This file is indexed.

/usr/share/doc/python-dask-doc/html/custom-collections.html is in python-dask-doc 0.16.0-1.

This file is owned by root:root, with mode 0o644.

The actual contents of the file can be viewed below.

  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
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
<!DOCTYPE html>
<!--[if IE 8]><html class="no-js lt-ie9" lang="en" > <![endif]-->
<!--[if gt IE 8]><!--> <html class="no-js" lang="en" > <!--<![endif]-->
<head>
  <meta charset="utf-8">
  
  <meta name="viewport" content="width=device-width, initial-scale=1.0">
  
  <title>Custom Collections &mdash; dask 0.16.0 documentation</title>
  

  
  
  
  

  

  
  
    

  

  
  
    <link rel="stylesheet" href="_static/css/theme.css" type="text/css" />
  

  
    <link rel="stylesheet" href="_static/style.css" type="text/css" />
  

  
        <link rel="index" title="Index"
              href="genindex.html"/>
        <link rel="search" title="Search" href="search.html"/>
    <link rel="top" title="dask 0.16.0 documentation" href="index.html"/>
        <link rel="next" title="Citations" href="cite.html"/>
        <link rel="prev" title="Remote Data Services" href="remote-data-services.html"/> 

  
  <script src="_static/js/modernizr.min.js"></script>

</head>

<body class="wy-body-for-nav" role="document">

   
  <div class="wy-grid-for-nav">

    
    <nav data-toggle="wy-nav-shift" class="wy-nav-side">
      <div class="wy-side-scroll">
        <div class="wy-side-nav-search">
          

          
            <a href="index.html" class="icon icon-home"> dask
          

          
          </a>

          
            
            
              <div class="version">
                0.16.0
              </div>
            
          

          
<div role="search">
  <form id="rtd-search-form" class="wy-form" action="search.html" method="get">
    <input type="text" name="q" placeholder="Search docs" />
    <input type="hidden" name="check_keywords" value="yes" />
    <input type="hidden" name="area" value="default" />
  </form>
</div>

          
        </div>

        <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
          
            
            
              
            
            
              <p class="caption"><span class="caption-text">Getting Started</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="install.html">Install Dask</a></li>
<li class="toctree-l1"><a class="reference internal" href="use-cases.html">Use Cases</a></li>
<li class="toctree-l1"><a class="reference internal" href="examples-tutorials.html">Examples</a></li>
<li class="toctree-l1"><a class="reference internal" href="cheatsheet.html">Dask Cheat Sheet</a></li>
</ul>
<p class="caption"><span class="caption-text">Collections</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="array.html">Array</a></li>
<li class="toctree-l1"><a class="reference internal" href="bag.html">Bag</a></li>
<li class="toctree-l1"><a class="reference internal" href="dataframe.html">DataFrame</a></li>
<li class="toctree-l1"><a class="reference internal" href="delayed.html">Delayed</a></li>
<li class="toctree-l1"><a class="reference internal" href="futures.html">Futures</a></li>
<li class="toctree-l1"><a class="reference internal" href="machine-learning.html">Machine Learning</a></li>
</ul>
<p class="caption"><span class="caption-text">Scheduling</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="distributed.html">Distributed Scheduling</a></li>
<li class="toctree-l1"><a class="reference internal" href="scheduler-overview.html">Scheduler Overview</a></li>
<li class="toctree-l1"><a class="reference internal" href="scheduler-choice.html">Choosing between Schedulers</a></li>
<li class="toctree-l1"><a class="reference internal" href="shared.html">Shared Memory</a></li>
<li class="toctree-l1"><a class="reference internal" href="scheduling-policy.html">Scheduling in Depth</a></li>
</ul>
<p class="caption"><span class="caption-text">Diagnostics</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="inspect.html">Inspecting Dask objects</a></li>
<li class="toctree-l1"><a class="reference internal" href="diagnostics.html">Diagnostics</a></li>
</ul>
<p class="caption"><span class="caption-text">Graphs</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="graphs.html">Overview</a></li>
<li class="toctree-l1"><a class="reference internal" href="spec.html">Specification</a></li>
<li class="toctree-l1"><a class="reference internal" href="custom-graphs.html">Custom Graphs</a></li>
<li class="toctree-l1"><a class="reference internal" href="optimize.html">Optimization</a></li>
</ul>
<p class="caption"><span class="caption-text">Help &amp; reference</span></p>
<ul class="current">
<li class="toctree-l1"><a class="reference internal" href="debugging.html">Debugging</a></li>
<li class="toctree-l1"><a class="reference internal" href="support.html">Contact and Support</a></li>
<li class="toctree-l1"><a class="reference internal" href="changelog.html">Changelog</a></li>
<li class="toctree-l1"><a class="reference internal" href="presentations.html">Presentations On Dask</a></li>
<li class="toctree-l1"><a class="reference internal" href="develop.html">Development Guidelines</a></li>
<li class="toctree-l1"><a class="reference internal" href="faq.html">Frequently Asked Questions</a></li>
<li class="toctree-l1"><a class="reference internal" href="spark.html">Comparison to PySpark</a></li>
<li class="toctree-l1"><a class="reference internal" href="caching.html">Opportunistic Caching</a></li>
<li class="toctree-l1"><a class="reference internal" href="bytes.html">Internal Data Ingestion</a></li>
<li class="toctree-l1"><a class="reference internal" href="remote-data-services.html">Remote Data Services</a></li>
<li class="toctree-l1 current"><a class="current reference internal" href="#">Custom Collections</a><ul>
<li class="toctree-l2"><a class="reference internal" href="#the-dask-collection-interface">The Dask Collection Interface</a></li>
<li class="toctree-l2"><a class="reference internal" href="#internals-of-the-core-dask-methods">Internals of the Core Dask Methods</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#compute">Compute</a></li>
<li class="toctree-l3"><a class="reference internal" href="#persist">Persist</a></li>
<li class="toctree-l3"><a class="reference internal" href="#visualize">Visualize</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="#adding-the-core-dask-methods-to-your-class">Adding the Core Dask Methods to Your Class</a></li>
<li class="toctree-l2"><a class="reference internal" href="#example-dask-collection">Example Dask Collection</a></li>
<li class="toctree-l2"><a class="reference internal" href="#checking-if-an-object-is-a-dask-collection">Checking if an object is a dask collection</a></li>
<li class="toctree-l2"><a class="reference internal" href="#implementing-deterministic-hashing">Implementing Deterministic Hashing</a><ul>
<li class="toctree-l3"><a class="reference internal" href="#example">Example</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l1"><a class="reference internal" href="cite.html">Citations</a></li>
<li class="toctree-l1"><a class="reference internal" href="funding.html">Funding</a></li>
</ul>

            
          
