##// END OF EJS Templates
- Added a fallback to urllib2 for python2...
Ben Kasel -
Show More
@@ -1,552 +1,558 b''
1 1 """Various display related classes.
2 2
3 3 Authors : MinRK, gregcaporaso, dannystaple
4 4 """
5 5 from os.path import exists, isfile, splitext, abspath, join, isdir
6 6 from os import walk, sep
7 7
8 8 from IPython.core.display import DisplayObject
9 9
10 10 __all__ = ['Audio', 'IFrame', 'YouTubeVideo', 'VimeoVideo', 'ScribdDocument',
11 11 'FileLink', 'FileLinks']
12 12
13 13
14 14 class Audio(DisplayObject):
15 15 """Create an audio object.
16 16
17 17 When this object is returned by an input cell or passed to the
18 18 display function, it will result in Audio controls being displayed
19 19 in the frontend (only works in the notebook).
20 20
21 21 Parameters
22 22 ----------
23 23 data : numpy array, list, unicode, str or bytes
24 24 Can be one of
25 25
26 26 * Numpy 1d array containing the desired waveform (mono)
27 27 * Numpy 2d array containing waveforms for each channel.
28 28 Shape=(NCHAN, NSAMPLES). For the standard channel order, see
29 29 http://msdn.microsoft.com/en-us/library/windows/hardware/dn653308(v=vs.85).aspx
30 30 * List of float or integer representing the waveform (mono)
31 31 * String containing the filename
32 32 * Bytestring containing raw PCM data or
33 33 * URL pointing to a file on the web.
34 34
35 35 If the array option is used the waveform will be normalized.
36 36
37 37 If a filename or url is used the format support will be browser
38 38 dependent.
39 39 url : unicode
40 40 A URL to download the data from.
41 41 filename : unicode
42 42 Path to a local file to load the data from.
43 43 embed : boolean
44 44 Should the image data be embedded using a data URI (True) or should
45 45 the original source be referenced. Set this to True if you want the
46 46 audio to playable later with no internet connection in the notebook.
47 47
48 48 Default is `True`, unless the keyword argument `url` is set, then
49 49 default value is `False`.
50 50 rate : integer
51 51 The sampling rate of the raw data.
52 52 Only required when data parameter is being used as an array
53 53 autoplay : bool
54 54 Set to True if the audio should immediately start playing.
55 55 Default is `False`.
56 56
57 57 Examples
58 58 --------
59 59 ::
60 60
61 61 # Generate a sound
62 62 import numpy as np
63 63 framerate = 44100
64 64 t = np.linspace(0,5,framerate*5)
65 65 data = np.sin(2*np.pi*220*t) + np.sin(2*np.pi*224*t))
66 66 Audio(data,rate=framerate)
67 67
68 68 # Can also do stereo or more channels
69 69 dataleft = np.sin(2*np.pi*220*t)
70 70 dataright = np.sin(2*np.pi*224*t)
71 71 Audio([dataleft, dataright],rate=framerate)
72 72
73 73 Audio("http://www.nch.com.au/acm/8k16bitpcm.wav") # From URL
74 74 Audio(url="http://www.w3schools.com/html/horse.ogg")
75 75
76 76 Audio('/path/to/sound.wav') # From file
77 77 Audio(filename='/path/to/sound.ogg')
78 78
79 79 Audio(b'RAW_WAV_DATA..) # From bytes
80 80 Audio(data=b'RAW_WAV_DATA..)
81 81
82 82 """
83 83 _read_flags = 'rb'
84 84
85 85 def __init__(self, data=None, filename=None, url=None, embed=None, rate=None, autoplay=False):
86 86 if filename is None and url is None and data is None:
87 87 raise ValueError("No image data found. Expecting filename, url, or data.")
88 88 if embed is False and url is None:
89 89 raise ValueError("No url found. Expecting url when embed=False")
90 90
91 91 if url is not None and embed is not True:
92 92 self.embed = False
93 93 else:
94 94 self.embed = True
95 95 self.autoplay = autoplay
96 96 super(Audio, self).__init__(data=data, url=url, filename=filename)
97 97
98 98 if self.data is not None and not isinstance(self.data, bytes):
99 99 self.data = self._make_wav(data,rate)
100 100
101 101 def reload(self):
102 102 """Reload the raw data from file or URL."""
