docutils/parsers/rst/directives/tables.py

   1 # $Id$
   2 # Authors: David Goodger <goodger@python.org>; David Priest
   3 # Copyright: This module has been placed in the public domain.
   4
   5 """
   6 Directives for table elements.
   7 """
   8
   9 __docformat__ = 'reStructuredText'
  10
  11
  12 import sys
  13 import os.path
  14 import csv
  15
  16 from docutils import io, nodes, statemachine, utils
  17 from docutils.error_reporting import SafeString
  18 from docutils.utils import SystemMessagePropagation
  19 from docutils.parsers.rst import Directive
  20 from docutils.parsers.rst import directives
  21
  22
  23 class Table(Directive):
  24
  25     """
  26     Generic table base class.
  27     """
  28
  29     required_arguments = 0
  30     optional_arguments = 1
  31     final_argument_whitespace = True
  32     option_spec = {'class': directives.class_option}
  33     has_content = True
  34
  35     def make_title(self):
  36         if self.arguments:
  37             title_text = self.arguments[0]
  38             text_nodes, messages = self.state.inline_text(title_text,
  39                                                           self.lineno)
  40             title = nodes.title(title_text, '', *text_nodes)
  41         else:
  42             title = None
  43             messages = []
  44         return title, messages
  45
  46     def process_header_option(self):
  47         source = self.state_machine.get_source(self.lineno - 1)
  48         table_head = []
  49         max_header_cols = 0
  50         if 'header' in self.options:   # separate table header in option
  51             rows, max_header_cols = self.parse_csv_data_into_rows(
  52                 self.options['header'].split('\n'), self.HeaderDialect(),
  53                 source)
  54             table_head.extend(rows)
  55         return table_head, max_header_cols
  56
  57     def check_table_dimensions(self, rows, header_rows, stub_columns):
  58         if len(rows) < header_rows:
  59             error = self.state_machine.reporter.error(
  60                 '%s header row(s) specified but only %s row(s) of data '
  61                 'supplied ("%s" directive).'
  62                 % (header_rows, len(rows), self.name), nodes.literal_block(
  63                 self.block_text, self.block_text), line=self.lineno)
  64             raise SystemMessagePropagation(error)
  65         if len(rows) == header_rows > 0:
  66             error = self.state_machine.reporter.error(
  67                 'Insufficient data supplied (%s row(s)); no data remaining '
  68                 'for table body, required by "%s" directive.'
  69                 % (len(rows), self.name), nodes.literal_block(
  70                 self.block_text, self.block_text), line=self.lineno)
  71             raise SystemMessagePropagation(error)
  72         for row in rows:
  73             if len(row) < stub_columns:
  74                 error = self.state_machine.reporter.error(
  75                     '%s stub column(s) specified but only %s columns(s) of '
  76                     'data supplied ("%s" directive).' %
  77                     (stub_columns, len(row), self.name), nodes.literal_block(
  78                     self.block_text, self.block_text), line=self.lineno)
  79                 raise SystemMessagePropagation(error)
  80             if len(row) == stub_columns > 0:
  81                 error = self.state_machine.reporter.error(
  82                     'Insufficient data supplied (%s columns(s)); no data remaining '
  83                     'for table body, required by "%s" directive.'
  84                     % (len(row), self.name), nodes.literal_block(
  85                     self.block_text, self.block_text), line=self.lineno)
  86                 raise SystemMessagePropagation(error)
  87
  88     def get_column_widths(self, max_cols):
  89         if 'widths' in self.options:
  90             col_widths = self.options['widths']
  91             if len(col_widths) != max_cols:
  92                 error = self.state_machine.reporter.error(
  93                     '"%s" widths do not match the number of columns in table '
  94                     '(%s).' % (self.name, max_cols), nodes.literal_block(
  95                     self.block_text, self.block_text), line=self.lineno)
  96                 raise SystemMessagePropagation(error)
  97         elif max_cols:
  98             col_widths = [100 // max_cols] * max_cols
  99         else:
 100             error = self.state_machine.reporter.error(
 101                 'No table data detected in CSV file.', nodes.literal_block(
 102                 self.block_text, self.block_text), line=self.lineno)
 103             raise SystemMessagePropagation(error)
 104         return col_widths
 105
 106     def extend_short_rows_with_empty_cells(self, columns, parts):
 107         for part in parts:
 108             for row in part:
 109                 if len(row) < columns:
 110                     row.extend([(0, 0, 0, [])] * (columns - len(row)))
 111
 112
 113 class RSTTable(Table):
 114
 115     def run(self):
 116         if not self.content:
 117             warning = self.state_machine.reporter.warning(
 118                 'Content block expected for the "%s" directive; none found.'
