Hide keyboard shortcuts

Hot-keys on this page

r m x p   toggle line displays

j k   next/prev highlighted chunk

0   (zero) top of page

1   (one) first highlighted chunk

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

#!/usr/bin/env python3 

# -*- coding: utf-8 -*- 

 

""" 

Visualizing the Output of LDA Models 

************************************ 

 

Functions and classes of this module are for visualizing LDA models. 

 

Contents 

******** 

*  

""" 

 

 

import logging 

from dariah_topics import postprocessing 

import matplotlib 

matplotlib.use('Agg') 

import matplotlib.pyplot as plt 

from matplotlib import cm 

import numpy as np 

import os 

import pandas as pd 

from bokeh.plotting import figure 

from bokeh import palettes 

from bokeh.models import ( 

ColumnDataSource, 

HoverTool, 

LinearColorMapper, 

BasicTicker, 

ColorBar 

) 

 

import regex 

from collections import defaultdict 

from wordcloud import WordCloud 

 

log = logging.getLogger(__name__) 

 

 

def plot_wordcloud(weights, enable_notebook=True, **kwargs): 

"""Plots a wordcloud based on tokens and frequencies. 

 

Args: 

weights (dict): A dictionary (or :module:``pandas`` Series) with tokens 

as keys and frequencies as values. 

enable_notebook (bool), optional: If True, enables :module:``matplotlib`` 

to show its figures within a Jupyter notebook. 

font_path (str), optional: Font path to the font that will be used (OTF or TTF). 

Defaults to DroidSansMono path on a Linux machine. If you are on 

another OS or don't have this font, you need to adjust this path. 

width (int), optional: Width of the canvas. Defaults to 400. 

height (int), optional: Height of the canvas. Defaults to 200. 

prefer_horizontal (float): The ratio of times to try horizontal fitting 

as opposed to vertical. If ``prefer_horizontal < 1``, the algorithm 

will try rotating the word if it doesn't fit. (There is currently no 

built-in way to get only vertical words. Defaults to 0.90. 

mask (nd-array), optional: If not None, gives a binary mask on where to draw words. 

If mask is not None, width and height will be ignored and the shape 

of mask will be used instead. All white (#FF or #FFFFFF) entries 

will be considerd 'masked out' while other entries will be free to 

draw on. Defaults to None. 

scale (float), optional: Scaling between computation and drawing. For large word-cloud 

images, using scale instead of larger canvas size is significantly 

faster, but might lead to a coarser fit for the words. Defaults to 1. 

min_font_size (int), optional: Smallest font size to use. Will stop when there is 

no more room in this size. Defaults to 4. 

font_step (int), optional: Step size for the font. ``font_step > 1`` might speed 

up computation but give a worse fit. Defaults to 1. 

max_words (int), optional: The maximum number of words. Defaults to 200. 

stopwords (set), optional: The words that will be eliminated. If None, the build-in 

stopwords list will be used. 

background_color (str), optional: Background color for the word cloud image. 

Defaults to ``black``. 

max_font_size (int), optional: Maximum font size for the largest word. If None, 

height of the image is used. 

mode (str), optional: Transparent background will be generated when mode is ``RGBA`` 

and background_color is None. Defaults to ``RGB``. 

relative_scaling (float), optional: Importance of relative word frequencies for 

font-size. With ``relative_scaling=0``, only word-ranks are considered. 

With ``relative_scaling=1``, a word that is twice as frequent will 

have twice the size. If you want to consider the word frequencies and 

not only their rank, ``relative_scaling`` around .5 often looks good. 

Defaults to 0.5. 

color_func (callable), optional: Callable with parameters ``word``, ``font_size``, 

``position``, ``orientation``, ``font_path``, ``random_state`` that 

returns a PIL color for each word. Overwrites ``colormap``. See ``colormap`` 

for specifying a :module:``matplotlib`` colormap instead. 

collocations (bool), optional: Whether to include collocations (bigrams) of two words. 