103 103 import mimetypes
104 104 if self.embed:
105 105 super(Audio, self).reload()
106 106
107 107 if self.filename is not None:
108 108 self.mimetype = mimetypes.guess_type(self.filename)[0]
109 109 elif self.url is not None:
110 110 self.mimetype = mimetypes.guess_type(self.url)[0]
111 111 else:
112 112 self.mimetype = "audio/wav"
113 113
114 114 def _make_wav(self, data, rate):
115 115 """ Transform a numpy array to a PCM bytestring """
116 116 import struct
117 117 from io import BytesIO
118 118 import wave
119 119
120 120 try:
121 121 import numpy as np
122 122
123 123 data = np.array(data, dtype=float)
124 124 if len(data.shape) == 1:
125 125 nchan = 1
126 126 elif len(data.shape) == 2:
127 127 # In wave files,channels are interleaved. E.g.,
128 128 # "L1R1L2R2..." for stereo. See
129 129 # http://msdn.microsoft.com/en-us/library/windows/hardware/dn653308(v=vs.85).aspx
130 130 # for channel ordering
131 131 nchan = data.shape[0]
132 132 data = data.T.ravel()
133 133 else:
134 134 raise ValueError('Array audio input must be a 1D or 2D array')
135 135 scaled = np.int16(data/np.max(np.abs(data))*32767).tolist()
136 136 except ImportError:
137 137 # check that it is a "1D" list
138 138 idata = iter(data) # fails if not an iterable
139 139 try:
140 140 iter(idata.next())
141 141 raise TypeError('Only lists of mono audio are '
142 142 'supported if numpy is not installed')
143 143 except TypeError:
144 144 # this means it's not a nested list, which is what we want
145 145 pass
146 146 maxabsvalue = float(max([abs(x) for x in data]))
147 147 scaled = [int(x/maxabsvalue*32767) for x in data]
148 148 nchan = 1
149 149
150 150 fp = BytesIO()
151 151 waveobj = wave.open(fp,mode='wb')
152 152 waveobj.setnchannels(nchan)
153 153 waveobj.setframerate(rate)
154 154 waveobj.setsampwidth(2)
155 155 waveobj.setcomptype('NONE','NONE')
156 156 waveobj.writeframes(b''.join([struct.pack('<h',x) for x in scaled]))
157 157 val = fp.getvalue()
158 158 waveobj.close()
159 159
160 160 return val
161 161
162 162 def _data_and_metadata(self):
163 163 """shortcut for returning metadata with url information, if defined"""
164 164 md = {}
165 165 if self.url:
166 166 md['url'] = self.url
167 167 if md:
168 168 return self.data, md
169 169 else:
170 170 return self.data
171 171
172 172 def _repr_html_(self):
173 173 src = """
174 174 <audio controls="controls" {autoplay}>
175 175 <source src="{src}" type="{type}" />
176 176 Your browser does not support the audio element.
177 177 </audio>
178 178 """
179 179 return src.format(src=self.src_attr(),type=self.mimetype, autoplay=self.autoplay_attr())
180 180
181 181 def src_attr(self):
182 182 import base64
183 183 if self.embed and (self.data is not None):
184 184 data = base64=base64.b64encode(self.data).decode('ascii')
185 185 return """data:{type};base64,{base64}""".format(type=self.mimetype,
186 186 base64=data)
187 187 elif self.url is not None:
188 188 return self.url
189 189 else:
190 190 return ""
191 191
192 192 def autoplay_attr(self):
193 193 if(self.autoplay):
194 194 return 'autoplay="autoplay"'
195 195 else:
196 196 return ''
197 197
198 198 class IFrame(object):
199 199 """
200 200 Generic class to embed an iframe in an IPython notebook
201 201 """
202 202
203 203 iframe = """
204 204 <iframe
205 205 width="{width}"
206 206 height="{height}"
207 207 src="{src}{params}"
208 208 frameborder="0"
209 209 allowfullscreen
210 210 ></iframe>
211 211 """
212 212
213 213 def __init__(self, src, width, height, **kwargs):
214 214 self.src = src
215 215 self.width = width
216 216 self.height = height
217 217 self.params = kwargs
218 218
219 219 def _repr_html_(self):
220 220 """return the embed iframe"""
221 221 if self.params:
222 222 try:
223 223 from urllib.parse import urlencode # Py 3
224 224 except ImportError:
225 225 from urllib import urlencode
226 226 params = "?" + urlencode(self.params)
227 227 else:
228 228 params = ""
229 229 return self.iframe.format(src=self.src,
230 230 width=self.width,
231 231 height=self.height,
232 232 params=params)
233 233
234 234 class YouTubeVideo(IFrame):
235 235 """Class for embedding a YouTube Video in an IPython session, based on its video id.