        </div>
      </div>
    </nav>

    <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">

      
      <nav class="wy-nav-top" role="navigation" aria-label="top navigation">
        
          <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
          <a href="index.html">dask</a>
        
      </nav>


      
      <div class="wy-nav-content">
        <div class="rst-content">
          















<div role="navigation" aria-label="breadcrumbs navigation">

  <ul class="wy-breadcrumbs">
    
      <li><a href="index.html">Docs</a> &raquo;</li>
        
      <li>Custom Collections</li>
    
    
      <li class="wy-breadcrumbs-aside">
        
            
            <a href="_sources/custom-collections.rst.txt" rel="nofollow"> View page source</a>
          
        
      </li>
    
  </ul>

  
  <hr/>
</div>
          <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
           <div itemprop="articleBody">
            
  <div class="section" id="custom-collections">
<h1>Custom Collections<a class="headerlink" href="#custom-collections" title="Permalink to this headline"></a></h1>
<p>For many problems the built-in dask collections (<code class="docutils literal"><span class="pre">dask.array</span></code>,
<code class="docutils literal"><span class="pre">dask.dataframe</span></code>, <code class="docutils literal"><span class="pre">dask.bag</span></code>, and <code class="docutils literal"><span class="pre">dask.delayed</span></code>) are sufficient. For
cases where they aren’t it’s possible to create your own dask collections. Here
we describe the required methods to fullfill the dask collection interface.</p>
<div class="admonition warning">
<p class="first admonition-title">Warning</p>
<p class="last">The custom collection API is experimental and subject to change
without going through a deprecation cycle.</p>
</div>
<div class="admonition note">
<p class="first admonition-title">Note</p>
<p class="last">This is considered an advanced feature. For most cases the built-in
collections are probably sufficient.</p>
</div>
<p>Before reading this you should read and underestand:</p>
<ul class="simple">
<li><a class="reference internal" href="graphs.html"><span class="doc">overview</span></a></li>
<li><a class="reference internal" href="spec.html"><span class="doc">graph specification</span></a></li>
<li><a class="reference internal" href="custom-graphs.html"><span class="doc">custom graphs</span></a></li>
</ul>
<p><strong>Contents</strong></p>
<ul class="simple">
<li><a class="reference internal" href="#collection-interface"><span class="std std-ref">Description of the dask collection interface</span></a></li>
<li><a class="reference internal" href="#core-method-internals"><span class="std std-ref">How this interface is used to implement the core dask
methods</span></a></li>
<li><a class="reference internal" href="#adding-methods-to-class"><span class="std std-ref">How to add the core methods to your class</span></a></li>
<li><a class="reference internal" href="#example-dask-collection"><span class="std std-ref">Example Dask Collection</span></a></li>
<li><a class="reference internal" href="#is-dask-collection"><span class="std std-ref">How to check if something is a dask collection</span></a></li>
<li><a class="reference internal" href="#deterministic-hashing"><span class="std std-ref">How to make tokenize work with your collection</span></a></li>
</ul>
<div class="section" id="the-dask-collection-interface">
<span id="collection-interface"></span><h2>The Dask Collection Interface<a class="headerlink" href="#the-dask-collection-interface" title="Permalink to this headline"></a></h2>
<p>To create your own dask collection, you need to fullfill the following
interface. Note that there is no required base class.</p>
<p>It’s recommended to also read <a class="reference internal" href="#core-method-internals"><span class="std std-ref">Internals of the Core Dask Methods</span></a> to see how this
interface is used inside dask.</p>
<dl class="method">
<dt id="__dask_graph__">
<code class="descname">__dask_graph__</code><span class="sig-paren">(</span><em>self</em><span class="sig-paren">)</span><a class="headerlink" href="#__dask_graph__" title="Permalink to this definition"></a></dt>
<dd><p>The dask graph.</p>
<table class="docutils field-list" frame="void" rules="none">
<col class="field-name" />
<col class="field-body" />
<tbody valign="top">
<tr class="field-odd field"><th class="field-name">Returns:</th><td class="field-body"><p class="first"><strong>dsk</strong> : MutableMapping, None</p>
<blockquote class="last">
<div><p>The dask graph. If <code class="docutils literal"><span class="pre">None</span></code>, this instance will not be interpreted as a
dask collection, and none of the remaining interface methods will be
called.</p>
</div></blockquote>
</td>
</tr>
</tbody>
</table>
</dd></dl>

<dl class="method">
<dt id="__dask_keys__">
<code class="descname">__dask_keys__</code><span class="sig-paren">(</span><em>self</em><span class="sig-paren">)</span><a class="headerlink" href="#__dask_keys__" title="Permalink to this definition"></a></dt>
<dd><p>The output keys for the dask graph.</p>
<table class="docutils field-list" frame="void" rules="none">
<col class="field-name" />
<col class="field-body" />
<tbody valign="top">
<tr class="field-odd field"><th class="field-name">Returns:</th><td class="field-body"><p class="first"><strong>keys</strong> : list</p>
<blockquote class="last">
<div><p>A possibly nested list of keys that represent the outputs of the graph.
After computation, the results will be returned in the same layout,
with the keys replaced with their corresponding outputs.</p>
</div></blockquote>
</td>
</tr>
</tbody>
</table>
</dd></dl>

<dl class="staticmethod">
<dt id="__dask_optimize__">
<em class="property">static </em><code class="descname">__dask_optimize__</code><span class="sig-paren">(</span><em>dsk</em>, <em>keys</em>, <em>**kwargs</em><span class="sig-paren">)</span><a class="headerlink" href="#__dask_optimize__" title="Permalink to this definition"></a></dt>
<dd><p>Given a graph and keys, return a new optimized graph.</p>
<p>This method can be either a <code class="docutils literal"><span class="pre">staticmethod</span></code> or a <code class="docutils literal"><span class="pre">classmethod</span></code>, but not
an instancemethod.</p>
<p>Note that graphs and keys are merged before calling <code class="docutils literal"><span class="pre">__dask_optimize__</span></code>;
as such the graph and keys passed to this method may represent more than
one collection sharing the same optimize method.</p>
<p>If not implemented, defaults to returning the graph unchanged.</p>
<table class="docutils field-list" frame="void" rules="none">
<col class="field-name" />
<col class="field-body" />
<tbody valign="top">
<tr class="field-odd field"><th class="field-name">Parameters:</th><td class="field-body"><p class="first"><strong>dsk</strong> : MutableMapping</p>
<blockquote>
<div><p>The merged graphs from all collections sharing the same
<code class="docutils literal"><span class="pre">__dask_optimize__</span></code> method.</p>
</div></blockquote>
<p><strong>keys</strong> : list</p>
<blockquote>
<div><p>A list of the outputs from <code class="docutils literal"><span class="pre">__dask_keys__</span></code> from all collections
sharing the same <code class="docutils literal"><span class="pre">__dask_optimize__</span></code> method.</p>
</div></blockquote>
<p><strong>**kwargs</strong></p>
<blockquote>
<div><p>Extra keyword arguments forwarded from the call to <code class="docutils literal"><span class="pre">compute</span></code> or
<code class="docutils literal"><span class="pre">persist</span></code>. Can be used or ignored as needed.</p>
</div></blockquote>
</td>
</tr>
<tr class="field-even field"><th class="field-name">Returns:</th><td class="field-body"><p class="first"><strong>optimized_dsk</strong> : MutableMapping</p>
<blockquote class="last">
<div><p>The optimized dask graph.</p>
</div></blockquote>
</td>
</tr>
</tbody>
</table>
</dd></dl>

