@recaptime-dev's working patches + fork for Phorge, a community fork of Phabricator. (Upstream dev and stable branches are at upstream/main and upstream/stable respectively.)
hq.recaptime.dev/wiki/Phorge
phorge
phabricator
In Jupyter notebooks, read strings stored in the raw as either "string" or "list<string>" more consistently
Summary:
Ref PHI1835. Generally, Jupyter notebooks in the wild may store source and markdown content as either a single string or a list of strings.
Make the renderer read these formats more consistently. In particular, this fixes rendering of code blocks stored as a single string.
This also fixes an issue where cell labels were double-rendered in diff views.
Test Plan:
Created a notebook with a code block represented on disk as a single string, rendered a diff from it.
{F7696071}
Differential Revision: https://secure.phabricator.com/D21434
+32
-48
+32
-48
src/applications/files/document/PhabricatorJupyterDocumentEngine.php
+32
-48
src/applications/files/document/PhabricatorJupyterDocumentEngine.php
···
85
85
if ($utype === $vtype) {
86
86
switch ($utype) {
87
87
case 'markdown':
88
-
$usource = idx($ucell, 'source');
89
-
if (is_array($usource)) {
90
-
$usource = implode('', $usource);
91
-
}
92
-
93
-
$vsource = idx($vcell, 'source');
94
-
if (is_array($vsource)) {
95
-
$vsource = implode('', $vsource);
96
-
}
88
+
$usource = $this->readString($ucell, 'source');
89
+
$vsource = $this->readString($vcell, 'source');
97
90
98
91
$diff = id(new PhutilProseDifferenceEngine())
99
92
->getDiff($usource, $vsource);
···
117
110
$vsource = idx($vcell, 'raw');
118
111
$udisplay = idx($ucell, 'display');
119
112
$vdisplay = idx($vcell, 'display');
120
-
$ulabel = idx($ucell, 'label');
121
-
$vlabel = idx($vcell, 'label');
122
113
123
114
$intraline_segments = ArcanistDiffUtils::generateIntralineDiff(
124
115
$usource,
···
142
133
$vdisplay,
143
134
$v_segments);
144
135
145
-
$u_content = $this->newCodeLineCell($ucell, $usource);
146
-
$v_content = $this->newCodeLineCell($vcell, $vsource);
136
+
list($u_label, $u_content) = $this->newCodeLineCell($ucell, $usource);
137
+
list($v_label, $v_content) = $this->newCodeLineCell($vcell, $vsource);
147
138
148
139
$classes = array(
149
140
'jupyter-cell-flush',
150
141
);
151
142
152
-
$u_content = $this->newJupyterCell($ulabel, $u_content, $classes);
153
-
$v_content = $this->newJupyterCell($vlabel, $v_content, $classes);
143
+
$u_content = $this->newJupyterCell($u_label, $u_content, $classes);
144
+
$v_content = $this->newJupyterCell($v_label, $v_content, $classes);
154
145
155
146
$u_content = $this->newCellContainer($u_content);
156
147
$v_content = $this->newCellContainer($v_content);
···
259
250
$hash_input = $cell['raw'];
260
251
break;
261
252
case 'markdown':
262
-
$hash_input = $cell['source'];
263
-
if (is_array($hash_input)) {
264
-
$hash_input = implode('', $cell['source']);
265
-
}
253
+
$hash_input = $this->readString($cell, 'source');
266
254
break;
267
255
default:
268
256
$hash_input = serialize($cell);
···
334
322
'be rendered as a Jupyter notebook.'));
335
323
}
336
324
337
-
338
325
$nbformat = idx($data, 'nbformat');
339
326
if (!strlen($nbformat)) {
340
327
throw new Exception(
···
376
363
foreach ($cells as $cell) {
377
364
$cell_type = idx($cell, 'cell_type');
378
365
if ($cell_type === 'markdown') {
379
-
$source = $cell['source'];
380
-
if (is_array($source)) {
381
-
$source = implode('', $source);
382
-
}
366
+
$source = $this->readString($cell, 'source');
383
367
384
368
// Attempt to split contiguous blocks of markdown into smaller
385
369
// pieces.
···
404
388
405
389
$label = $this->newCellLabel($cell);
406
390
407
-
$lines = idx($cell, 'source');
408
-
if (!is_array($lines)) {
409
-
$lines = array();
410
-
}
411
-
391
+
$lines = $this->readStringList($cell, 'source');
412
392
$content = $this->highlightLines($lines);
413
393
414
394
$count = count($lines);
···
526
506
}
527
507
528
508
private function newMarkdownCell(array $cell) {
529
-
$content = idx($cell, 'source');
530
-
if (!is_array($content)) {
531
-
$content = array();
532
-
}
509
+
$content = $this->readStringList($cell, 'source');
533
510
534
511
// TODO: This should ideally highlight as Markdown, but the "md"
535
512
// highlighter in Pygments is painfully slow and not terribly useful.
···
549
526
private function newCodeCell(array $cell) {
550
527
$label = $this->newCellLabel($cell);
551
528
552
-
$content = idx($cell, 'source');
553
-
if (!is_array($content)) {
554
-
$content = array();
555
-
}
556
-
529
+
$content = $this->readStringList($cell, 'source');
557
530
$content = $this->highlightLines($content);
558
531
559
532
$outputs = array();
···
660
633
continue;
661
634
}
662
635
663
-
$raw_data = $data[$image_format];
664
-
if (!is_array($raw_data)) {
665
-
$raw_data = array($raw_data);
666
-
}
667
-
$raw_data = implode('', $raw_data);
636
+
$raw_data = $this->readString($data, $image_format);
668
637
669
638
$content = phutil_tag(
670
639
'img',
···
695
664
break;
696
665
case 'stream':
697
666
default:
698
-
$content = idx($output, 'text');
699
-
if (!is_array($content)) {
700
-
$content = array();
701
-
}
702
-
$content = implode('', $content);
667
+
$content = $this->readString($output, 'text');
703
668
break;
704
669
}
705
670
···
759
724
760
725
public function shouldSuggestEngine(PhabricatorDocumentRef $ref) {
761
726
return true;
727
+
}
728
+
729
+
private function readString(array $src, $key) {
730
+
$list = $this->readStringList($src, $key);
731
+
return implode('', $list);
732
+
}
733
+
734
+
private function readStringList(array $src, $key) {
735
+
$list = idx($src, $key);
736
+
737
+
if (is_array($list)) {
738
+
$list = $list;
739
+
} else if (is_string($list)) {
740
+
$list = array($list);
741
+
} else {
742
+
$list = array();
743
+
}
744
+
745
+
return $list;
762
746
}
763
747
764
748
}