236 236
237 237 e.g. to embed the video from https://www.youtube.com/watch?v=foo , you would
238 238 do::
239 239
240 240 vid = YouTubeVideo("foo")
241 241 display(vid)
242 242
243 243 To start from 30 seconds::
244 244
245 245 vid = YouTubeVideo("abc", start=30)
246 246 display(vid)
247 247
248 248 To calculate seconds from time as hours, minutes, seconds use
249 249 :class:`datetime.timedelta`::
250 250
251 251 start=int(timedelta(hours=1, minutes=46, seconds=40).total_seconds())
252 252
253 253 Other parameters can be provided as documented at
254 254 https://developers.google.com/youtube/player_parameters#parameter-subheader
255 255
256 256 When converting the notebook using nbconvert, a jpeg representation of the video
257 257 will be inserted in the document.
258 258 """
259 259
260 260 def __init__(self, id, width=400, height=300, **kwargs):
261 261 self.id=id
262 262 src = "https://www.youtube.com/embed/{0}".format(id)
263 263 super(YouTubeVideo, self).__init__(src, width, height, **kwargs)
264 264
265 265 def _repr_jpeg_(self):
266 from urllib import urlopen
267 return urlopen("https://img.youtube.com/vi/{id}/hqdefault.jpg".format(id=self.id)).read()
266 try:
267 from urllib.request import urlopen # Py3
268 except ImportError:
269 from urllib2 import urlopen
270 try:
271 return urlopen("https://img.youtube.com/vi/{id}/hqdefault.jpg".format(id=self.id)).read()
272 except IOError:
273 return None
268 274
269 275 class VimeoVideo(IFrame):
270 276 """
271 277 Class for embedding a Vimeo video in an IPython session, based on its video id.
272 278 """
273 279
274 280 def __init__(self, id, width=400, height=300, **kwargs):
275 281 src="https://player.vimeo.com/video/{0}".format(id)
276 282 super(VimeoVideo, self).__init__(src, width, height, **kwargs)
277 283
278 284 class ScribdDocument(IFrame):
279 285 """
280 286 Class for embedding a Scribd document in an IPython session
281 287
282 288 Use the start_page params to specify a starting point in the document
283 289 Use the view_mode params to specify display type one off scroll | slideshow | book
284 290
285 291 e.g to Display Wes' foundational paper about PANDAS in book mode from page 3
286 292
287 293 ScribdDocument(71048089, width=800, height=400, start_page=3, view_mode="book")
288 294 """
289 295
290 296 def __init__(self, id, width=400, height=300, **kwargs):
291 297 src="https://www.scribd.com/embeds/{0}/content".format(id)
292 298 super(ScribdDocument, self).__init__(src, width, height, **kwargs)
293 299
294 300 class FileLink(object):
295 301 """Class for embedding a local file link in an IPython session, based on path
296 302
297 303 e.g. to embed a link that was generated in the IPython notebook as my/data.txt
298 304
299 305 you would do::
300 306
301 307 local_file = FileLink("my/data.txt")
302 308 display(local_file)
303 309
304 310 or in the HTML notebook, just::
305 311
306 312 FileLink("my/data.txt")
307 313 """
308 314
309 315 html_link_str = "<a href='%s' target='_blank'>%s</a>"
310 316
311 317 def __init__(self,
312 318 path,
313 319 url_prefix='',
314 320 result_html_prefix='',
315 321 result_html_suffix='<br>'):
316 322 """
317 323 Parameters
318 324 ----------
319 325 path : str
320 326 path to the file or directory that should be formatted
321 327 directory_prefix : str
322 328 prefix to be prepended to all files to form a working link [default:
323 329 'files']