<dl class="staticmethod">
<dt id="__dask_scheduler__">
<em class="property">static </em><code class="descname">__dask_scheduler__</code><span class="sig-paren">(</span><em>dsk</em>, <em>keys</em>, <em>**kwargs</em><span class="sig-paren">)</span><a class="headerlink" href="#__dask_scheduler__" title="Permalink to this definition"></a></dt>
<dd><p>The default scheduler <code class="docutils literal"><span class="pre">get</span></code> to use for this object.</p>
<p>Usually attached to the class as a staticmethod, e.g.</p>
<div class="highlight-default"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="kn">import</span> <span class="nn">dask.threaded</span>
<span class="gp">&gt;&gt;&gt; </span><span class="k">class</span> <span class="nc">MyCollection</span><span class="p">(</span><span class="nb">object</span><span class="p">):</span>
<span class="gp">... </span>    <span class="c1"># Use the threaded scheduler by default</span>
<span class="gp">... </span>    <span class="n">__dask_scheduler__</span> <span class="o">=</span> <span class="nb">staticmethod</span><span class="p">(</span><span class="n">dask</span><span class="o">.</span><span class="n">threaded</span><span class="o">.</span><span class="n">get</span><span class="p">)</span>
</pre></div>
</div>
</dd></dl>

<dl class="method">
<dt id="__dask_postcompute__">
<code class="descname">__dask_postcompute__</code><span class="sig-paren">(</span><em>self</em><span class="sig-paren">)</span><a class="headerlink" href="#__dask_postcompute__" title="Permalink to this definition"></a></dt>
<dd><p>Return the finalizer and (optional) extra arguments to convert the computed
results into their in-memory representation.</p>
<p>Used to implement <code class="docutils literal"><span class="pre">dask.compute</span></code>.</p>
<table class="docutils field-list" frame="void" rules="none">
<col class="field-name" />
<col class="field-body" />
<tbody valign="top">
<tr class="field-odd field"><th class="field-name">Returns:</th><td class="field-body"><p class="first"><strong>finalize</strong> : callable</p>
<blockquote>
<div><p>A function with the signature <code class="docutils literal"><span class="pre">finalize(results,</span> <span class="pre">*extra_args)</span></code>.
Called with the computed results in the same structure as the
corresponding keys from <code class="docutils literal"><span class="pre">__dask_keys__</span></code>, as well as any extra
arguments as specified in <code class="docutils literal"><span class="pre">extra_args</span></code>.  Should perform any necessary
finalization before returning the corresponding in-memory collection
from <code class="docutils literal"><span class="pre">compute</span></code>. For example, the <code class="docutils literal"><span class="pre">finalize</span></code> function for
<code class="docutils literal"><span class="pre">dask.array.Array</span></code> concatenates all the individual array chunks into
one large numpy array, which is then the result of <code class="docutils literal"><span class="pre">compute</span></code>.</p>
</div></blockquote>
<p><strong>extra_args</strong> : tuple</p>
<blockquote class="last">
<div><p>Any extra arguments to pass to <code class="docutils literal"><span class="pre">finalize</span></code> after <code class="docutils literal"><span class="pre">results</span></code>. If no
extra arguments should be an empty tuple.</p>
</div></blockquote>
</td>
</tr>
</tbody>
</table>
</dd></dl>

<dl class="method">
<dt id="__dask_postpersist__">
<code class="descname">__dask_postpersist__</code><span class="sig-paren">(</span><em>self</em><span class="sig-paren">)</span><a class="headerlink" href="#__dask_postpersist__" title="Permalink to this definition"></a></dt>
<dd><p>Return the rebuilder and (optional) extra arguments to rebuild an equivalent
dask collection from a persisted graph.</p>
<p>Used to implement <code class="docutils literal"><span class="pre">dask.persist</span></code>.</p>
<table class="docutils field-list" frame="void" rules="none">
<col class="field-name" />
<col class="field-body" />
<tbody valign="top">
<tr class="field-odd field"><th class="field-name">Returns:</th><td class="field-body"><p class="first"><strong>rebuild</strong> : callable</p>
<blockquote>
<div><p>A function with the signature <code class="docutils literal"><span class="pre">rebuild(dsk,</span> <span class="pre">*extra_args)</span></code>. Called
with a persisted graph containing only the keys and results from
<code class="docutils literal"><span class="pre">__dask_keys__</span></code>, as well as any extra arguments as specified in
<code class="docutils literal"><span class="pre">extra_args</span></code>. Should return an equivalent dask collection with the
same keys as <code class="docutils literal"><span class="pre">self</span></code>, but with the results already computed. For
example, the <code class="docutils literal"><span class="pre">rebuild</span></code> function for <code class="docutils literal"><span class="pre">dask.array.Array</span></code> is just the
<code class="docutils literal"><span class="pre">__init__</span></code> method called with the new graph but the same metadata.</p>
</div></blockquote>
<p><strong>extra_args</strong> : tuple</p>
<blockquote class="last">
<div><p>Any extra arguments to pass to <code class="docutils literal"><span class="pre">rebuild</span></code> after <code class="docutils literal"><span class="pre">dsk</span></code>. If no extra
arguments should be an empty tuple.</p>
</div></blockquote>
</td>
</tr>
</tbody>
</table>
</dd></dl>