Defaults to True. 

colormap (str), optional: :module:``matplotlib`` colormap to randomly draw colors 

from for each word. Ignored if ``color_func`` is specified. Defaults to 

``viridis``. 

normalize_plurals (bool), optional: Whether to remove trailing 's' from words. If 

True and a word appears with and without a trailing 's', the one with 

trailing 's' is removed and its counts are added to the version without 

trailing 's' -- unless the word ends with 'ss'. Defaults to True. 

 

Returns: 

WordCloud object. 

 

Example: 

>>> weights = {'an': 2, 'example': 1} 

>>> plot_wordcloud(weights, enable_notebook=False) # doctest: +ELLIPSIS 

<wordcloud.wordcloud.WordCloud object at ...> 

""" 

wordcloud = WordCloud(**kwargs).fit_words(weights) 

if enable_notebook: 

from IPython import get_ipython 

get_ipython().run_line_magic('matplotlib', 'inline') 

try: 

fig, ax = plt.subplots(figsize=(kwargs['width'] / 96, kwargs['height'] / 96)) 

except KeyError: 

fig, ax = plt.subplots(figsize=(400 / 96, 200 / 96)) 

ax.axis('off') 

ax.imshow(wordcloud) 

return wordcloud 

 

 

def plot_key_frequencies(keys=None, overall_freqs=None, within_topic_freqs=None, 

within_topic_color='#FF1727', document_term_matrix=None, 

model=None, vocabulary=None, topic_no=None, overall_color='#053967', 

figsize=(15, 7), dpi=None, overall_edgecolor=None, 

overall_linewidth=None, overall_alpha=0.9, within_topic_edgecolor=None, 

within_topic_linewidth=None, within_topic_alpha=0.9, 

label_fontsize=15, num_keys=None, tick_fontsize=14, legend_fontsize=15, 

legend=True, enable_notebook=True): 

"""Plots key frequencies overall and from within topic. 

 

Args: 

keys (list): A list of tokens. Defaults to None. 

overall_freqs (list): A list of frequencies. Defaults to None. 

within_topic_freqs (list): A list of frequencies. Defaults to None. 

within_topic_color (str), optional: Color for topic frequencies bar. Defaults to 

``#FF1727``. 

document_term_matrix (pandas DataFrame), optional: A document-term matrix. Defaults 

to None. 

model, optional: A LDA model. Defaults to None. 

vocabulary (list), optional: Vocabulary of the corpus. Defaults to None. 

topic_no (int), optional: Number of topic. Defaults to None. 

overall_color (str), optional: Color for overall frequencies bar. Defaults to ``#053967``. 

figsize (tuple), optional: Size of the figure. Defaults to ``(15, 7)``. 

dpi (int), optional: Dots per inch. Defaults to None. 

overall_edgecolor (str), optional: Color for edgecolors of overall frequencies bar. 

Defaults to None. 

overall_linewidth (int), optional: Linewidth of overall frequencies bar. Defaults to 

None. 

overall_alpha (int), optional: Alpha for overall frequencies bar. Defaults to 0.9. 

within_topic_edgecolor (str), optional: Color for edgecolors of overall frequencies bar. 

Defaults to None. 

within_topic_linewidth (int), optional: Linewidth of overall frequencies bar. Defaults to 

None. 

within_topic_alpha (int), optional: Alpha for overall frequencies bar. Defaults to 0.9. 

label_fontsize (int), optional: Fontsize of x-axis and y-axis labels. Defaults to 15. 

num_keys (int), optional: Number of tokens for y-axis. Defaults to None. 

tick_fontsize (int), optional: Fontsize of x- and y-ticks. Defaults to 14. 

legend_fontsize (int), optional: Fontsize of the legend. Defaults to 15. 

legend (bool), optional: If True, legend will be displayed. Defaults to True. 

enable_notebook (bool), optional: If True, enables :module:``matplotlib`` 

to show its figures within a Jupyter notebook. 