 119                 % self.name, nodes.literal_block(
 120                 self.block_text, self.block_text), line=self.lineno)
 121             return [warning]
 122         title, messages = self.make_title()
 123         node = nodes.Element()          # anonymous container for parsing
 124         self.state.nested_parse(self.content, self.content_offset, node)
 125         if len(node) != 1 or not isinstance(node[0], nodes.table):
 126             error = self.state_machine.reporter.error(
 127                 'Error parsing content block for the "%s" directive: exactly '
 128                 'one table expected.' % self.name, nodes.literal_block(
 129                 self.block_text, self.block_text), line=self.lineno)
 130             return [error]
 131         table_node = node[0]
 132         table_node['classes'] += self.options.get('class', [])
 133         if title:
 134             table_node.insert(0, title)
 135         return [table_node] + messages
 136
 137
 138 class CSVTable(Table):
 139
 140     option_spec = {'header-rows': directives.nonnegative_int,
 141                    'stub-columns': directives.nonnegative_int,
 142                    'header': directives.unchanged,
 143                    'widths': directives.positive_int_list,
 144                    'file': directives.path,
 145                    'url': directives.uri,
 146                    'encoding': directives.encoding,
 147                    'class': directives.class_option,
 148                    # field delimiter char
 149                    'delim': directives.single_char_or_whitespace_or_unicode,
 150                    # treat whitespace after delimiter as significant
 151                    'keepspace': directives.flag,
 152                    # text field quote/unquote char:
 153                    'quote': directives.single_char_or_unicode,
 154                    # char used to escape delim & quote as-needed:
 155                    'escape': directives.single_char_or_unicode,}
 156
 157     class DocutilsDialect(csv.Dialect):
 158
 159         """CSV dialect for `csv_table` directive."""
 160
 161         delimiter = ','
 162         quotechar = '"'
 163         doublequote = True
 164         skipinitialspace = True
 165         lineterminator = '\n'
 166         quoting = csv.QUOTE_MINIMAL
 167
 168         def __init__(self, options):
 169             if 'delim' in options:
 170                 self.delimiter = str(options['delim'])
 171             if 'keepspace' in options:
 172                 self.skipinitialspace = False
 173             if 'quote' in options:
 174                 self.quotechar = str(options['quote'])
 175             if 'escape' in options:
 176                 self.doublequote = False
 177                 self.escapechar = str(options['escape'])
 178             csv.Dialect.__init__(self)
 179
 180
 181     class HeaderDialect(csv.Dialect):
 182
 183         """CSV dialect to use for the "header" option data."""
 184
 185         delimiter = ','
 186         quotechar = '"'
 187         escapechar = '\\'
 188         doublequote = False
 189         skipinitialspace = True
 190         lineterminator = '\n'
 191         quoting = csv.QUOTE_MINIMAL
 192
 193     def check_requirements(self):
 194         pass
 195
 196     def run(self):
 197         try:
 198             if (not self.state.document.settings.file_insertion_enabled
 199                 and ('file' in self.options
 200                      or 'url' in self.options)):
 201                 warning = self.state_machine.reporter.warning(
 202                     'File and URL access deactivated; ignoring "%s" '
 203                     'directive.' % self.name, nodes.literal_block(
 204                     self.block_text, self.block_text), line=self.lineno)
 205                 return [warning]
 206             self.check_requirements()
 207             title, messages = self.make_title()
 208             csv_data, source = self.get_csv_data()
 209             table_head, max_header_cols = self.process_header_option()
 210             rows, max_cols = self.parse_csv_data_into_rows(
 211                 csv_data, self.DocutilsDialect(self.options), source)
 212             max_cols = max(max_cols, max_header_cols)
 213             header_rows = self.options.get('header-rows', 0)
 214             stub_columns = self.options.get('stub-columns', 0)
 215             self.check_table_dimensions(rows, header_rows, stub_columns)
 216             table_head.extend(rows[:header_rows])
 217             table_body = rows[header_rows:]
 218             col_widths = self.get_column_widths(max_cols)
 219             self.extend_short_rows_with_empty_cells(max_cols,
 220                                                     (table_head, table_body))
 221         except SystemMessagePropagation, detail:
 222             return [detail.args[0]]
 223         except csv.Error, detail:
 224             error = self.state_machine.reporter.error(
 225                 'Error with CSV data in "%s" directive:\n%s'
 226                 % (self.name, detail), nodes.literal_block(
 227                 self.block_text, self.block_text), line=self.lineno)
 228             return [error]
 229         table = (col_widths, table_head, table_body)
 230         table_node = self.state.build_table(table, self.content_offset,
 231                                             stub_columns)
 232         table_node['classes'] += self.options.get('class', [])
 233         if title:
 234             table_node.insert(0, title)
 235         return [table_node] + messages
 236
 237     def get_csv_data(self):
 238         """
 239         Get CSV data from the directive content, from an external
 240         file, or from a URL reference.