<div class="admonition note">
<p class="first admonition-title">Note</p>
<p class="last">It’s also recommended to define <code class="docutils literal"><span class="pre">__dask_tokenize__</span></code>,
see <a class="reference internal" href="#deterministic-hashing"><span class="std std-ref">Implementing Deterministic Hashing</span></a>.</p>
</div>
</div>
<div class="section" id="internals-of-the-core-dask-methods">
<span id="core-method-internals"></span><h2>Internals of the Core Dask Methods<a class="headerlink" href="#internals-of-the-core-dask-methods" title="Permalink to this headline"></a></h2>
<p>Dask has a few <em>core</em> functions (and corresponding methods) that implement
common operations:</p>
<ul class="simple">
<li><code class="docutils literal"><span class="pre">compute</span></code>: convert one or more dask collections into their in-memory
counterparts</li>
<li><code class="docutils literal"><span class="pre">persist</span></code>: convert one or more dask collections into equivalent dask
collections with their results already computed and cached in memory.</li>
<li><code class="docutils literal"><span class="pre">visualize</span></code>: given one or more dask collections, draw out the graph that
would be passed to the scheduler during a call to <code class="docutils literal"><span class="pre">compute</span></code> or <code class="docutils literal"><span class="pre">persist</span></code></li>
</ul>
<p>Here we briefly describe the internals of these functions to illustrate how
they relate to the above interface.</p>
<div class="section" id="compute">
<h3>Compute<a class="headerlink" href="#compute" title="Permalink to this headline"></a></h3>
<p>The operation of <code class="docutils literal"><span class="pre">compute</span></code> can be broken into three stages:</p>
<ol class="arabic">
<li><p class="first"><strong>Graph Merging &amp; Optimization</strong></p>
<p>First the individual collections are converted to a single large graph and
nested list of keys. How this happens depends on the value of the
<code class="docutils literal"><span class="pre">optimize_graph</span></code> keyword, which each function takes:</p>
<ul class="simple">
<li>If <code class="docutils literal"><span class="pre">optimize_graph</span></code> is <code class="docutils literal"><span class="pre">True</span></code> (default) then the collections are first
grouped by their <code class="docutils literal"><span class="pre">__dask_optimize__</span></code> methods. All collections with the
same <code class="docutils literal"><span class="pre">__dask_optimize__</span></code> method have their graphs merged and keys
concatenated, and then a single call to each respective
<code class="docutils literal"><span class="pre">__dask_optimize__</span></code> is made with the merged graphs and keys. The
resulting graphs are then merged.</li>
<li>If <code class="docutils literal"><span class="pre">optimize_graph</span></code> is <code class="docutils literal"><span class="pre">False</span></code> then all the graphs are merged and all
the keys concatenated.</li>
</ul>
<p>After this stage there is a single large graph and nested list of keys which
represents all the collections.</p>
</li>
<li><p class="first"><strong>Computation</strong></p>
<p>After the graphs are merged and any optimizations performed, the resulting
large graph and nested list of keys are passed on to the scheduler. The
scheduler to use is chosen as follows:</p>
<ul class="simple">
<li>If a <code class="docutils literal"><span class="pre">get</span></code> function is specified directly as a keyword, use that.</li>
<li>Otherwise, if a global scheduler is set, use that.</li>
<li>Otherwise fall back to the default scheduler for the given collections.
Note that if all collections don’t share the same <code class="docutils literal"><span class="pre">__dask_scheduler__</span></code>
then an error will be raised.</li>
</ul>
<p>Once the appropriate scheduler <code class="docutils literal"><span class="pre">get</span></code> function is determined, it’s called
with the merged graph, keys, and extra keyword arguments. After this stage
<code class="docutils literal"><span class="pre">results</span></code> is a nested list of values. The structure of this list mirrors
that of <code class="docutils literal"><span class="pre">keys</span></code>, with each key substituted with its corresponding result.</p>
</li>
<li><p class="first"><strong>Postcompute</strong></p>
<p>After the results are generated the output values of <code class="docutils literal"><span class="pre">compute</span></code> need to be
built. This is what the <code class="docutils literal"><span class="pre">__dask_postcompute__</span></code> method is for.
<code class="docutils literal"><span class="pre">__dask_postcompute__</span></code> returns two things:</p>
<ul class="simple">
<li>A <code class="docutils literal"><span class="pre">finalize</span></code> function, which takes in the results for the corresponding
keys</li>
<li>A tuple of extra arguments to pass to <code class="docutils literal"><span class="pre">finalize</span></code> after the results</li>
</ul>
<p>To build the outputs, the list of collections and results is iterated over,
and the finalizer for each collection is called on its respective results.</p>
</li>
</ol>
<p>In pseudocode this process looks like:</p>
<div class="code python highlight-default"><div class="highlight"><pre><span></span><span class="k">def</span> <span class="nf">compute</span><span class="p">(</span><span class="o">*</span><span class="n">collections</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
    <span class="c1"># 1. Graph Merging &amp; Optimization</span>
    <span class="c1"># -------------------------------</span>
    <span class="k">if</span> <span class="n">kwargs</span><span class="o">.</span><span class="n">pop</span><span class="p">(</span><span class="s1">&#39;optimize_graph&#39;</span><span class="p">,</span> <span class="kc">True</span><span class="p">):</span>
        <span class="c1"># If optimization is turned on, group the collections by</span>
        <span class="c1"># optimization method, and apply each method only once to the merged</span>
        <span class="c1"># sub-graphs.</span>
        <span class="n">optimization_groups</span> <span class="o">=</span> <span class="n">groupby_optimization_methods</span><span class="p">(</span><span class="n">collections</span><span class="p">)</span>
        <span class="n">graphs</span> <span class="o">=</span> <span class="p">[]</span>
        <span class="k">for</span> <span class="n">optimize_method</span><span class="p">,</span> <span class="n">cols</span> <span class="ow">in</span> <span class="n">optimization_groups</span><span class="p">:</span>
            <span class="c1"># Merge the graphs and keys for the subset of collections that</span>
            <span class="c1"># share this optimization method</span>
            <span class="n">sub_graph</span> <span class="o">=</span> <span class="n">merge_graphs</span><span class="p">([</span><span class="n">x</span><span class="o">.</span><span class="n">__dask_graph__</span><span class="p">()</span> <span class="k">for</span> <span class="n">x</span> <span class="ow">in</span> <span class="n">cols</span><span class="p">])</span>
            <span class="n">sub_keys</span> <span class="o">=</span> <span class="p">[</span><span class="n">x</span><span class="o">.</span><span class="n">__dask_keys__</span><span class="p">()</span> <span class="k">for</span> <span class="n">x</span> <span class="ow">in</span> <span class="n">cols</span><span class="p">]</span>
            <span class="c1"># kwargs are forwarded to ``__dask_optimize__`` from compute</span>
            <span class="n">optimized_graph</span> <span class="o">=</span> <span class="n">optimize_method</span><span class="p">(</span><span class="n">sub_graph</span><span class="p">,</span> <span class="n">sub_keys</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span>
            <span class="n">graphs</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">optimized_graph</span><span class="p">)</span>
        <span class="n">graph</span> <span class="o">=</span> <span class="n">merge_graphs</span><span class="p">(</span><span class="n">graphs</span><span class="p">)</span>
    <span class="k">else</span><span class="p">:</span>
        <span class="n">graph</span> <span class="o">=</span> <span class="n">merge_graphs</span><span class="p">([</span><span class="n">x</span><span class="o">.</span><span class="n">__dask_graph__</span><span class="p">()</span> <span class="k">for</span> <span class="n">x</span> <span class="ow">in</span> <span class="n">collections</span><span class="p">])</span>
    <span class="c1"># Keys are always the same</span>
    <span class="n">keys</span> <span class="o">=</span> <span class="p">[</span><span class="n">x</span><span class="o">.</span><span class="n">__dask_keys__</span><span class="p">()</span> <span class="k">for</span> <span class="n">x</span> <span class="ow">in</span> <span class="n">collections</span><span class="p">]</span>

