tensorflow/tensorflow/tools/docs/parser.py at master · codeplaysoftware/tensorflow · GitHub

Repository navigation

History

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

884

885

886

887

888

889

890

891

892

893

894

895

896

897

898

899

900

901

902

903

904

905

906

907

908

909

910

911

912

913

914

915

916

917

918

919

920

921

922

923

924

925

926

927

928

929

930

931

932

933

934

935

936

937

938

939

940

941

942

943

944

945

946

947

948

949

950

951

952

953

954

955

956

957

958

959

960

961

962

963

964

965

966

967

968

969

970

971

972

973

974

975

976

977

978

979

980

981

982

983

984

985

986

987

988

989

990

991

992

993

994

995

996

997

998

999

1000

# Licensed under the Apache License, Version 2.0 (the "License");

# you may not use this file except in compliance with the License.

# You may obtain a copy of the License at

# http://www.apache.org/licenses/LICENSE-2.0

# Unless required by applicable law or agreed to in writing, software

# distributed under the License is distributed on an "AS IS" BASIS,

# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.

# See the License for the specific language governing permissions and

# limitations under the License.

# ==============================================================================

"""Turn Python docstrings into Markdown for TensorFlow documentation."""

from __future__ import absolute_import

from __future__ import division

from __future__ import print_function

import ast

import collections

import functools

import itertools

import json

import os

import re

import sys

import astor

import six

from google.protobuf.message import Message as ProtoMessage

from tensorflow.python.util import tf_inspect

# A regular expression capturing a python identifier.

IDENTIFIER_RE = r'[a-zA-Z_]\w*'

class TFDocsError(Exception):

pass

class _Errors(object):

"""A collection of errors."""

def __init__(self):

self._errors = []

def log_all(self):

"""Log all the collected errors to the standard error."""

template = 'ERROR:\n output file name: %s\n %s\n\n'

for full_name, message in self._errors:

print(template % (full_name, message), file=sys.stderr)

def append(self, full_name, message):

"""Add an error to the collection.

Args:

full_name: The path to the file in which the error occurred.

message: The message to display with the error.

"""

self._errors.append((full_name, message))

def __len__(self):

return len(self._errors)

def __eq__(self, other):

if not isinstance(other, _Errors):

return False

return self._errors == other._errors # pylint: disable=protected-access

def documentation_path(full_name):

"""Returns the file path for the documentation for the given API symbol.

Given the fully qualified name of a library symbol, compute the path to which

to write the documentation for that symbol (relative to a base directory).

Documentation files are organized into directories that mirror the python

module/class structure.

Args:

full_name: Fully qualified name of a library symbol.

Returns:

The file path to which to write the documentation for `full_name`.

"""

dirs = full_name.split('.')

return os.path.join(*dirs) + '.md'

def _get_raw_docstring(py_object):

"""Get the docs for a given python object.

Args:

py_object: A python object to retrieve the docs for (class, function/method,

or module).

Returns:

The docstring, or the empty string if no docstring was found.

"""

# For object instances, tf_inspect.getdoc does give us the docstring of their

# type, which is not what we want. Only return the docstring if it is useful.

if (tf_inspect.isclass(py_object) or tf_inspect.ismethod(py_object) or

tf_inspect.isfunction(py_object) or tf_inspect.ismodule(py_object) or

isinstance(py_object, property)):

return tf_inspect.getdoc(py_object) or ''

else:

return ''

# A regular expression for capturing a @{symbol} reference.

SYMBOL_REFERENCE_RE = re.compile(

r"""

# Start with a literal "@{".

@\{

# Group at least 1 symbol, not "}".

([^}]+)

# Followed by a closing "}"

""",

flags=re.VERBOSE)

AUTO_REFERENCE_RE = re.compile(r'`([a-zA-Z0-9_.]+?)`')

class ReferenceResolver(object):

"""Class for replacing @{...} references with Markdown links.

Attributes:

current_doc_full_name: A string (or None) indicating the name of the

document currently being processed, so errors can reference the broken

doc.

"""

def __init__(self, duplicate_of, doc_index, is_class, is_module,

py_module_names):

"""Initializes a Reference Resolver.