 

Returns: 

Figure object. 

 

Example: 

>>> keys = ['one', 'example'] 

>>> overall_freqs = [20, 10] 

>>> within_topic_freqs = [10, 5] 

>>> plot_key_frequencies(keys=keys, 

... overall_freqs=overall_freqs, 

... within_topic_freqs=within_topic_freqs, 

... enable_notebook=False) # doctest: +ELLIPSIS 

<matplotlib.figure.Figure object at ...> 

""" 

if enable_notebook: 

from IPython import get_ipython 

get_ipython().run_line_magic('matplotlib', 'inline') 

if model: 

within_topic_freqs = postprocessing.get_sorted_values_from_distribution(model.components_[topic_no], 

model.components_[topic_no], 

num_keys) 

within_topic_freqs = [dist * len(vocabulary) for dist in within_topic_freqs] 

total = [document_term_matrix[token].sum() for token in vocabulary] 

overall_freqs = postprocessing.get_sorted_values_from_distribution(total, 

model.components_[topic_no], 

num_keys) 

keys = postprocessing.get_sorted_values_from_distribution(vocabulary, 

model.components_[topic_no], 

num_keys) 

fig, ax = plt.subplots(figsize=figsize, dpi=dpi) 

y_axis = np.arange(len(keys)) 

overall = ax.barh(y_axis, overall_freqs, color=overall_color, edgecolor=overall_edgecolor, 

linewidth=overall_linewidth, alpha=overall_alpha) 

within = ax.barh(y_axis, within_topic_freqs, color=within_topic_color, 

edgecolor=within_topic_edgecolor, linewidth=within_topic_linewidth, 

alpha=within_topic_alpha) 

ax.set_xlabel('Frequency', fontsize=label_fontsize) 

ax.set_ylabel('Key', fontsize=label_fontsize) 

ax.set_yticks(y_axis) 

ax.set_yticklabels(keys, fontsize=tick_fontsize) 

ax.tick_params(axis='x', labelsize=tick_fontsize) 

if legend: 

ax.legend(handles=[overall, within], labels=['Overall', 'Within Topic'], loc='best', 

fontsize=legend_fontsize) 

return fig 

 

 

class PlotDocumentTopics: 

""" 

Class to visualize document-topic matrix. 

""" 

def __init__(self, document_topics, enable_notebook=True): 

self.document_topics = document_topics 

self.enable_notebook = enable_notebook 

if enable_notebook: 

self.show = self.notebook_handling() 

 

@staticmethod 

def notebook_handling(): 

"""Runs cell magic for Jupyter notebooks 

""" 

from IPython import get_ipython 

get_ipython().run_line_magic('matplotlib', 'inline') 

from bokeh.io import output_notebook, show 

output_notebook() 

return show 

 

def static_heatmap(self, figsize=(1000 / 96, 600 / 96), dpi=None, 

labels_fontsize=13, cmap='Blues', ticks_fontsize=12, 

xlabel='Document', ylabel='Topic', xticks_bottom=0.1, 

xticks_rotation=50, xticks_ha='right', colorbar=False): 

"""Plots a static heatmap. 

 

Args: 

figsize (tuple), optional: Size of the figure in inches. Defaults to 

``(1000 / 96, 500 / 96)``. 

dpi (int), optional: Dots per inch. Defaults to None. 

labels_fontsize (int), optional: Fontsize of the figure labels. Defaults 

to 13. 

cmap (str), optional: Colormap for the figure. Defaults to ``Blues``. 

ticks_fontsize (int), optional: Fontsize of axis ticks. Defaults to 12. 

xlabel (str), optional: Label of x-axis. Defaults to ``Document``. 

ylabel (str), optional: Label of y-axis. Defaults to ``Topic``. 

xticks_bottom (str), optional: Distance to bottom of x-ticks. Defaults 

to 0.1. 

xticks_rotation (int), optional: Rotation degree of x-ticks. Defaults 

to 50. 

xticks_ha (str), optional: The horizontal alignment of the x-tick labels. 