 241         """
 242         encoding = self.options.get(
 243             'encoding', self.state.document.settings.input_encoding)
 244         if self.content:
 245             # CSV data is from directive content.
 246             if 'file' in self.options or 'url' in self.options:
 247                 error = self.state_machine.reporter.error(
 248                     '"%s" directive may not both specify an external file and'
 249                     ' have content.' % self.name, nodes.literal_block(
 250                     self.block_text, self.block_text), line=self.lineno)
 251                 raise SystemMessagePropagation(error)
 252             source = self.content.source(0)
 253             csv_data = self.content
 254         elif 'file' in self.options:
 255             # CSV data is from an external file.
 256             if 'url' in self.options:
 257                 error = self.state_machine.reporter.error(
 258                       'The "file" and "url" options may not be simultaneously'
 259                       ' specified for the "%s" directive.' % self.name,
 260                       nodes.literal_block(self.block_text, self.block_text),
 261                       line=self.lineno)
 262                 raise SystemMessagePropagation(error)
 263             source_dir = os.path.dirname(
 264                 os.path.abspath(self.state.document.current_source))
 265             source = os.path.normpath(os.path.join(source_dir,
 266                                                    self.options['file']))
 267             source = utils.relative_path(None, source)
 268             try:
 269                 self.state.document.settings.record_dependencies.add(source)
 270                 csv_file = io.FileInput(
 271                     source_path=source, encoding=encoding,
 272                     error_handler=(self.state.document.settings.\
 273                                    input_encoding_error_handler),
 274                     handle_io_errors=None)
 275                 csv_data = csv_file.read().splitlines()
 276             except IOError, error:
 277                 severe = self.state_machine.reporter.severe(
 278                     u'Problems with "%s" directive path:\n%s.'
 279                     % (self.name, SafeString(error)),
 280                     nodes.literal_block(self.block_text, self.block_text),
 281                     line=self.lineno)
 282                 raise SystemMessagePropagation(severe)
 283         elif 'url' in self.options:
 284             # CSV data is from a URL.
 285             # Do not import urllib2 at the top of the module because
 286             # it may fail due to broken SSL dependencies, and it takes
 287             # about 0.15 seconds to load.
 288             import urllib2
 289             source = self.options['url']
 290             try:
 291                 csv_text = urllib2.urlopen(source).read()
 292             except (urllib2.URLError, IOError, OSError, ValueError), error:
 293                 severe = self.state_machine.reporter.severe(
 294                       'Problems with "%s" directive URL "%s":\n%s.'
 295                       % (self.name, self.options['url'], SafeString(error)),
 296                       nodes.literal_block(self.block_text, self.block_text),
 297                       line=self.lineno)
 298                 raise SystemMessagePropagation(severe)
 299             csv_file = io.StringInput(
 300                 source=csv_text, source_path=source, encoding=encoding,
 301                 error_handler=(self.state.document.settings.\
 302                                input_encoding_error_handler))
 303             csv_data = csv_file.read().splitlines()
 304         else:
 305             error = self.state_machine.reporter.warning(
 306                 'The "%s" directive requires content; none supplied.'
 307                 % self.name, nodes.literal_block(
 308                 self.block_text, self.block_text), line=self.lineno)
 309             raise SystemMessagePropagation(error)
 310         return csv_data, source
 311
 312     if sys.version_info < (3,):
 313         # 2.x csv module doesn't do Unicode
 314         def decode_from_csv(s):
 315             return s.decode('utf-8')
 316         def encode_for_csv(s):
 317             return s.encode('utf-8')
 318     else:
 319         def decode_from_csv(s):
 320             return s
 321         def encode_for_csv(s):
 322             return s
 323     decode_from_csv = staticmethod(decode_from_csv)
 324     encode_for_csv = staticmethod(encode_for_csv)
 325
 326     def parse_csv_data_into_rows(self, csv_data, dialect, source):
 327         # csv.py doesn't do Unicode; encode temporarily as UTF-8
 328         csv_reader = csv.reader([self.encode_for_csv(line + '\n')
 329                                  for line in csv_data],
 330                                 dialect=dialect)
 331         rows = []
 332         max_cols = 0
 333         for row in csv_reader:
 334             row_data = []
 335             for cell in row:
 336                 # decode UTF-8 back to Unicode
 337                 cell_text = self.decode_from_csv(cell)
 338                 cell_data = (0, 0, 0, statemachine.StringList(
 339                     cell_text.splitlines(), source=source))
 340                 row_data.append(cell_data)
 341             rows.append(row_data)
 342             max_cols = max(max_cols, len(row))
 343         return rows, max_cols
 344
 345
 346 class ListTable(Table):
 347
 348     """