Args:

duplicate_of: A map from duplicate names to preferred names of API

symbols.

doc_index: A `dict` mapping symbol name strings to objects with `url`

and `title` fields. Used to resolve @{$doc} references in docstrings.

is_class: A map from full names to bool for each symbol.

is_module: A map from full names to bool for each symbol.

py_module_names: A list of string names of Python modules.

"""

self._duplicate_of = duplicate_of

self._doc_index = doc_index

self._is_class = is_class

self._is_module = is_module

self._all_names = set(is_class.keys())

self._py_module_names = py_module_names

self.current_doc_full_name = None

self._errors = _Errors()

def add_error(self, message):

self._errors.append(self.current_doc_full_name, message)

def log_errors(self):

self._errors.log_all()

def num_errors(self):

return len(self._errors)

@classmethod

def from_visitor(cls, visitor, doc_index, **kwargs):

"""A factory function for building a ReferenceResolver from a visitor.

Args:

visitor: an instance of `DocGeneratorVisitor`

doc_index: a dictionary mapping document names to references objects with

"title" and "url" fields

**kwargs: all remaining args are passed to the constructor

Returns:

an instance of `ReferenceResolver` ()

"""

is_class = {

for name, obj in visitor.index.items()

}

is_module = {

for name, obj in visitor.index.items()

}

return cls(

duplicate_of=visitor.duplicate_of,

doc_index=doc_index,

is_class=is_class,

is_module=is_module,

**kwargs)

@classmethod

def from_json_file(cls, filepath, doc_index):

with open(filepath) as f:

json_dict = json.load(f)

return cls(doc_index=doc_index, **json_dict)

def to_json_file(self, filepath):

"""Converts the RefenceResolver to json and writes it to the specified file.

Args:

filepath: The file path to write the json to.

"""

json_dict = {}

for key, value in self.__dict__.items():

# Drop these two fields. `_doc_index` is not serializable. `_all_names` is

# generated by the constructor.

if key in ('_doc_index', '_all_names',

'_errors', 'current_doc_full_name'):

continue

# Strip off any leading underscores on field names as these are not

# recognized by the constructor.

json_dict[key.lstrip('_')] = value

with open(filepath, 'w') as f:

json.dump(json_dict, f)

def replace_references(self, string, relative_path_to_root):

"""Replace "@{symbol}" references with links to symbol's documentation page.

This functions finds all occurrences of "@{symbol}" in `string`

and replaces them with markdown links to the documentation page

for "symbol".

`relative_path_to_root` is the relative path from the document

that contains the "@{symbol}" reference to the root of the API

documentation that is linked to. If the containing page is part of

the same API docset, `relative_path_to_root` can be set to

`os.path.dirname(documentation_path(name))`, where `name` is the

python name of the object whose documentation page the reference

lives on.

Args:

string: A string in which "@{symbol}" references should be replaced.

relative_path_to_root: The relative path from the containing document to

the root of the API documentation that is being linked to.

Returns:

`string`, with "@{symbol}" references replaced by Markdown links.

"""

def strict_one_ref(match):

try:

return self._one_ref(match, relative_path_to_root)

except TFDocsError as e:

self.add_error(e.message)

return 'BAD_LINK'

string = re.sub(SYMBOL_REFERENCE_RE, strict_one_ref, string)

def sloppy_one_ref(match):

try:

return self._one_ref(match, relative_path_to_root)

except TFDocsError:

return match.group(0)

string = re.sub(AUTO_REFERENCE_RE, sloppy_one_ref, string)

return string

def python_link(self, link_text, ref_full_name, relative_path_to_root,

code_ref=True):

"""Resolve a "@{python symbol}" reference to a Markdown link.

This will pick the canonical location for duplicate symbols. The

input to this function should already be stripped of the '@' and

'{}'. This function returns a Markdown link. If `code_ref` is

true, it is assumed that this is a code reference, so the link

text will be rendered as code (using backticks).

`link_text` should refer to a library symbol, starting with 'tf.'.

Args:

link_text: The text of the Markdown link.

ref_full_name: The fully qualified name of the symbol to link to.

relative_path_to_root: The relative path from the location of the current

document to the root of the API documentation.

code_ref: If true (the default), put `link_text` in `...`.