Defaulst to ``right``. 

colorbar (bool), optional: If True, include colorbar. Defaults to True. 

 

Returns: 

Figure object. 

""" 

fig, ax = plt.subplots(figsize=figsize, dpi=dpi) 

heatmap = ax.pcolor(self.document_topics, cmap=cmap) 

ax.set_xlabel(xlabel, fontsize=labels_fontsize) 

ax.set_ylabel(ylabel, fontsize=labels_fontsize) 

ax.set_xticks(np.arange(self.document_topics.shape[1]) + 0.5) 

ax.set_yticks(np.arange(self.document_topics.shape[0]) + 0.5) 

ax.set_xticklabels(list(self.document_topics.columns), fontsize=ticks_fontsize) 

ax.set_yticklabels(list(self.document_topics.index), fontsize=ticks_fontsize) 

fig.autofmt_xdate(bottom=xticks_bottom, rotation=xticks_rotation, ha=xticks_ha) 

if colorbar: 

cax = ax.imshow(self.document_topics, interpolation='nearest', cmap=cmap) 

cbar = fig.colorbar(cax, ticks=np.arange(0, 1, 0.1)) 

return fig 

 

def __static_barchart(self, index, describer, figsize=(11, 7), color='#053967', 

edgecolor=None, linewidth=None, alpha=None, labels_fontsize=15, 

ticks_fontsize=14, title=True, title_fontsize=17, 

dpi=None, transpose_data=False): 

"""Plots a static barchart. 

 

Args: 

index Union(int, str): Index of document-topics matrix column or 

name of column. 

describer (str): Describer of what the plot shows, e.g. either document 

or topic. 

title (bool), optional: If True, figure will have a title in the format 

``describer: index``. 

title_fontsize (int), optional: Fontsize of figure title. 

transpose_data (bool): If True. document-topics matrix will be transposed. 

Defaults to False. 

color (str), optional: Color of the bins. Defaults to ``#053967``. 

edgecolor (str), optional: Color of the bin edges. Defaults to None. 

lindewidth (float), optional: Width of bin lines. Defaults to None. 

alpha (float): Alpha value used for blending. Defaults to None. 

figsize (tuple), optional: Size of the figure in inches. Defaults to 

``(1000 / 96, 500 / 96)``. 

dpi (int), optional: Dots per inch. Defaults to None. 

labels_fontsize (int), optional: Fontsize of the figure labels. Defaults 

to 15. 

ticks_fontsize (int), optional: Fontsize of axis ticks. Defaults to 14. 

 

Returns: 

Figure object. 

""" 

fig, ax = plt.subplots(figsize=figsize, dpi=dpi) 

if isinstance(index, int): 

if transpose_data: 

proportions = self.document_topics.T.iloc[index] 

else: 

proportions = self.document_topics.iloc[index] 

if title: 

plot_title = '{0}: {1}'.format(describer, proportions.name) 

ax.set_title(plot_title, fontsize=title_fontsize) 

elif isinstance(index, str): 

if transpose: 

proportions = self.document_topics.T.loc[index] 

else: 

proportions = self.document_topics.loc[index] 

if title: 

plot_title = '{}: {}'.format(describer, index) 

ax.set_title(plot_title, fontsize=title_fontsize) 

else: 

raise ValueError("{} must be int or str.".format(index)) 

 

y_axis = np.arange(len(proportions)) 

x_axis = proportions 

y_ticks_labels = proportions.index 

ax.barh(y_axis, x_axis, color=color, edgecolor=edgecolor, linewidth=linewidth, alpha=alpha) 

ax.set_xlabel('Proportion', fontsize=labels_fontsize) 

ax.set_ylabel(describer, fontsize=labels_fontsize) 

ax.set_yticks(y_axis) 

ax.set_yticklabels(y_ticks_labels, fontsize=ticks_fontsize) 

ax.tick_params(axis='x', labelsize=ticks_fontsize) 

return fig 

 

def static_barchart_per_topic(self, **kwargs): 

"""Plots a static barchart per topic. 