    <span class="c1"># 2. Computation</span>
    <span class="c1"># --------------</span>
    <span class="c1"># Determine appropriate get function based on collections, global</span>
    <span class="c1"># settings, and keyword arguments</span>
    <span class="n">get</span> <span class="o">=</span> <span class="n">determine_get_function</span><span class="p">(</span><span class="n">collections</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span>
    <span class="c1"># Pass the merged graph, keys, and kwargs to ``get``</span>
    <span class="n">results</span> <span class="o">=</span> <span class="n">get</span><span class="p">(</span><span class="n">graph</span><span class="p">,</span> <span class="n">keys</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span>

    <span class="c1"># 3. Postcompute</span>
    <span class="c1"># --------------</span>
    <span class="n">output</span> <span class="o">=</span> <span class="p">[]</span>
    <span class="c1"># Iterate over the results and collections</span>
    <span class="k">for</span> <span class="n">res</span><span class="p">,</span> <span class="n">collection</span> <span class="ow">in</span> <span class="nb">zip</span><span class="p">(</span><span class="n">results</span><span class="p">,</span> <span class="n">collections</span><span class="p">):</span>
        <span class="n">finalize</span><span class="p">,</span> <span class="n">extra_args</span> <span class="o">=</span> <span class="n">collection</span><span class="o">.</span><span class="n">__dask_postcompute__</span><span class="p">()</span>
        <span class="n">out</span> <span class="o">=</span> <span class="n">finalize</span><span class="p">(</span><span class="n">res</span><span class="p">,</span> <span class="o">**</span><span class="n">extra_args</span><span class="p">)</span>
        <span class="n">output</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">out</span><span class="p">)</span>

    <span class="c1"># `dask.compute` always returns tuples</span>
    <span class="k">return</span> <span class="nb">tuple</span><span class="p">(</span><span class="n">output</span><span class="p">)</span>
</pre></div>
</div>
</div>
<div class="section" id="persist">
<h3>Persist<a class="headerlink" href="#persist" title="Permalink to this headline"></a></h3>
<p>Persist is very similar to <code class="docutils literal"><span class="pre">compute</span></code>, except for how the return values are
created. It too has three stages:</p>
<ol class="arabic">
<li><p class="first"><strong>Graph Merging &amp; Optimization</strong></p>
<p>Same as in <code class="docutils literal"><span class="pre">compute</span></code>.</p>
</li>
<li><p class="first"><strong>Computation</strong></p>
<p>Same as in <code class="docutils literal"><span class="pre">compute</span></code>, except in the case of the distributed scheduler,
where the values in <code class="docutils literal"><span class="pre">results</span></code> are futures instead of values.</p>
</li>
<li><p class="first"><strong>Postpersist</strong></p>
<p>Similar to <code class="docutils literal"><span class="pre">__dask_postcompute__</span></code>, <code class="docutils literal"><span class="pre">__dask_postpersist__</span></code> is used to
rebuild values in a call to <code class="docutils literal"><span class="pre">persist</span></code>. <code class="docutils literal"><span class="pre">__dask_postpersist__</span></code> returns
two things:</p>
<ul class="simple">
<li>A <code class="docutils literal"><span class="pre">rebuild</span></code> function, which takes in a persisted graph. The keys of
this graph are the same as <code class="docutils literal"><span class="pre">__dask_keys__</span></code> for the corresponding
collection, and the values are computed results (for the single machine
scheduler) or futures (for the distributed scheduler).</li>
<li>A tuple of extra arguments to pass to <code class="docutils literal"><span class="pre">rebuild</span></code> after the graph</li>
</ul>
<p>To build the outputs of <code class="docutils literal"><span class="pre">persist</span></code>, the list of collections and results is
iterated over, and the rebuilder for each collection is called on the graph
for its respective results.</p>
</li>
</ol>
<p>In pseudocode this looks like:</p>
<div class="code python highlight-default"><div class="highlight"><pre><span></span><span class="k">def</span> <span class="nf">persist</span><span class="p">(</span><span class="o">*</span><span class="n">collections</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
    <span class="c1"># 1. Graph Merging &amp; Optimization</span>
    <span class="c1"># -------------------------------</span>
    <span class="c1"># **Same as in compute**</span>
    <span class="n">graph</span> <span class="o">=</span> <span class="o">...</span>
    <span class="n">keys</span> <span class="o">=</span> <span class="o">...</span>

    <span class="c1"># 2. Computation</span>
    <span class="c1"># --------------</span>
    <span class="c1"># **Same as in compute**</span>
    <span class="n">results</span> <span class="o">=</span> <span class="o">...</span>

    <span class="c1"># 3. Postpersist</span>
    <span class="c1"># --------------</span>
    <span class="n">output</span> <span class="o">=</span> <span class="p">[]</span>
    <span class="c1"># Iterate over the results and collections</span>
    <span class="k">for</span> <span class="n">res</span><span class="p">,</span> <span class="n">collection</span> <span class="ow">in</span> <span class="nb">zip</span><span class="p">(</span><span class="n">results</span><span class="p">,</span> <span class="n">collections</span><span class="p">):</span>
        <span class="c1"># res has the same structure as keys</span>
        <span class="n">keys</span> <span class="o">=</span> <span class="n">collection</span><span class="o">.</span><span class="n">__dask_keys__</span><span class="p">()</span>
        <span class="c1"># Get the computed graph for this collection.</span>
        <span class="c1"># Here flatten converts a nested list into a single list</span>
        <span class="n">graph</span> <span class="o">=</span> <span class="p">{</span><span class="n">k</span><span class="p">:</span> <span class="n">r</span> <span class="k">for</span> <span class="p">(</span><span class="n">k</span><span class="p">,</span> <span class="n">r</span><span class="p">)</span> <span class="ow">in</span> <span class="nb">zip</span><span class="p">(</span><span class="n">flatten</span><span class="p">(</span><span class="n">keys</span><span class="p">),</span> <span class="n">flatten</span><span class="p">(</span><span class="n">res</span><span class="p">))}</span>