 349     Implement tables whose data is encoded as a uniform two-level bullet list.
 350     For further ideas, see
 351     http://docutils.sf.net/docs/dev/rst/alternatives.html#list-driven-tables
 352     """
 353
 354     option_spec = {'header-rows': directives.nonnegative_int,
 355                    'stub-columns': directives.nonnegative_int,
 356                    'widths': directives.positive_int_list,
 357                    'class': directives.class_option}
 358
 359     def run(self):
 360         if not self.content:
 361             error = self.state_machine.reporter.error(
 362                 'The "%s" directive is empty; content required.' % self.name,
 363                 nodes.literal_block(self.block_text, self.block_text),
 364                 line=self.lineno)
 365             return [error]
 366         title, messages = self.make_title()
 367         node = nodes.Element()          # anonymous container for parsing
 368         self.state.nested_parse(self.content, self.content_offset, node)
 369         try:
 370             num_cols, col_widths = self.check_list_content(node)
 371             table_data = [[item.children for item in row_list[0]]
 372                           for row_list in node[0]]
 373             header_rows = self.options.get('header-rows', 0)
 374             stub_columns = self.options.get('stub-columns', 0)
 375             self.check_table_dimensions(table_data, header_rows, stub_columns)
 376         except SystemMessagePropagation, detail:
 377             return [detail.args[0]]
 378         table_node = self.build_table_from_list(table_data, col_widths,
 379                                                 header_rows, stub_columns)
 380         table_node['classes'] += self.options.get('class', [])
 381         if title:
 382             table_node.insert(0, title)
 383         return [table_node] + messages
 384
 385     def check_list_content(self, node):
 386         if len(node) != 1 or not isinstance(node[0], nodes.bullet_list):
 387             error = self.state_machine.reporter.error(
 388                 'Error parsing content block for the "%s" directive: '
 389                 'exactly one bullet list expected.' % self.name,
 390                 nodes.literal_block(self.block_text, self.block_text),
 391                 line=self.lineno)
 392             raise SystemMessagePropagation(error)
 393         list_node = node[0]
 394         # Check for a uniform two-level bullet list:
 395         for item_index in range(len(list_node)):
 396             item = list_node[item_index]
 397             if len(item) != 1 or not isinstance(item[0], nodes.bullet_list):
 398                 error = self.state_machine.reporter.error(
 399                     'Error parsing content block for the "%s" directive: '
 400                     'two-level bullet list expected, but row %s does not '
 401                     'contain a second-level bullet list.'
 402                     % (self.name, item_index + 1), nodes.literal_block(
 403                     self.block_text, self.block_text), line=self.lineno)
 404                 raise SystemMessagePropagation(error)
 405             elif item_index:
 406                 # ATTN pychecker users: num_cols is guaranteed to be set in the
 407                 # "else" clause below for item_index==0, before this branch is
 408                 # triggered.
 409                 if len(item[0]) != num_cols:
 410                     error = self.state_machine.reporter.error(
 411                         'Error parsing content block for the "%s" directive: '
 412                         'uniform two-level bullet list expected, but row %s '
 413                         'does not contain the same number of items as row 1 '
 414                         '(%s vs %s).'
 415                         % (self.name, item_index + 1, len(item[0]), num_cols),
 416                         nodes.literal_block(self.block_text, self.block_text),
 417                         line=self.lineno)
 418                     raise SystemMessagePropagation(error)
 419             else:
 420                 num_cols = len(item[0])
 421         col_widths = self.get_column_widths(num_cols)
 422         return num_cols, col_widths
 423
 424     def build_table_from_list(self, table_data, col_widths, header_rows, stub_columns):
 425         table = nodes.table()
 426         tgroup = nodes.tgroup(cols=len(col_widths))
 427         table += tgroup
 428         for col_width in col_widths:
 429             colspec = nodes.colspec(colwidth=col_width)
 430             if stub_columns:
 431                 colspec.attributes['stub'] = 1
 432                 stub_columns -= 1
 433             tgroup += colspec
 434         rows = []
 435         for row in table_data:
 436             row_node = nodes.row()
 437             for cell in row:
 438                 entry = nodes.entry()
 439                 entry += cell
 440                 row_node += entry
 441             rows.append(row_node)
 442         if header_rows:
 443             thead = nodes.thead()
 444             thead.extend(rows[:header_rows])
 445             tgroup += thead
 446         tbody = nodes.tbody()
 447         tbody.extend(rows[header_rows:])
 448         tgroup += tbody
 449         return table