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.utils.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     optional_arguments = 1
  30     final_argument_whitespace = True
  31     option_spec = {'class': directives.class_option,
  32                    'name': directives.unchanged}
  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         self.add_name(table_node)
 134         if title:
 135             table_node.insert(0, title)
 136         return [table_node] + messages
 137
 138
 139 class CSVTable(Table):
 140
 141     option_spec = {'header-rows': directives.nonnegative_int,
 142                    'stub-columns': directives.nonnegative_int,
 143                    'header': directives.unchanged,
 144                    'widths': directives.positive_int_list,
 145                    'file': directives.path,
 146                    'url': directives.uri,
 147                    'encoding': directives.encoding,
 148                    'class': directives.class_option,
 149                    'name': directives.unchanged,
 150                    # field delimiter char
 151                    'delim': directives.single_char_or_whitespace_or_unicode,
 152                    # treat whitespace after delimiter as significant
 153                    'keepspace': directives.flag,
 154                    # text field quote/unquote char:
 155                    'quote': directives.single_char_or_unicode,
 156                    # char used to escape delim & quote as-needed:
 157                    'escape': directives.single_char_or_unicode,}
 158
 159     class DocutilsDialect(csv.Dialect):
 160
 161         """CSV dialect for `csv_table` directive."""
 162
 163         delimiter = ','
 164         quotechar = '"'
 165         doublequote = True
 166         skipinitialspace = True
 167         lineterminator = '\n'
 168         quoting = csv.QUOTE_MINIMAL
 169
 170         def __init__(self, options):
 171             if 'delim' in options:
 172                 self.delimiter = str(options['delim'])
 173             if 'keepspace' in options:
 174                 self.skipinitialspace = False
 175             if 'quote' in options:
 176                 self.quotechar = str(options['quote'])
 177             if 'escape' in options:
 178                 self.doublequote = False
 179                 self.escapechar = str(options['escape'])
 180             csv.Dialect.__init__(self)
 181
 182
 183     class HeaderDialect(csv.Dialect):
 184
 185         """CSV dialect to use for the "header" option data."""
 186
 187         delimiter = ','
 188         quotechar = '"'
 189         escapechar = '\\'
 190         doublequote = False
 191         skipinitialspace = True
 192         lineterminator = '\n'
 193         quoting = csv.QUOTE_MINIMAL
 194
 195     def check_requirements(self):
 196         pass
 197
 198     def run(self):
 199         try:
 200             if (not self.state.document.settings.file_insertion_enabled
 201                 and ('file' in self.options
 202                      or 'url' in self.options)):
 203                 warning = self.state_machine.reporter.warning(
 204                     'File and URL access deactivated; ignoring "%s" '
 205                     'directive.' % self.name, nodes.literal_block(
 206                     self.block_text, self.block_text), line=self.lineno)
 207                 return [warning]
 208             self.check_requirements()
 209             title, messages = self.make_title()
 210             csv_data, source = self.get_csv_data()
 211             table_head, max_header_cols = self.process_header_option()
 212             rows, max_cols = self.parse_csv_data_into_rows(
 213                 csv_data, self.DocutilsDialect(self.options), source)
 214             max_cols = max(max_cols, max_header_cols)
 215             header_rows = self.options.get('header-rows', 0)
 216             stub_columns = self.options.get('stub-columns', 0)
 217             self.check_table_dimensions(rows, header_rows, stub_columns)
 218             table_head.extend(rows[:header_rows])
 219             table_body = rows[header_rows:]
 220             col_widths = self.get_column_widths(max_cols)
 221             self.extend_short_rows_with_empty_cells(max_cols,
 222                                                     (table_head, table_body))
 223         except SystemMessagePropagation, detail:
 224             return [detail.args[0]]
 225         except csv.Error, detail:
 226             error = self.state_machine.reporter.error(
 227                 'Error with CSV data in "%s" directive:\n%s'
 228                 % (self.name, detail), nodes.literal_block(
 229                 self.block_text, self.block_text), line=self.lineno)
 230             return [error]
 231         table = (col_widths, table_head, table_body)
 232         table_node = self.state.build_table(table, self.content_offset,
 233                                             stub_columns)
 234         table_node['classes'] += self.options.get('class', [])
 235         self.add_name(table_node)
 236         if title:
 237             table_node.insert(0, title)
 238         return [table_node] + messages
 239
 240     def get_csv_data(self):
 241         """
 242         Get CSV data from the directive content, from an external
 243         file, or from a URL reference.
 244         """
 245         encoding = self.options.get(
 246             'encoding', self.state.document.settings.input_encoding)
 247         if self.content:
 248             # CSV data is from directive content.
 249             if 'file' in self.options or 'url' in self.options:
 250                 error = self.state_machine.reporter.error(
 251                     '"%s" directive may not both specify an external file and'
 252                     ' have content.' % self.name, nodes.literal_block(
 253                     self.block_text, self.block_text), line=self.lineno)
 254                 raise SystemMessagePropagation(error)
 255             source = self.content.source(0)
 256             csv_data = self.content
 257         elif 'file' in self.options:
 258             # CSV data is from an external file.