        <span class="c1"># Rebuild the output dask collection with the computed graph</span>
        <span class="n">rebuild</span><span class="p">,</span> <span class="n">extra_args</span> <span class="o">=</span> <span class="n">collection</span><span class="o">.</span><span class="n">__dask_postpersist__</span><span class="p">()</span>
        <span class="n">out</span> <span class="o">=</span> <span class="n">rebuild</span><span class="p">(</span><span class="n">graph</span><span class="p">,</span> <span class="o">*</span><span class="n">extra_args</span><span class="p">)</span>

        <span class="n">output</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">out</span><span class="p">)</span>

    <span class="c1"># dask.persist always returns tuples</span>
    <span class="k">return</span> <span class="nb">tuple</span><span class="p">(</span><span class="n">output</span><span class="p">)</span>
</pre></div>
</div>
</div>
<div class="section" id="visualize">
<h3>Visualize<a class="headerlink" href="#visualize" title="Permalink to this headline"></a></h3>
<p>Visualize is the simplest of the 3 core methods. It only has two stages:</p>
<ol class="arabic">
<li><p class="first"><strong>Graph Merging &amp; Optimization</strong></p>
<p>Same as in <code class="docutils literal"><span class="pre">compute</span></code></p>
</li>
<li><p class="first"><strong>Graph Drawing</strong></p>
<p>The resulting merged graph is drawn using <code class="docutils literal"><span class="pre">graphviz</span></code> and output to the
specified file.</p>
</li>
</ol>
<p>In pseudocode this looks like:</p>
<div class="code python highlight-default"><div class="highlight"><pre><span></span><span class="k">def</span> <span class="nf">visualize</span><span class="p">(</span><span class="o">*</span><span class="n">collections</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
    <span class="c1"># 1. Graph Merging &amp; Optimization</span>
    <span class="c1"># -------------------------------</span>
    <span class="c1"># **Same as in compute**</span>
    <span class="n">graph</span> <span class="o">=</span> <span class="o">...</span>
    <span class="n">keys</span> <span class="o">=</span> <span class="o">...</span>

    <span class="c1"># 2. Graph Drawing</span>
    <span class="c1"># ----------------</span>
    <span class="c1"># Draw the graph with graphviz&#39;s `dot` tool and return the result.</span>
    <span class="k">return</span> <span class="n">dot_graph</span><span class="p">(</span><span class="n">graph</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span>
</pre></div>
</div>
</div>
</div>
<div class="section" id="adding-the-core-dask-methods-to-your-class">
<span id="adding-methods-to-class"></span><h2>Adding the Core Dask Methods to Your Class<a class="headerlink" href="#adding-the-core-dask-methods-to-your-class" title="Permalink to this headline"></a></h2>
<p>Defining the above interface will allow your object to used by the core dask
functions (<code class="docutils literal"><span class="pre">dask.compute</span></code>, <code class="docutils literal"><span class="pre">dask.persist</span></code>, <code class="docutils literal"><span class="pre">dask.visualize</span></code>, etc…). To
add corresponding method versions of these subclass from
<code class="docutils literal"><span class="pre">dask.base.DaskMethodsMixin</span></code>, which adds implementations of <code class="docutils literal"><span class="pre">compute</span></code>,
<code class="docutils literal"><span class="pre">persist</span></code>, and <code class="docutils literal"><span class="pre">visualize</span></code> based on the interface above.</p>
</div>
<div class="section" id="example-dask-collection">
<span id="id1"></span><h2>Example Dask Collection<a class="headerlink" href="#example-dask-collection" title="Permalink to this headline"></a></h2>
<p>Here we create a dask collection representing a tuple. Every element in the
tuple is represented as a task in the graph. Note that this is for illustration
purposes only - the same user experience could be done using normal tuples with
elements of <code class="docutils literal"><span class="pre">dask.delayed</span></code>.</p>
<div class="code python highlight-default"><div class="highlight"><pre><span></span><span class="c1"># Saved as dask_tuple.py</span>
<span class="kn">from</span> <span class="nn">dask.base</span> <span class="k">import</span> <span class="n">DaskMethodsMixin</span>
<span class="kn">from</span> <span class="nn">dask.optimize</span> <span class="k">import</span> <span class="n">cull</span>

<span class="c1"># We subclass from DaskMethodsMixin to add common dask methods to our</span>
<span class="c1"># class. This is nice but not necessary for creating a dask collection.</span>
<span class="k">class</span> <span class="nc">Tuple</span><span class="p">(</span><span class="n">DaskMethodsMixin</span><span class="p">):</span>
    <span class="k">def</span> <span class="nf">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">dsk</span><span class="p">,</span> <span class="n">keys</span><span class="p">):</span>
        <span class="c1"># The init method takes in a dask graph and a set of keys to use</span>
        <span class="c1"># as outputs.</span>
        <span class="bp">self</span><span class="o">.</span><span class="n">_dsk</span> <span class="o">=</span> <span class="n">dsk</span>
        <span class="bp">self</span><span class="o">.</span><span class="n">_keys</span> <span class="o">=</span> <span class="n">keys</span>

    <span class="k">def</span> <span class="nf">__dask_graph__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
        <span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">_dsk</span>

    <span class="k">def</span> <span class="nf">__dask_keys__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
        <span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">_keys</span>

    <span class="nd">@staticmethod</span>
    <span class="k">def</span> <span class="nf">__dask_optimize__</span><span class="p">(</span><span class="n">dsk</span><span class="p">,</span> <span class="n">keys</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
        <span class="c1"># We cull unnecessary tasks here. Note that this isn&#39;t necessary,</span>
        <span class="c1"># dask will do this automatically, this just shows one optimization</span>
        <span class="c1"># you could do.</span>
        <span class="n">dsk2</span><span class="p">,</span> <span class="n">_</span> <span class="o">=</span> <span class="n">cull</span><span class="p">(</span><span class="n">dsk</span><span class="p">,</span> <span class="n">keys</span><span class="p">)</span>
        <span class="k">return</span> <span class="n">dsk2</span>

    <span class="c1"># Use the threaded scheduler by default.</span>
    <span class="n">__dask_scheduler__</span> <span class="o">=</span> <span class="nb">staticmethod</span><span class="p">(</span><span class="n">dask</span><span class="o">.</span><span class="n">threaded</span><span class="o">.</span><span class="n">get</span><span class="p">)</span>

    <span class="k">def</span> <span class="nf">__dask_postcompute__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
        <span class="c1"># We want to return the results as a tuple, so our finalize</span>
        <span class="c1"># function is `tuple`. There are no extra arguments, so we also</span>
        <span class="c1"># return an empty tuple.</span>
        <span class="k">return</span> <span class="nb">tuple</span><span class="p">,</span> <span class="p">()</span>