 

Args: 

index Union(int, str): Index of document-topics matrix column or 

name of column. 

describer (str): Describer of what the plot shows, e.g. either document 

or topic. 

title (bool), optional: If True, figure will have a title in the format 

``describer: index``. 

title_fontsize (int), optional: Fontsize of figure title. 

transpose_data (bool): If True. document-topics matrix will be transposed. 

Defaults to False. 

color (str), optional: Color of the bins. Defaults to ``#053967``. 

edgecolor (str), optional: Color of the bin edges. Defaults to None. 

lindewidth (float), optional: Width of bin lines. Defaults to None. 

alpha (float): Alpha value used for blending. Defaults to None. 

figsize (tuple), optional: Size of the figure in inches. Defaults to 

``(1000 / 96, 500 / 96)``. 

dpi (int), optional: Dots per inch. Defaults to None. 

labels_fontsize (int), optional: Fontsize of the figure labels. Defaults 

to 15. 

ticks_fontsize (int), optional: Fontsize of axis ticks. Defaults to 14. 

 

Returns: 

Figure object. 

""" 

return self.__static_barchart(**kwargs) 

 

def static_barchart_per_document(self, **kwargs): 

"""Plots a static barchart per document. 

 

Args: 

index Union(int, str): Index of document-topics matrix column or 

name of column. 

describer (str): Describer of what the plot shows, e.g. either document 

or topic. 

title (bool), optional: If True, figure will have a title in the format 

``describer: index``. 

title_fontsize (int), optional: Fontsize of figure title. 

transpose_data (bool): If True. document-topics matrix will be transposed. 

Defaults to False. 

color (str), optional: Color of the bins. Defaults to ``#053967``. 

edgecolor (str), optional: Color of the bin edges. Defaults to None. 

lindewidth (float), optional: Width of bin lines. Defaults to None. 

alpha (float): Alpha value used for blending. Defaults to None. 

figsize (tuple), optional: Size of the figure in inches. Defaults to 

``(1000 / 96, 500 / 96)``. 

dpi (int), optional: Dots per inch. Defaults to None. 

labels_fontsize (int), optional: Fontsize of the figure labels. Defaults 

to 15. 

ticks_fontsize (int), optional: Fontsize of axis ticks. Defaults to 14. 

 

Returns: 

Figure object. 

""" 

return self.__static_barchart(**kwargs, transpose_data=True) 

 

def interactive_heatmap(self, palette=palettes.Blues[9], reverse_palette=True, 

tools='hover, pan, reset, save, wheel_zoom, zoom_in, zoom_out', 

width=1000, height=550, x_axis_location='below', toolbar_location='above', 

sizing_mode='fixed', line_color=None, grid_line_color=None, axis_line_color=None, 

major_tick_line_color=None, major_label_text_font_size='9pt', 

major_label_standoff=0, major_label_orientation=3.14/3, colorbar=True): 

"""Plots an interactive heatmap. 

 

Args: 

palette (list), optional: A list of color values. Defaults to ``palettes.Blues[9]``. 

reverse_palette (bool), optional: If True, color values of ``palette`` will 

be reversed. Defaults to True. 

tools (str), optional: Tools, which will be includeded. Defaults to ``hover, 

pan, reset, save, wheel_zoom, zoom_in, zoom_out``. 

width (int), optional: Width of the figure. Defaults to 1000. 

height (int), optional: Height of the figure. Defaults to 550. 

x_axis_location (str), optional: Location of the x-axis. Defaults to 

``below``. 

toolbar_location (str), optional: Location of the toolbar. Defaults to 

``above``. 

sizing_mode (str), optional: Size fixed or width oriented. Defaults to ``fixed``. 