324 330 result_html_prefix : str
325 331 text to append to beginning to link [default: none]
326 332 result_html_suffix : str
327 333 text to append at the end of link [default: '<br>']
328 334 """
329 335 if isdir(path):
330 336 raise ValueError("Cannot display a directory using FileLink. "
331 337 "Use FileLinks to display '%s'." % path)
332 338 self.path = path
333 339 self.url_prefix = url_prefix
334 340 self.result_html_prefix = result_html_prefix
335 341 self.result_html_suffix = result_html_suffix
336 342
337 343 def _format_path(self):
338 344 fp = ''.join([self.url_prefix,self.path])
339 345 return ''.join([self.result_html_prefix,
340 346 self.html_link_str % (fp, self.path),
341 347 self.result_html_suffix])
342 348
343 349 def _repr_html_(self):
344 350 """return html link to file
345 351 """
346 352 if not exists(self.path):
347 353 return ("Path (<tt>%s</tt>) doesn't exist. "
348 354 "It may still be in the process of "
349 355 "being generated, or you may have the "
350 356 "incorrect path." % self.path)
351 357
352 358 return self._format_path()
353 359
354 360 def __repr__(self):
355 361 """return absolute path to file
356 362 """
357 363 return abspath(self.path)
358 364
359 365 class FileLinks(FileLink):
360 366 """Class for embedding local file links in an IPython session, based on path
361 367
362 368 e.g. to embed links to files that were generated in the IPython notebook
363 369 under ``my/data``, you would do::
364 370
365 371 local_files = FileLinks("my/data")
366 372 display(local_files)
367 373
368 374 or in the HTML notebook, just::
369 375
370 376 FileLinks("my/data")
371 377 """
372 378 def __init__(self,
373 379 path,
374 380 url_prefix='',
375 381 included_suffixes=None,
376 382 result_html_prefix='',
377 383 result_html_suffix='<br>',
378 384 notebook_display_formatter=None,
379 385 terminal_display_formatter=None,
380 386 recursive=True):
381 387 """
382 388 See :class:`FileLink` for the ``path``, ``url_prefix``,
383 389 ``result_html_prefix`` and ``result_html_suffix`` parameters.
384 390
385 391 included_suffixes : list
386 392 Filename suffixes to include when formatting output [default: include
387 393 all files]
388 394
389 395 notebook_display_formatter : function
390 396 Used to format links for display in the notebook. See discussion of
391 397 formatter functions below.
392 398
393 399 terminal_display_formatter : function
394 400 Used to format links for display in the terminal. See discussion of
395 401 formatter functions below.
396 402
397 403 Formatter functions must be of the form::
398 404
399 405 f(dirname, fnames, included_suffixes)
400 406
401 407 dirname : str
402 408 The name of a directory
403 409 fnames : list
404 410 The files in that directory
405 411 included_suffixes : list
406 412 The file suffixes that should be included in the output (passing None
407 413 meansto include all suffixes in the output in the built-in formatters)
408 414 recursive : boolean
409 415 Whether to recurse into subdirectories. Default is True.
410 416
411 417 The function should return a list of lines that will be printed in the
412 418 notebook (if passing notebook_display_formatter) or the terminal (if
413 419 passing terminal_display_formatter). This function is iterated over for
414 420 each directory in self.path. Default formatters are in place, can be
415 421 passed here to support alternative formatting.