    <span class="k">def</span> <span class="nf">__dask_postpersist__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
        <span class="c1"># Since our __init__ takes a graph as its first argument, our</span>
        <span class="c1"># rebuild function can just be the class itself. For extra</span>
        <span class="c1"># arguments we also return a tuple containing just the keys.</span>
        <span class="k">return</span> <span class="n">Tuple</span><span class="p">,</span> <span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">_keys</span><span class="p">,)</span>

    <span class="k">def</span> <span class="nf">__dask_tokenize__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
        <span class="c1"># For tokenize to work we want to return a value that fully</span>
        <span class="c1"># represents this object. In this case it&#39;s the list of keys</span>
        <span class="c1"># to be computed.</span>
        <span class="k">return</span> <span class="nb">tuple</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">_keys</span><span class="p">)</span>
</pre></div>
</div>
<p>Demonstrating this class:</p>
<div class="code python highlight-default"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="kn">from</span> <span class="nn">dask_tuple</span> <span class="k">import</span> <span class="n">Tuple</span>
<span class="gp">&gt;&gt;&gt; </span><span class="kn">from</span> <span class="nn">operator</span> <span class="k">import</span> <span class="n">add</span><span class="p">,</span> <span class="n">mul</span>

<span class="go"># Define a dask graph</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">dsk</span> <span class="o">=</span> <span class="p">{</span><span class="s1">&#39;a&#39;</span><span class="p">:</span> <span class="mi">1</span><span class="p">,</span>
<span class="gp">... </span>       <span class="s1">&#39;b&#39;</span><span class="p">:</span> <span class="mi">2</span><span class="p">,</span>
<span class="gp">... </span>       <span class="s1">&#39;c&#39;</span><span class="p">:</span> <span class="p">(</span><span class="n">add</span><span class="p">,</span> <span class="s1">&#39;a&#39;</span><span class="p">,</span> <span class="s1">&#39;b&#39;</span><span class="p">),</span>
<span class="gp">... </span>       <span class="s1">&#39;d&#39;</span><span class="p">:</span> <span class="p">(</span><span class="n">mul</span><span class="p">,</span> <span class="s1">&#39;b&#39;</span><span class="p">,</span> <span class="mi">2</span><span class="p">),</span>
<span class="gp">... </span>       <span class="s1">&#39;e&#39;</span><span class="p">:</span> <span class="p">(</span><span class="n">add</span><span class="p">,</span> <span class="s1">&#39;b&#39;</span><span class="p">,</span> <span class="s1">&#39;c&#39;</span><span class="p">)}</span>

<span class="go"># The output keys for this graph</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">keys</span> <span class="o">=</span> <span class="p">[</span><span class="s1">&#39;b&#39;</span><span class="p">,</span> <span class="s1">&#39;c&#39;</span><span class="p">,</span> <span class="s1">&#39;d&#39;</span><span class="p">,</span> <span class="s1">&#39;e&#39;</span><span class="p">]</span>

<span class="gp">&gt;&gt;&gt; </span><span class="n">x</span> <span class="o">=</span> <span class="n">Tuple</span><span class="p">(</span><span class="n">dsk</span><span class="p">,</span> <span class="n">keys</span><span class="p">)</span>

<span class="go"># Compute turns Tuple into a tuple</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">x</span><span class="o">.</span><span class="n">compute</span><span class="p">()</span>
<span class="go">(2, 3, 4, 5)</span>

<span class="go"># Persist turns Tuple into a Tuple, with each task already computed</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">x2</span> <span class="o">=</span> <span class="n">x</span><span class="o">.</span><span class="n">persist</span><span class="p">()</span>
<span class="gp">&gt;&gt;&gt; </span><span class="nb">isinstance</span><span class="p">(</span><span class="n">x2</span><span class="p">,</span> <span class="n">Tuple</span><span class="p">)</span>
<span class="go">True</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">x2</span><span class="o">.</span><span class="n">__dask_graph__</span><span class="p">()</span>
<span class="go">{&#39;b&#39;: 2,</span>
<span class="go"> &#39;c&#39;: 3,</span>
<span class="go"> &#39;d&#39;: 4,</span>
<span class="go"> &#39;e&#39;: 5}</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">x2</span><span class="o">.</span><span class="n">compute</span><span class="p">()</span>
<span class="go">(2, 3, 4, 5)</span>
</pre></div>
</div>
</div>
<div class="section" id="checking-if-an-object-is-a-dask-collection">
<span id="is-dask-collection"></span><h2>Checking if an object is a dask collection<a class="headerlink" href="#checking-if-an-object-is-a-dask-collection" title="Permalink to this headline"></a></h2>
<p>To check if an object is a dask collection, use
<code class="docutils literal"><span class="pre">dask.base.is_dask_collection</span></code>:</p>
<div class="code python highlight-default"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="kn">from</span> <span class="nn">dask.base</span> <span class="k">import</span> <span class="n">is_dask_collection</span>
<span class="gp">&gt;&gt;&gt; </span><span class="kn">from</span> <span class="nn">dask</span> <span class="k">import</span> <span class="n">delayed</span>

<span class="gp">&gt;&gt;&gt; </span><span class="n">x</span> <span class="o">=</span> <span class="n">delayed</span><span class="p">(</span><span class="nb">sum</span><span class="p">)([</span><span class="mi">1</span><span class="p">,</span> <span class="mi">2</span><span class="p">,</span> <span class="mi">3</span><span class="p">])</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">is_dask_collection</span><span class="p">(</span><span class="n">x</span><span class="p">)</span>
<span class="go">True</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">is_dask_collection</span><span class="p">(</span><span class="mi">1</span><span class="p">)</span>
<span class="go">False</span>
</pre></div>
</div>
</div>
<div class="section" id="implementing-deterministic-hashing">
<span id="deterministic-hashing"></span><h2>Implementing Deterministic Hashing<a class="headerlink" href="#implementing-deterministic-hashing" title="Permalink to this headline"></a></h2>
<p>Dask implements its own deterministic hash function to generate keys based on
the value of arguments. This function is available as <code class="docutils literal"><span class="pre">dask.base.tokenize</span></code>.
Many common types already have implementations of <code class="docutils literal"><span class="pre">tokenize</span></code>, which can be
found in <code class="docutils literal"><span class="pre">dask/base.py</span></code>.</p>
<p>When creating your own custom classes you may need to register a <code class="docutils literal"><span class="pre">tokenize</span></code>
implementation. There are two ways to do this:</p>
<div class="admonition note">
<p class="first admonition-title">Note</p>
<p class="last">Both dask collections and normal python objects can have
implementations of <code class="docutils literal"><span class="pre">tokenize</span></code> using either of the methods
described below.</p>
</div>
<ol class="arabic">
<li><p class="first">The <code class="docutils literal"><span class="pre">__dask_tokenize__</span></code> method</p>
<p>Where possible, it’s recommended to define the <code class="docutils literal"><span class="pre">__dask_tokenize__</span></code> method.
This method takes no arguments and should return a value fully
representative of the object.</p>
</li>
<li><p class="first">Register a function with <code class="docutils literal"><span class="pre">dask.base.normalize_token</span></code></p>
<p>If defining a method on the class isn’t possible, you can register a tokenize
function with the <code class="docutils literal"><span class="pre">normalize_token</span></code> dispatch. The function should have the
same signature as described above.</p>
</li>
</ol>
<p>In both cases the implementation should be the same, only the location of the
definition is different.</p>
<div class="section" id="example">
<h3>Example<a class="headerlink" href="#example" title="Permalink to this headline"></a></h3>
<div class="code python highlight-default"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="kn">from</span> <span class="nn">dask.base</span> <span class="k">import</span> <span class="n">tokenize</span><span class="p">,</span> <span class="n">normalize_token</span>