Returns:

A markdown link to the documentation page of `ref_full_name`.

"""

url = self.reference_to_url(ref_full_name, relative_path_to_root)

if code_ref:

link_text = link_text.join(['<code>', '</code>'])

else:

link_text = self._link_text_to_html(link_text)

return '<a href="{}">{}</a>'.format(url, link_text)

@staticmethod

def _link_text_to_html(link_text):

code_re = '`(.*?)`'

return re.sub(code_re, r'<code>\1</code>', link_text)

def py_master_name(self, full_name):

"""Return the master name for a Python symbol name."""

return self._duplicate_of.get(full_name, full_name)

def reference_to_url(self, ref_full_name, relative_path_to_root):

"""Resolve a "@{python symbol}" reference to a relative path.

The input to this function should already be stripped of the '@'

and '{}', and its output is only the link, not the full Markdown.

If `ref_full_name` is the name of a class member, method, or property, the

link will point to the page of the containing class, and it will include the

method name as an anchor. For example, `tf.module.MyClass.my_method` will be

translated into a link to

`os.join.path(relative_path_to_root, 'tf/module/MyClass.md#my_method')`.

Args:

ref_full_name: The fully qualified name of the symbol to link to.

relative_path_to_root: The relative path from the location of the current

document to the root of the API documentation.

Returns:

A relative path that links from the documentation page of `from_full_name`

to the documentation page of `ref_full_name`.

Raises:

RuntimeError: If `ref_full_name` is not documented.

TFDocsError: If the @{} syntax cannot be decoded.

"""

master_name = self._duplicate_of.get(ref_full_name, ref_full_name)

# Check whether this link exists

if master_name not in self._all_names:

raise TFDocsError(

'Cannot make link to "%s": Not in index.' % master_name)

# If this is a member of a class, link to the class page with an anchor.

ref_path = None

if not (self._is_class[master_name] or self._is_module[master_name]):

idents = master_name.split('.')

if len(idents) > 1:

class_name = '.'.join(idents[:-1])

assert class_name in self._all_names

if self._is_class[class_name]:

ref_path = documentation_path(class_name) + '#%s' % idents[-1]

if not ref_path:

ref_path = documentation_path(master_name)

return os.path.join(relative_path_to_root, ref_path)

def _one_ref(self, match, relative_path_to_root):

"""Return a link for a single "@{symbol}" reference."""

string = match.group(1)

# Look for link text after $.

dollar = string.rfind('$')

if dollar > 0: # Ignore $ in first character

link_text = string[dollar + 1:]

string = string[:dollar]

manual_link_text = True

else:

link_text = string

manual_link_text = False

# Handle different types of references.

if string.startswith('$'): # Doc reference

return self._doc_link(string, link_text, manual_link_text,

relative_path_to_root)

elif string.startswith('tensorflow::'):

# C++ symbol

return self._cc_link(string, link_text, manual_link_text,

relative_path_to_root)

else:

is_python = False

for py_module_name in self._py_module_names:

if string == py_module_name or string.startswith(py_module_name + '.'):

is_python = True

break

if is_python: # Python symbol

return self.python_link(

link_text,

string,

relative_path_to_root,

code_ref=not manual_link_text)

# Error!

raise TFDocsError('Did not understand "%s"' % match.group(0),

'BROKEN_LINK')

def _doc_link(self, string, link_text, manual_link_text,

relative_path_to_root):

"""Generate a link for a @{$...} reference."""

string = string[1:] # remove leading $

# If string has a #, split that part into `hash_tag`

hash_pos = string.find('#')

if hash_pos > -1:

hash_tag = string[hash_pos:]

string = string[:hash_pos]

else:

hash_tag = ''

if string in self._doc_index:

if not manual_link_text: link_text = self._doc_index[string].title

url = os.path.normpath(os.path.join(

relative_path_to_root, '../..', self._doc_index[string].url))

link_text = self._link_text_to_html(link_text)

return '<a href="{}{}">{}</a>'.format(url, hash_tag, link_text)

return self._doc_missing(string, hash_tag, link_text, manual_link_text,

relative_path_to_root)

def _doc_missing(self, string, unused_hash_tag, unused_link_text,

unused_manual_link_text, unused_relative_path_to_root):

"""Generate an error for unrecognized @{$...} references."""

raise TFDocsError('Unknown Document "%s"' % string)

def _cc_link(self, string, link_text, unused_manual_link_text,

relative_path_to_root):

"""Generate a link for a @{tensorflow::...} reference."""