line_color (str): Color for lines. Defaults to None. 

grid_line_color (str): Color for grid lines. Defaults to None. 

axis_line_color (str): Color for axis lines. Defaults to None. 

major_tick_line_color (str): Color for major tick lines. Defaults to None. 

major_label_text_font_size (str): Font size for major label text. Defaults 

to ``9pt``. 

major_label_standoff (int): Standoff for major labels. Defaults to 0. 

major_label_orientation (float): Orientation for major labels. Defaults 

to ``3.14 / 3``. 

colorbar (bool): If True, colorbar will be included. 

 

Returns: 

Figure object. 

""" 

if reverse_palette: 

palette = list(reversed(palette)) 

 

x_range = list(self.document_topics.columns) 

y_range = list(self.document_topics.index) 

 

stacked_data = pd.DataFrame(self.document_topics.stack()).reset_index() 

stacked_data.columns = ['Topics', 'Documents', 'Distributions'] 

mapper = LinearColorMapper(palette=palette, 

low=stacked_data.Distributions.min(), 

high=stacked_data.Distributions.max()) 

source = ColumnDataSource(stacked_data) 

 

fig = figure(x_range=x_range, 

y_range=y_range, 

x_axis_location=x_axis_location, 

plot_width=width, plot_height=height, 

tools=tools, toolbar_location=toolbar_location, 

sizing_mode=sizing_mode, 

logo=None) 

fig.rect(x='Documents', y='Topics', source=source, width=1, height=1, 

fill_color={'field': 'Distributions', 'transform': mapper}, 

line_color=line_color) 

 

fig.grid.grid_line_color = grid_line_color 

fig.axis.axis_line_color = axis_line_color 

fig.axis.major_tick_line_color = major_tick_line_color 

fig.axis.major_label_text_font_size = major_label_text_font_size 

fig.axis.major_label_standoff = major_label_standoff 

fig.xaxis.major_label_orientation = major_label_orientation 

 

if 'hover' in tools: 

fig.select_one(HoverTool).tooltips = [('Document', '@Documents'), 

('Topic', '@Topics'), 

('Score', '@Distributions')] 

 

if colorbar: 

feature = ColorBar(color_mapper=mapper, major_label_text_font_size=major_label_text_font_size, 

ticker=BasicTicker(desired_num_ticks=len(palette)), 

label_standoff=6, border_line_color=None, location=(0, 0)) 

fig.add_layout(feature, 'right') 

if self.enable_notebook: 

self.show(fig, notebook_handle=True) 

return fig 

 

 

def __interactive_barchart(self, index, describer, tools='hover, pan, reset, save, wheel_zoom, zoom_in, zoom_out', 

width=1000, height=400, toolbar_location='above', 

sizing_mode='fixed', line_color=None, grid_line_color=None, axis_line_color=None, 

major_tick_line_color=None, major_label_text_font_size='9pt', 

major_label_standoff=0, title=True, bin_height=0.5, 

transpose_data=False, bar_color='#053967'): 

"""Plots an interactive barchart. 

 

Args: 

index Union(int, str): Index of document-topics matrix column or 

name of column. 

describer (str): Describer of what the plot shows, e.g. either document 

or topic. 

bar_color (str), optional: Color of bars. Defaults to ``#053967``. 

transpose_data (bool): If True. document-topics matrix will be transposed. 

Defaults to False. 

title (bool), optional: If True, figure will have a title in the format 

``describer: index``. 

tools (str), optional: Tools, which will be includeded. Defaults to ``hover, 

pan, reset, save, wheel_zoom, zoom_in, zoom_out``. 

width (int), optional: Width of the figure. Defaults to 1000. 

height (int), optional: Height of the figure. Defaults to 400. 

x_axis_location (str), optional: Location of the x-axis. Defaults to 

``below``. 

toolbar_location (str), optional: Location of the toolbar. Defaults to 

``above``. 

sizing_mode (str), optional: Size fixed or width oriented. Defaults to ``fixed``. 

line_color (str): Color for lines. Defaults to None. 

grid_line_color (str): Color for grid lines. Defaults to None. 

axis_line_color (str): Color for axis lines. Defaults to None. 

major_tick_line_color (str): Color for major tick lines. Defaults to None. 

major_label_text_font_size (str): Font size for major label text. Defaults 

to ``9pt``. 

major_label_standoff (int): Standoff for major labels. Defaults to 0. 