<span class="go"># Define a tokenize implementation using a method.</span>
<span class="gp">&gt;&gt;&gt; </span><span class="k">class</span> <span class="nc">Foo</span><span class="p">(</span><span class="nb">object</span><span class="p">):</span>
<span class="gp">... </span>    <span class="k">def</span> <span class="nf">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">a</span><span class="p">,</span> <span class="n">b</span><span class="p">):</span>
<span class="gp">... </span>        <span class="bp">self</span><span class="o">.</span><span class="n">a</span> <span class="o">=</span> <span class="n">a</span>
<span class="gp">... </span>        <span class="bp">self</span><span class="o">.</span><span class="n">b</span> <span class="o">=</span> <span class="n">b</span>
<span class="gp">...</span>
<span class="gp">... </span>    <span class="k">def</span> <span class="nf">__dask_tokenize__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="gp">... </span>        <span class="c1"># This tuple fully represents self</span>
<span class="gp">... </span>        <span class="k">return</span> <span class="p">(</span><span class="n">Foo</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">a</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">b</span><span class="p">)</span>

<span class="gp">&gt;&gt;&gt; </span><span class="n">x</span> <span class="o">=</span> <span class="n">Foo</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="mi">2</span><span class="p">)</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">tokenize</span><span class="p">(</span><span class="n">x</span><span class="p">)</span>
<span class="go">&#39;5988362b6e07087db2bc8e7c1c8cc560&#39;</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">tokenize</span><span class="p">(</span><span class="n">x</span><span class="p">)</span> <span class="o">==</span> <span class="n">tokenize</span><span class="p">(</span><span class="n">x</span><span class="p">)</span>  <span class="c1"># token is deterministic</span>
<span class="go">True</span>

<span class="go"># Register an implementation with normalize_token</span>
<span class="gp">&gt;&gt;&gt; </span><span class="k">class</span> <span class="nc">Bar</span><span class="p">(</span><span class="nb">object</span><span class="p">):</span>
<span class="gp">... </span>    <span class="k">def</span> <span class="nf">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">x</span><span class="p">,</span> <span class="n">y</span><span class="p">):</span>
<span class="gp">... </span>        <span class="bp">self</span><span class="o">.</span><span class="n">x</span> <span class="o">=</span> <span class="n">x</span>
<span class="gp">... </span>        <span class="bp">self</span><span class="o">.</span><span class="n">y</span> <span class="o">=</span> <span class="n">y</span>

<span class="gp">&gt;&gt;&gt; </span><span class="nd">@normalize_token</span><span class="o">.</span><span class="n">register</span><span class="p">(</span><span class="n">Bar</span><span class="p">)</span>
<span class="gp">... </span><span class="k">def</span> <span class="nf">tokenize_bar</span><span class="p">(</span><span class="n">x</span><span class="p">):</span>
<span class="gp">... </span>    <span class="k">return</span> <span class="p">(</span><span class="n">Bar</span><span class="p">,</span> <span class="n">x</span><span class="o">.</span><span class="n">x</span><span class="p">,</span> <span class="n">x</span><span class="o">.</span><span class="n">x</span><span class="p">)</span>

<span class="gp">&gt;&gt;&gt; </span><span class="n">y</span> <span class="o">=</span> <span class="n">Bar</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="mi">2</span><span class="p">)</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">tokenize</span><span class="p">(</span><span class="n">y</span><span class="p">)</span>
<span class="go">&#39;5a7e9c3645aa44cf13d021c14452152e&#39;</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">tokenize</span><span class="p">(</span><span class="n">y</span><span class="p">)</span> <span class="o">==</span> <span class="n">tokenize</span><span class="p">(</span><span class="n">y</span><span class="p">)</span>
<span class="go">True</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">tokenize</span><span class="p">(</span><span class="n">y</span><span class="p">)</span> <span class="o">==</span> <span class="n">tokenize</span><span class="p">(</span><span class="n">x</span><span class="p">)</span>  <span class="c1"># tokens for different objects aren&#39;t equal</span>
<span class="go">False</span>
</pre></div>
</div>
<p>For more examples please see <code class="docutils literal"><span class="pre">dask/base.py</span></code> or any of the built-in dask
collections.</p>
</div>
</div>
</div>


           </div>
           <div class="articleComments">
            
           </div>
          </div>
          <footer>
  
    <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
      
        <a href="cite.html" class="btn btn-neutral float-right" title="Citations" accesskey="n" rel="next">Next <span class="fa fa-arrow-circle-right"></span></a>
      
      
        <a href="remote-data-services.html" class="btn btn-neutral" title="Remote Data Services" accesskey="p" rel="prev"><span class="fa fa-arrow-circle-left"></span> Previous</a>
      
    </div>
  

  <hr/>

  <div role="contentinfo">
    <p>
        &copy; Copyright 2017, Anaconda.

    </p>
  </div>
  Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a <a href="https://github.com/snide/sphinx_rtd_theme">theme</a> provided by <a href="https://readthedocs.org">Read the Docs</a>. 

</footer>

        </div>
      </div>

    </section>

  </div>
  


  

    <script type="text/javascript">
        var DOCUMENTATION_OPTIONS = {
            URL_ROOT:'./',
            VERSION:'0.16.0',
            COLLAPSE_INDEX:false,
            FILE_SUFFIX:'.html',
            HAS_SOURCE:  true,
            SOURCELINK_SUFFIX: '.txt'
        };
    </script>
      <script type="text/javascript" src="_static/jquery.js"></script>
      <script type="text/javascript" src="_static/underscore.js"></script>
      <script type="text/javascript" src="_static/doctools.js"></script>
      <script type="text/javascript" src="file:///usr/share/javascript/mathjax/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>

  

  
  
    <script type="text/javascript" src="_static/js/theme.js"></script>
  

  
  
  <script type="text/javascript">
      jQuery(function () {
          SphinxRtdTheme.StickyNav.enable();
      });
  </script>
   

</body>
</html>