# TODO(josh11b): Fix this hard-coding of paths.

if string == 'tensorflow::ClientSession':

ret = 'class/tensorflow/client-session.md'

elif string == 'tensorflow::Scope':

ret = 'class/tensorflow/scope.md'

elif string == 'tensorflow::Status':

ret = 'class/tensorflow/status.md'

elif string == 'tensorflow::Tensor':

ret = 'class/tensorflow/tensor.md'

elif string == 'tensorflow::ops::Const':

ret = 'namespace/tensorflow/ops.md#const'

else:

raise TFDocsError('C++ reference not understood: "%s"' % string)

# relative_path_to_root gets you to api_docs/python, we go from there

# to api_docs/cc, and then add ret.

cc_relative_path = os.path.normpath(os.path.join(

relative_path_to_root, '../cc', ret))

return '<a href="{}"><code>{}</code></a>'.format(cc_relative_path,

link_text)

# TODO(aselle): Collect these into a big list for all modules and functions

# and make a rosetta stone page.

def _handle_compatibility(doc):

"""Parse and remove compatibility blocks from the main docstring.

Args:

doc: The docstring that contains compatibility notes"

Returns:

a tuple of the modified doc string and a hash that maps from compatibility

note type to the text of the note.

"""

compatibility_notes = {}

match_compatibility = re.compile(r'[ \t]*@compatibility$(\w+)$\s*\n'

r'((?:[^@\n]*\n)+)'

r'\s*@end_compatibility')

for f in match_compatibility.finditer(doc):

compatibility_notes[f.group(1)] = f.group(2)

return match_compatibility.subn(r'', doc)[0], compatibility_notes

def _gen_pairs(items):

"""Given an list of items [a,b,a,b...], generate pairs [(a,b),(a,b)...].

Args:

items: A list of items (length must be even)

Yields:

The original items, in pairs

"""

assert len(items) % 2 == 0

items = iter(items)

while True:

yield next(items), next(items)

class _FunctionDetail(

collections.namedtuple('_FunctionDetail', ['keyword', 'header', 'items'])):

"""A simple class to contain function details.

Composed of a "keyword", a possibly empty "header" string, and a possibly

empty

list of key-value pair "items".

"""

__slots__ = []

def __str__(self):

"""Return the original string that represents the function detail."""

parts = [self.keyword + ':\n']

parts.append(self.header)

for key, value in self.items:

parts.append(' ' + key + ': ')

parts.append(value)

return ''.join(parts)

def _parse_function_details(docstring):

r"""Given a docstring, split off the header and parse the function details.

For example the docstring of tf.nn.relu:

'''Computes rectified linear: `max(features, 0)`.

Args:

features: A `Tensor`. Must be one of the following types: `float32`,

`float64`, `int32`, `int64`, `uint8`, `int16`, `int8`, `uint16`,

`half`.

Returns:

A `Tensor`. Has the same type as `features`.

'''

This is parsed, and returned as:

```

('Computes rectified linear: `max(features, 0)`.\n\n', [

_FunctionDetail(

keyword='Args',

header='',

items=[

('features', ' A `Tensor`. Must be ...'),

('name', ' A name for the operation (optional).\n\n')]),

_FunctionDetail(

keyword='Returns',

header=' A `Tensor`. Has the same type as `features`.',

items=[])

])

```

Args:

docstring: The docstring to parse

Returns:

A (header, function_details) pair, where header is a string and

function_details is a (possibly empty) list of `_FunctionDetail` objects.