416 422
417 423 """
418 424 if isfile(path):
419 425 raise ValueError("Cannot display a file using FileLinks. "
420 426 "Use FileLink to display '%s'." % path)
421 427 self.included_suffixes = included_suffixes
422 428 # remove trailing slashs for more consistent output formatting
423 429 path = path.rstrip('/')
424 430
425 431 self.path = path
426 432 self.url_prefix = url_prefix
427 433 self.result_html_prefix = result_html_prefix
428 434 self.result_html_suffix = result_html_suffix
429 435
430 436 self.notebook_display_formatter = \
431 437 notebook_display_formatter or self._get_notebook_display_formatter()
432 438 self.terminal_display_formatter = \
433 439 terminal_display_formatter or self._get_terminal_display_formatter()
434 440
435 441 self.recursive = recursive
436 442
437 443 def _get_display_formatter(self,
438 444 dirname_output_format,
439 445 fname_output_format,
440 446 fp_format,
441 447 fp_cleaner=None):
442 448 """ generate built-in formatter function
443 449
444 450 this is used to define both the notebook and terminal built-in
445 451 formatters as they only differ by some wrapper text for each entry
446 452
447 453 dirname_output_format: string to use for formatting directory
448 454 names, dirname will be substituted for a single "%s" which
449 455 must appear in this string
450 456 fname_output_format: string to use for formatting file names,
451 457 if a single "%s" appears in the string, fname will be substituted
452 458 if two "%s" appear in the string, the path to fname will be
453 459 substituted for the first and fname will be substituted for the
454 460 second
455 461 fp_format: string to use for formatting filepaths, must contain
456 462 exactly two "%s" and the dirname will be subsituted for the first
457 463 and fname will be substituted for the second
458 464 """
459 465 def f(dirname, fnames, included_suffixes=None):
460 466 result = []
461 467 # begin by figuring out which filenames, if any,
462 468 # are going to be displayed
463 469 display_fnames = []
464 470 for fname in fnames:
465 471 if (isfile(join(dirname,fname)) and
466 472 (included_suffixes is None or
467 473 splitext(fname)[1] in included_suffixes)):
468 474 display_fnames.append(fname)
469 475
470 476 if len(display_fnames) == 0:
471 477 # if there are no filenames to display, don't print anything
472 478 # (not even the directory name)
473 479 pass
474 480 else:
475 481 # otherwise print the formatted directory name followed by
476 482 # the formatted filenames
477 483 dirname_output_line = dirname_output_format % dirname
478 484 result.append(dirname_output_line)
479 485 for fname in display_fnames:
480 486 fp = fp_format % (dirname,fname)
481 487 if fp_cleaner is not None:
482 488 fp = fp_cleaner(fp)
483 489 try:
484 490 # output can include both a filepath and a filename...
485 491 fname_output_line = fname_output_format % (fp, fname)
486 492 except TypeError:
487 493 # ... or just a single filepath
488 494 fname_output_line = fname_output_format % fname
489 495 result.append(fname_output_line)
490 496 return result
491 497 return f
492 498
493 499 def _get_notebook_display_formatter(self,
494 500 spacer="&nbsp;&nbsp;"):
495 501 """ generate function to use for notebook formatting
496 502 """
497 503 dirname_output_format = \
498 504 self.result_html_prefix + "%s/" + self.result_html_suffix
499 505 fname_output_format = \
500 506 self.result_html_prefix + spacer + self.html_link_str + self.result_html_suffix
501 507 fp_format = self.url_prefix + '%s/%s'
502 508 if sep == "\\":
503 509 # Working on a platform where the path separator is "\", so
504 510 # must convert these to "/" for generating a URI
505 511 def fp_cleaner(fp):
506 512 # Replace all occurences of backslash ("\") with a forward
507 513 # slash ("/") - this is necessary on windows when a path is
508 514 # provided as input, but we must link to a URI
509 515 return fp.replace('\\','/')
510 516 else:
511 517 fp_cleaner = None
512 518
513 519 return self._get_display_formatter(dirname_output_format,
514 520 fname_output_format,
515 521 fp_format,
516 522 fp_cleaner)
517 523
518 524 def _get_terminal_display_formatter(self,
519 525 spacer=" "):
520 526 """ generate function to use for terminal formatting
521 527 """
522 528 dirname_output_format = "%s/"
523 529 fname_output_format = spacer + "%s"
524 530 fp_format = '%s/%s'
525 531
526 532 return self._get_display_formatter(dirname_output_format,
527 533 fname_output_format,
528 534 fp_format)
529 535
530 536 def _format_path(self):
531 537 result_lines = []
532 538 if self.recursive:
533 539 walked_dir = list(walk(self.path))
534 540 else:
535 541 walked_dir = [next(walk(self.path))]
536 542 walked_dir.sort()
537 543 for dirname, subdirs, fnames in walked_dir:
538 544 result_lines += self.notebook_display_formatter(dirname, fnames, self.included_suffixes)
539 545 return '\n'.join(result_lines)
540 546
541 547 def __repr__(self):
542 548 """return newline-separated absolute paths
543 549 """
544 550 result_lines = []
545 551 if self.recursive:
546 552 walked_dir = list(walk(self.path))
547 553 else:
548 554 walked_dir = [next(walk(self.path))]
549 555 walked_dir.sort()
550 556 for dirname, subdirs, fnames in walked_dir:
551 557 result_lines += self.terminal_display_formatter(dirname, fnames, self.included_suffixes)
552 558 return '\n'.join(result_lines)
General Comments 0
You need to be logged in to leave comments. Login now