-
-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathmain.js
More file actions
197 lines (179 loc) · 6.28 KB
/
main.js
File metadata and controls
197 lines (179 loc) · 6.28 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
/**
* @license Apache-2.0
*
* Copyright (c) 2024 The Stdlib Authors.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
'use strict';
// MODULES //
var isPlainObject = require( '@stdlib/assert-is-plain-object' );
var isFunction = require( '@stdlib/assert-is-function' );
var isndarrayLike = require( '@stdlib/assert-is-ndarray-like' );
var isOrder = require( '@stdlib/ndarray-base-assert-is-order' );
var hasOwnProp = require( '@stdlib/assert-has-own-property' );
var ctors = require( '@stdlib/ndarray-base-buffer-ctors' );
var zeros = require( '@stdlib/array-base-zeros' );
var getShape = require( '@stdlib/ndarray-shape' );
var getDType = require( '@stdlib/ndarray-dtype' );
var getOrder = require( '@stdlib/ndarray-order' );
var numel = require( '@stdlib/ndarray-base-numel' );
var nextCartesianIndex = require( '@stdlib/ndarray-base-next-cartesian-index' ).assign;
var gcopy = require( '@stdlib/blas-base-gcopy' );
var format = require( '@stdlib/string-format' );
// MAIN //
/**
* Filters and maps elements in an input ndarray to elements in a new output ndarray according to a callback function.
*
* @param {ndarray} x - input ndarray
* @param {Options} [options] - function options
* @param {string} [options.dtype] - output array data type
* @param {string} [options.order] - index iteration order
* @param {Callback} fcn - callback function
* @param {*} [thisArg] - callback execution context
* @throws {TypeError} first argument must be an ndarray-like object
* @throws {TypeError} callback argument must be a function
* @throws {TypeError} options argument must be an object
* @returns {ndarray} output ndarray
*
* @example
* var Float64Array = require( '@stdlib/array-float64' );
* var ndarray = require( '@stdlib/ndarray-ctor' );
*
* function fcn( v ) {
* if ( v > 5.0 ) {
* return v * 10.0;
* }
* }
*
* var buffer = new Float64Array( [ 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0 ] );
* var shape = [ 2, 3 ];
* var strides = [ 6, 1 ];
* var offset = 1;
*
* var x = ndarray( 'float64', buffer, shape, strides, offset, 'row-major' );
* // returns <ndarray>
*
* var y = filterMap( x, fcn );
* // returns <ndarray>[ 80.0, 90.0, 100.0 ]
*/
function filterMap( x, options, fcn, thisArg ) {
var hasOpts;
var ndims;
var cache;
var clbk;
var opts;
var ctor;
var ctx;
var ord;
var dim;
var idx;
var buf;
var dt;
var sh;
var N;
var y;
var v;
var i;
if ( !isndarrayLike( x ) ) {
throw new TypeError( format( 'invalid argument. First argument must be an ndarray-like object. Value: `%s`.', x ) );
}
if ( arguments.length < 3 ) {
clbk = options;
} else if ( arguments.length === 3 ) {
if ( isFunction( options ) ) {
clbk = options;
ctx = fcn;
} else {
hasOpts = true;
opts = options;
clbk = fcn;
}
} else {
hasOpts = true;
opts = options;
clbk = fcn;
ctx = thisArg;
}
if ( !isFunction( clbk ) ) {
throw new TypeError( format( 'invalid argument. Callback argument must be a function. Value: `%s`.', clbk ) );
}
if ( hasOpts ) {
if ( !isPlainObject( opts ) ) {
throw new TypeError( format( 'invalid argument. Options argument must be an object. Value: `%s`.', opts ) );
}
if ( hasOwnProp( opts, 'dtype' ) ) {
dt = opts.dtype;
} else {
dt = getDType( x );
}
if ( hasOwnProp( opts, 'order' ) ) {
if ( !isOrder( opts.order ) ) {
throw new TypeError( format( 'invalid option. `%s` option must be a recognized order. Option: `%s`.', 'order', opts.order ) );
}
ord = opts.order;
}
} else {
dt = getDType( x );
}
// Resolve an output array buffer constructor:
ctor = ctors( dt );
if ( ctor === null ) {
// The only way we should get here is if the user provided an unsupported data type, as `getDType` should error if the input array has an unrecognized/unsupported data type...
throw new TypeError( format( 'invalid option. `%s` option must be a recognized data type. Option: `%s`.', 'dtype', opts.dtype ) );
}
// Resolve the iteration order:
if ( ord === void 0 ) {
ord = getOrder( x );
}
// Resolve the input array shape:
sh = getShape( x );
// Compute the number of array elements:
N = numel( sh );
// Retrieve the number of dimensions:
ndims = sh.length;
// Resolve the dimension in which indices should iterate fastest:
if ( ord === 'row-major' ) {
dim = ndims - 1;
} else { // ord === 'column-major'
dim = 0;
}
// Initialize an index array workspace:
idx = zeros( ndims );
// Initialize a value cache for those elements which pass a callback function (note: unfortunately, we use an associative array here, as no other good options. If we use a "generic" array, we are limited to 2^32-1 elements. If we allocate, say, a Float64Array buffer for storing indices, we risk materializing lazily-materialized input ndarray values again (e.g., lazy accessor ndarray), which could be expensive. If we allocate a workspace buffer of equal size to the input ndarray to store materialized values, we'd then need to perform another copy in order to shrink the buffer, as, otherwise, could be holding on to significantly more memory than needed for the returned ndarray. There are likely other options, but all involve complexity, so the simplest option is used here.):
cache = {
'length': 0
};
// Filter and map elements according to a callback function...
for ( i = 0; i < N; i++ ) {
if ( i > 0 ) {
idx = nextCartesianIndex( sh, ord, idx, dim, idx );
}
v = clbk.call( ctx, x.get.apply( x, idx ), idx.slice(), x );
if ( v !== void 0 ) {
cache[ cache.length ] = v;
cache.length += 1;
}
}
// Retrieve the number of cached elements:
N = cache.length;
// Allocate an output array buffer:
buf = new ctor( N );
// Copy cached elements to the output array buffer:
gcopy( N, cache, 1, buf, 1 );
// Create an output ndarray:
y = new x.constructor( dt, buf, [ N ], [ 1 ], 0, ord );
return y;
}
// EXPORTS //
module.exports = filterMap;