"""

detail_keywords = '|'.join([

'Args', 'Arguments', 'Fields', 'Returns', 'Yields', 'Raises', 'Attributes'

])

tag_re = re.compile('(?<=\n)(' + detail_keywords + '):\n', re.MULTILINE)

parts = tag_re.split(docstring)

# The first part is the main docstring

docstring = parts[0]

# Everything else alternates keyword-content

pairs = list(_gen_pairs(parts[1:]))

function_details = []

item_re = re.compile(r'^ ? ?(\*?\*?\w[\w.]*?\s*):\s', re.MULTILINE)

for keyword, content in pairs:

content = item_re.split(content)

header = content[0]

items = list(_gen_pairs(content[1:]))

function_details.append(_FunctionDetail(keyword, header, items))

return docstring, function_details

_DocstringInfo = collections.namedtuple('_DocstringInfo', [

'brief', 'docstring', 'function_details', 'compatibility'

])

def _parse_md_docstring(py_object, relative_path_to_root, reference_resolver):

"""Parse the object's docstring and return a `_DocstringInfo`.

This function clears @@'s from the docstring, and replaces @{} references

with markdown links.

For links within the same set of docs, the `relative_path_to_root` for a

docstring on the page for `full_name` can be set to:

```python

relative_path_to_root = os.path.relpath(

path='.', start=os.path.dirname(documentation_path(full_name)) or '.')

```

Args:

py_object: A python object to retrieve the docs for (class, function/method,

or module).

relative_path_to_root: The relative path from the location of the current

document to the root of the Python API documentation. This is used to

compute links for "@{symbol}" references.

reference_resolver: An instance of ReferenceResolver.

Returns:

A _DocstringInfo object, all fields will be empty if no docstring was found.

"""

# TODO(wicke): If this is a partial, use the .func docstring and add a note.

raw_docstring = _get_raw_docstring(py_object)

raw_docstring = reference_resolver.replace_references(

raw_docstring, relative_path_to_root)

atat_re = re.compile(r' *@@[a-zA-Z_.0-9]+ *$')

raw_docstring = '\n'.join(

line for line in raw_docstring.split('\n') if not atat_re.match(line))

docstring, compatibility = _handle_compatibility(raw_docstring)

docstring, function_details = _parse_function_details(docstring)

if 'Generated by: tensorflow/tools/api/generator' in docstring:

docstring = ''

return _DocstringInfo(

docstring.split('\n')[0], docstring, function_details, compatibility)

def _get_arg_spec(func):

"""Extracts signature information from a function or functools.partial object.

For functions, uses `tf_inspect.getfullargspec`. For `functools.partial`

objects, corrects the signature of the underlying function to take into

account the removed arguments.

Args:

func: A function whose signature to extract.

Returns:

An `FullArgSpec` namedtuple `(args, varargs, varkw, defaults, etc.)`,

as returned by `tf_inspect.getfullargspec`.

"""

# getfullargspec does not work for functools.partial objects directly.

if isinstance(func, functools.partial):

argspec = tf_inspect.getfullargspec(func.func)

# Remove the args from the original function that have been used up.

first_default_arg = (

len(argspec.args or []) - len(argspec.defaults or []))

partial_args = len(func.args)

argspec_args = []

if argspec.args:

argspec_args = list(argspec.args[partial_args:])

argspec_defaults = list(argspec.defaults or ())

if argspec.defaults and partial_args > first_default_arg:

argspec_defaults = list(argspec.defaults[partial_args-first_default_arg:])

first_default_arg = max(0, first_default_arg - partial_args)

for kwarg in (func.keywords or []):

if kwarg in (argspec.args or []):

i = argspec_args.index(kwarg)

argspec_args.pop(i)

if i >= first_default_arg:

argspec_defaults.pop(i-first_default_arg)

else:

first_default_arg -= 1

return tf_inspect.FullArgSpec(

args=argspec_args,

varargs=argspec.varargs,

varkw=argspec.varkw,

defaults=tuple(argspec_defaults),

kwonlyargs=[],

kwonlydefaults=None,

annotations={})

else: # Regular function or method, getargspec will work fine.

return tf_inspect.getfullargspec(func)

def _remove_first_line_indent(string):

indent = len(re.match(r'^\s*', string).group(0))

return '\n'.join([line[indent:] for line in string.split('\n')])

PAREN_NUMBER_RE = re.compile(r'^$([0-9.e-]+)$')

def _generate_signature(func, reverse_index):

"""Given a function, returns a list of strings representing its args.