 

Returns: 

Figure object. 

""" 

if isinstance(index, int): 

if transpose_data: 

proportions = self.document_topics.T.iloc[index] 

else: 

proportions = self.document_topics.iloc[index] 

if title: 

plot_title = '{}: {}'.format(describer, proportions.name) 

elif isinstance(index, str): 

if transpose_data: 

proportions = self.document_topics.T.loc[index] 

else: 

proportions = self.document_topics.loc[index] 

if title: 

plot_title = '{}: {}'.format(describer, index) 

else: 

raise ValueError("{} must be int or str.".format(index)) 

 

x_axis = proportions 

y_range = list(proportions.index) 

 

source = ColumnDataSource(dict(Describer=y_range, Proportion=x_axis)) 

 

fig = figure(y_range=y_range, title=plot_title, plot_width=width, plot_height=height, 

tools=tools, toolbar_location=toolbar_location, 

sizing_mode=sizing_mode, logo=None) 

fig.hbar(y='Describer', right='Proportion', height=bin_height, source=source, 

line_color=line_color, color=bar_color) 

 

fig.xgrid.grid_line_color = None 

fig.x_range.start = 0 

fig.grid.grid_line_color = grid_line_color 

fig.axis.axis_line_color = axis_line_color 

fig.axis.major_tick_line_color = major_tick_line_color 

fig.axis.major_label_text_font_size = major_label_text_font_size 

fig.axis.major_label_standoff = major_label_standoff 

 

if 'hover' in tools: 

fig.select_one(HoverTool).tooltips = [('Proportion', '@Proportion')] 

if self.enable_notebook: 

self.show(fig, notebook_handle=True) 

return fig 

 

def interactive_barchart_per_topic(self, **kwargs): 

"""Plots an interactive barchart per topic. 

 

Args: 

index Union(int, str): Index of document-topics matrix column or 

name of column. 

describer (str): Describer of what the plot shows, e.g. either document 

or topic. 

bar_color (str), optional: Color of bars. Defaults to ``#053967``. 

transpose_data (bool): If True. document-topics matrix will be transposed. 

Defaults to False. 

title (bool), optional: If True, figure will have a title in the format 

``describer: index``. 

tools (str), optional: Tools, which will be includeded. Defaults to ``hover, 

pan, reset, save, wheel_zoom, zoom_in, zoom_out``. 

width (int), optional: Width of the figure. Defaults to 1000. 

height (int), optional: Height of the figure. Defaults to 400. 

x_axis_location (str), optional: Location of the x-axis. Defaults to 

``below``. 

toolbar_location (str), optional: Location of the toolbar. Defaults to 

``above``. 

sizing_mode (str), optional: Size fixed or width oriented. Defaults to ``fixed``. 

line_color (str): Color for lines. Defaults to None. 

grid_line_color (str): Color for grid lines. Defaults to None. 

axis_line_color (str): Color for axis lines. Defaults to None. 

major_tick_line_color (str): Color for major tick lines. Defaults to None. 

major_label_text_font_size (str): Font size for major label text. Defaults 

to ``9pt``. 

major_label_standoff (int): Standoff for major labels. Defaults to 0. 

 

Returns: 

Figure object. 

""" 

return self.__interactive_barchart(**kwargs) 

 

def interactive_barchart_per_document(self, **kwargs): 

"""Plots an interactive barchart per document. 

