultFields_ field identifiers are objects with the following
properties:
* `key` : <strong>(required)</strong> The property name you want
the data value assigned to in the result object (String)
* `parser`: A function or the name of a function on `Y.Parsers` used
to convert the input value into a normalized type. Parser
functions are passed the value as input and are expected to
return a value.
If no value parsing is needed, you can use just the desired property
name string as the field identifier instead of an object (see example
below).
@example
// Process simple csv
var schema = {
resultDelimiter: "\n",
fieldDelimiter: ",",
resultFields: [ 'fruit', 'color' ]
},
data = "Banana,yellow\nOrange,orange\nEggplant,purple";
var response = Y.DataSchema.Text.apply(schema, data);
// response.results[0] is { fruit: "Banana", color: "yellow" }
// Use parsers
schema.resultFields = [
{
key: 'fruit',
parser: function (val) { return val.toUpperCase(); }
},
'color' // mix and match objects and strings
];
response = Y.DataSchema.Text.apply(schema, data);
// response.results[0] is { fruit: "BANANA", color: "yellow" }
@method apply
@param {Object} schema Schema to apply. Supported configuration
properties are:
@param {String} schema.resultDelimiter Character or character
sequence that marks the end of one record and the start of
another.
@param {String} [schema.fieldDelimiter] Character or character
sequence that marks the end of a field and the start of
another within the same record.
@param {Array} [schema.resultFields] Field identifiers to
assign values in the response records. See above for details.
@param {String} data Text data.
@return {Object} An Object with properties `results` and `meta`
@static
**/
apply: function(schema, data) {
var data_in = data,
data_out = { results: [], meta: {} };
if (isString(data) && schema && isString(schema.resultDelimiter)) {
// Parse results data
data_out = SchemaText._parseResults.call(this, schema, data_in, data_out);
} else {
Y.log("Text data could not be schema-parsed: " + Y.dump(data) + " " + Y.dump(data), "error", "dataschema-text");
data_out.error = new Error("Text schema parse failure");
}
return data_out;
},
/**
* Schema-parsed list of results from full data
*
* @method _parseResults
* @param schema {Array} Schema to parse against.
* @param text_in {String} Text to parse.
* @param data_out {Object} In-progress parsed data to update.
* @return {Object} Parsed data object.
* @static
* @protected
*/
_parseResults: function(schema, text_in, data_out) {
var resultDelim = schema.resultDelimiter,
fieldDelim = isString(schema.fieldDelimiter) &&
schema.fieldDelimiter,
fields = schema.resultFields || [],
results = [],
parse = Y.DataSchema.Base.parse,
results_in, fields_in, result, item,
field, key, value, i, j;
// Delete final delimiter at end of string if there
if (text_in.slice(-resultDelim.length) === resultDelim) {
text_in = text_in.slice(0, -resultDelim.length);
}
// Split into results
results_in = text_in.split(schema.resultDelimiter);
if (fieldDelim) {
for (i = results_in.length - 1; i >= 0; --i) {
result = {};
item = results_in[i];
fields_in = item.split(schema.fieldDelimiter);
for (j = fields.length - 1; j >= 0; --j) {
field = fields[j];
key = (!isUndef(field.key)) ? field.key : field;
// FIXME: unless the key is an array index, this test
// for fields_in[key] is useless.
value = (!isUndef(fields_in[key])) ?
fields_in[key] :
fields_in[j];
result[key] = parse.call(this, value, field);
}
results[i] = result;
}
} else {
results = results_in;
}
data_out.results = results;
return data_out;
}
};
Y.DataSchema.Text = Y.mix(SchemaText, Y.DataSchema.Base);