This function produces a list of strings representing the arguments to a

python function. It uses tf_inspect.getfullargspec, which

does not generalize well to Python 3.x, which is more flexible in how *args

and **kwargs are handled. This is not a problem in TF, since we have to remain

compatible to Python 2.7 anyway.

This function uses `__name__` for callables if it is available. This can lead

to poor results for functools.partial and other callable objects.

The returned string is Python code, so if it is included in a Markdown

document, it should be typeset as code (using backticks), or escaped.

Args:

func: A function, method, or functools.partial to extract the signature for.

reverse_index: A map from object ids to canonical full names to use.

Returns:

A list of strings representing the argument signature of `func` as python

code.

"""

args_list = []

argspec = _get_arg_spec(func)

first_arg_with_default = (

len(argspec.args or []) - len(argspec.defaults or []))

# Python documentation skips `self` when printing method signatures.

# Note we cannot test for ismethod here since unbound methods do not register

# as methods (in Python 3).

first_arg = 1 if 'self' in argspec.args[:1] else 0

# Add all args without defaults.

for arg in argspec.args[first_arg:first_arg_with_default]:

args_list.append(arg)

# Add all args with defaults.

if argspec.defaults:

try:

source = _remove_first_line_indent(tf_inspect.getsource(func))

func_ast = ast.parse(source)

ast_defaults = func_ast.body[0].args.defaults

except IOError: # If this is a builtin, getsource fails with IOError

# If we cannot get the source, assume the AST would be equal to the repr

# of the defaults.

ast_defaults = [None] * len(argspec.defaults)

for arg, default, ast_default in zip(

argspec.args[first_arg_with_default:], argspec.defaults, ast_defaults):

if id(default) in reverse_index:

default_text = reverse_index[id(default)]

elif ast_default is not None:

default_text = (

astor.to_source(ast_default).rstrip('\n').replace('\t', '\\t')

.replace('\n', '\\n').replace('"""', "'"))

default_text = PAREN_NUMBER_RE.sub('\\1', default_text)

if default_text != repr(default):

# This may be an internal name. If so, handle the ones we know about.

# TODO(wicke): This should be replaced with a lookup in the index.

# TODO(wicke): (replace first ident with tf., check if in index)

internal_names = {

'ops.GraphKeys': 'tf.GraphKeys',

'_ops.GraphKeys': 'tf.GraphKeys',

'init_ops.zeros_initializer': 'tf.zeros_initializer',

'init_ops.ones_initializer': 'tf.ones_initializer',

'saver_pb2.SaverDef': 'tf.train.SaverDef',

}

full_name_re = '^%s(.%s)+' % (IDENTIFIER_RE, IDENTIFIER_RE)

match = re.match(full_name_re, default_text)

if match:

lookup_text = default_text

for internal_name, public_name in six.iteritems(internal_names):

if match.group(0).startswith(internal_name):

lookup_text = public_name + default_text[len(internal_name):]

break

if default_text is lookup_text:

print('WARNING: Using default arg, failed lookup: %s, repr: %r' %

(default_text, default))

else:

default_text = lookup_text

else:

default_text = repr(default)

args_list.append('%s=%s' % (arg, default_text))

# Add *args and *kwargs.

if argspec.varargs:

args_list.append('*' + argspec.varargs)

if argspec.varkw:

args_list.append('**' + argspec.varkw)

return args_list

def _get_guides_markdown(duplicate_names, guide_index, relative_path):

all_guides = []

for name in duplicate_names:

all_guides.extend(guide_index.get(name, []))

if not all_guides: return ''

prefix = '../' * (relative_path.count('/') + 3)

links = sorted(set([guide_ref.make_md_link(prefix)

for guide_ref in all_guides]))

return 'See the guide%s: %s\n\n' % (

's' if len(links) > 1 else '', ', '.join(links))

def _get_defining_class(py_class, name):

for cls in tf_inspect.getmro(py_class):

if name in cls.__dict__:

return cls

return None

class _LinkInfo(

collections.namedtuple(

'_LinkInfo', ['short_name', 'full_name', 'obj', 'doc', 'url'])):

__slots__ = []

def is_link(self):

return True

class _OtherMemberInfo(

collections.namedtuple('_OtherMemberInfo',

['short_name', 'full_name', 'obj', 'doc'])):