 

Args: 

index Union(int, str): Index of document-topics matrix column or 

name of column. 

describer (str): Describer of what the plot shows, e.g. either document 

or topic. 

bar_color (str), optional: Color of bars. Defaults to ``#053967``. 

transpose_data (bool): If True. document-topics matrix will be transposed. 

Defaults to False. 

title (bool), optional: If True, figure will have a title in the format 

``describer: index``. 

tools (str), optional: Tools, which will be includeded. Defaults to ``hover, 

pan, reset, save, wheel_zoom, zoom_in, zoom_out``. 

width (int), optional: Width of the figure. Defaults to 1000. 

height (int), optional: Height of the figure. Defaults to 400. 

x_axis_location (str), optional: Location of the x-axis. Defaults to 

``below``. 

toolbar_location (str), optional: Location of the toolbar. Defaults to 

``above``. 

sizing_mode (str), optional: Size fixed or width oriented. Defaults to ``fixed``. 

line_color (str): Color for lines. Defaults to None. 

grid_line_color (str): Color for grid lines. Defaults to None. 

axis_line_color (str): Color for axis lines. Defaults to None. 

major_tick_line_color (str): Color for major tick lines. Defaults to None. 

major_label_text_font_size (str): Font size for major label text. Defaults 

to ``9pt``. 

major_label_standoff (int): Standoff for major labels. Defaults to 0. 

 

Returns: 

Figure object. 

""" 

return self.__interactive_barchart(**kwargs, transpose_data=True) 

 

def topic_over_time(self, pattern = r"\d{4}", threshold=0.1, starttime=1841, endtime=1920): 

"""Creates a visualization that shows topics over time. 

 

Description: 

With this function you can plot topics over time using metadata stored in the documents name. 

Only works with mallet output. 

 

Args: 

labels(list): first three keys in a topic to select 

threshold(float): threshold set to define if a topic in a document is viable 

starttime(int): sets starting point for visualization 

endtime(int): sets ending point for visualization 

 

 

Returns:  

matplotlib plot 

 

Note: this function is created for a corpus with filenames that looks like: 

1866_ArticleName.txt 

 

ToDo: make it compatible with gensim output 

Doctest 

 

""" 

years=list(range(starttime,endtime)) 

#doc_topicT = doc_topics.T 

topiclabels = [] 

reg = regex.compile(pattern) 

for topiclabel in self.document_topics.index.values: 

for topiclabel in topiclabels: 

topic_over_threshold_per_year = [] 

mask = doc_topics.loc[topiclabel] > threshold 

df = doc_topics.loc[topiclabel].loc[mask] 

#df = doc_topics.loc[doc_topics.loc[topiclabel] > threshold] 

#print (df) 

d = defaultdict(int) 

for item in df.index.values: 

year = reg.findall(item) 

d[year[0]]+=1 

for year in years: 

topic_over_threshold_per_year.append(d[str(year)]) 

plt.plot(years, topic_over_threshold_per_year, label=topiclabel) 

 

plt.xlabel('Year') 

plt.ylabel('count topics over threshold') 

plt.legend() 

fig = plt.gcf() 

fig.set_size_inches(18.5, 10.5) 

return fig 

 

@staticmethod 

def to_file(fig, filename): 

"""Saves a figure object to file. 

 

Args: 

fig Union(bokeh.figure, matplotlib.figure): Figure produced by either 

bokeh or matplotlib. 

filename (str): Name of the file with an extension, e.g. ``plot.png``. 

 

Returns: 

None. 

""" 

import matplotlib 

import bokeh 

if isinstance(fig, bokeh.plotting.figure.Figure): 

ext = os.path.splitext(filename)[1] 

if ext == '.png': 

export_png(fig, filename) 

elif ext == '.svg': 

fig.output_backend = 'svg' 

export_svgs(fig, filename) 

elif ext == '.html': 

output_file(filename) 

elif isinstance(fig, matplotlib.figure.Figure): 

fig.savefig(filename) 

return None