__slots__ = []

def is_link(self):

return False

_PropertyInfo = collections.namedtuple(

'_PropertyInfo', ['short_name', 'full_name', 'obj', 'doc'])

_MethodInfo = collections.namedtuple('_MethodInfo', [

'short_name', 'full_name', 'obj', 'doc', 'signature', 'decorators'

])

class _FunctionPageInfo(object):

"""Collects docs For a function Page."""

def __init__(self, full_name):

self._full_name = full_name

self._defined_in = None

self._aliases = None

self._doc = None

self._guides = None

self._signature = None

self._decorators = []

def for_function(self):

return True

def for_class(self):

return False

def for_module(self):

return False

@property

def full_name(self):

return self._full_name

@property

def short_name(self):

return self._full_name.split('.')[-1]

@property

def defined_in(self):

return self._defined_in

def set_defined_in(self, defined_in):

assert self.defined_in is None

self._defined_in = defined_in

@property

def aliases(self):

return self._aliases

def set_aliases(self, aliases):

assert self.aliases is None

self._aliases = aliases

@property

def doc(self):

return self._doc

def set_doc(self, doc):

assert self.doc is None

self._doc = doc

@property

def guides(self):

return self._guides

def set_guides(self, guides):

assert self.guides is None

self._guides = guides

@property

def signature(self):

return self._signature

def set_signature(self, function, reverse_index):

"""Attach the function's signature.

Args:

function: The python function being documented.

reverse_index: A map from object ids in the index to full names.

"""

assert self.signature is None

self._signature = _generate_signature(function, reverse_index)

@property

def decorators(self):

return list(self._decorators)

def add_decorator(self, dec):

self._decorators.append(dec)

def get_metadata_html(self):

return _Metadata(self.full_name).build_html()

class _ClassPageInfo(object):

"""Collects docs for a class page.

Attributes:

full_name: The fully qualified name of the object at the master

location. Aka `master_name`. For example: `tf.nn.sigmoid`.

short_name: The last component of the `full_name`. For example: `sigmoid`.

defined_in: The path to the file where this object is defined.

aliases: The list of all fully qualified names for the locations where the

object is visible in the public api. This includes the master location.

doc: A `_DocstringInfo` object representing the object's docstring (can be

created with `_parse_md_docstring`).

guides: A markdown string, of back links pointing to the api_guides that

reference this object.

bases: A list of `_LinkInfo` objects pointing to the docs for the parent

classes.

properties: A list of `_PropertyInfo` objects documenting the class'

properties (attributes that use `@property`).

methods: A list of `_MethodInfo` objects documenting the class' methods.

classes: A list of `_LinkInfo` objects pointing to docs for any nested

classes.

other_members: A list of `_OtherMemberInfo` objects documenting any other

object's defined inside the class object (mostly enum style fields).

"""

def __init__(self, full_name):

self._full_name = full_name

self._defined_in = None

self._aliases = None

self._doc = None

self._guides = None

self._bases = None

self._properties = []

self._methods = []

self._classes = []

self._other_members = []

def for_function(self):

"""Returns true if this object documents a function."""

return False

def for_class(self):

"""Returns true if this object documents a class."""

return True

def for_module(self):

"""Returns true if this object documents a module."""

return False

@property

def full_name(self):

"""Returns the documented object's fully qualified name."""

return self._full_name

@property

def short_name(self):

"""Returns the documented object's short name."""

return self._full_name.split('.')[-1]

@property

def defined_in(self):

"""Returns the path to the file where the documented object is defined."""

return self._defined_in

def set_defined_in(self, defined_in):

"""Sets the `defined_in` path."""

assert self.defined_in is None

self._defined_in = defined_in

@property

def aliases(self):

"""Returns a list of all full names for the documented object."""

return self._aliases

def set_aliases(self, aliases):

"""Sets the `aliases` list.

Args:

aliases: A list of strings. Containing all the object's full names.

"""

assert self.aliases is None

self._aliases = aliases

View remainder of file in raw view

tensorflow/tensorflow/tools/docs/parser.py at master · codeplaysoftware/tensorflow · GitHub

Repository navigation

Expand file tree

Breadcrumbs

parser.py

Latest commit

History

Breadcrumbs

parser.py

File metadata and controls