 259             if 'url' in self.options:
 260                 error = self.state_machine.reporter.error(
 261                       'The "file" and "url" options may not be simultaneously'
 262                       ' specified for the "%s" directive.' % self.name,
 263                       nodes.literal_block(self.block_text, self.block_text),
 264                       line=self.lineno)
 265                 raise SystemMessagePropagation(error)
 266             source_dir = os.path.dirname(
 267                 os.path.abspath(self.state.document.current_source))
 268             source = os.path.normpath(os.path.join(source_dir,
 269                                                    self.options['file']))
 270             source = utils.relative_path(None, source)
 271             try:
 272                 self.state.document.settings.record_dependencies.add(source)
 273                 csv_file = io.FileInput(
 274                     source_path=source, encoding=encoding,
 275                     error_handler=(self.state.document.settings.\
 276                                    input_encoding_error_handler),
 277                     handle_io_errors=None)
 278                 csv_data = csv_file.read().splitlines()
 279             except IOError, error:
 280                 severe = self.state_machine.reporter.severe(
 281                     u'Problems with "%s" directive path:\n%s.'
 282                     % (self.name, SafeString(error)),
 283                     nodes.literal_block(self.block_text, self.block_text),
 284                     line=self.lineno)
 285                 raise SystemMessagePropagation(severe)
 286         elif 'url' in self.options:
 287             # CSV data is from a URL.
 288             # Do not import urllib2 at the top of the module because
 289             # it may fail due to broken SSL dependencies, and it takes
 290             # about 0.15 seconds to load.
 291             import urllib2
 292             source = self.options['url']
 293             try:
 294                 csv_text = urllib2.urlopen(source).read()
 295             except (urllib2.URLError, IOError, OSError, ValueError), error:
 296                 severe = self.state_machine.reporter.severe(
 297                       'Problems with "%s" directive URL "%s":\n%s.'
 298                       % (self.name, self.options['url'], SafeString(error)),
 299                       nodes.literal_block(self.block_text, self.block_text),
 300                       line=self.lineno)
 301                 raise SystemMessagePropagation(severe)
 302             csv_file = io.StringInput(
 303                 source=csv_text, source_path=source, encoding=encoding,
 304                 error_handler=(self.state.document.settings.\
 305                                input_encoding_error_handler))
 306             csv_data = csv_file.read().splitlines()
 307         else:
 308             error = self.state_machine.reporter.warning(
 309                 'The "%s" directive requires content; none supplied.'
 310                 % self.name, nodes.literal_block(
 311                 self.block_text, self.block_text), line=self.lineno)
 312             raise SystemMessagePropagation(error)
 313         return csv_data, source
 314
 315     if sys.version_info < (3,):
 316         # 2.x csv module doesn't do Unicode
 317         def decode_from_csv(s):
 318             return s.decode('utf-8')
 319         def encode_for_csv(s):
 320             return s.encode('utf-8')
 321     else:
 322         def decode_from_csv(s):
 323             return s
 324         def encode_for_csv(s):
 325             return s
 326     decode_from_csv = staticmethod(decode_from_csv)
 327     encode_for_csv = staticmethod(encode_for_csv)
 328
 329     def parse_csv_data_into_rows(self, csv_data, dialect, source):
 330         # csv.py doesn't do Unicode; encode temporarily as UTF-8
 331         csv_reader = csv.reader([self.encode_for_csv(line + '\n')
 332                                  for line in csv_data],
 333                                 dialect=dialect)
 334         rows = []
 335         max_cols = 0
 336         for row in csv_reader:
 337             row_data = []
 338             for cell in row:
 339                 # decode UTF-8 back to Unicode
 340                 cell_text = self.decode_from_csv(cell)
 341                 cell_data = (0, 0, 0, statemachine.StringList(
 342                     cell_text.splitlines(), source=source))
 343                 row_data.append(cell_data)
 344             rows.append(row_data)
 345             max_cols = max(max_cols, len(row))
 346         return rows, max_cols
 347
 348
 349 class ListTable(Table):
 350
 351     """
 352     Implement tables whose data is encoded as a uniform two-level bullet list.
 353     For further ideas, see
 354     http://docutils.sf.net/docs/dev/rst/alternatives.html#list-driven-tables
 355     """
 356
 357     option_spec = {'header-rows': directives.nonnegative_int,
 358                    'stub-columns': directives.nonnegative_int,
 359                    'widths': directives.positive_int_list,
 360                    'class': directives.class_option,
 361                    'name': directives.unchanged}
 362
 363     def run(self):
 364         if not self.content:
 365             error = self.state_machine.reporter.error(
 366                 'The "%s" directive is empty; content required.' % self.name,
 367                 nodes.literal_block(self.block_text, self.block_text),
 368                 line=self.lineno)
 369             return [error]
 370         title, messages = self.make_title()
 371         node = nodes.Element()          # anonymous container for parsing
 372         self.state.nested_parse(self.content, self.content_offset, node)
 373         try:
 374             num_cols, col_widths = self.check_list_content(node)
 375             table_data = [[item.children for item in row_list[0]]
 376                           for row_list in node[0]]
 377             header_rows = self.options.get('header-rows', 0)
 378             stub_columns = self.options.get('stub-columns', 0)
 379             self.check_table_dimensions(table_data, header_rows, stub_columns)
 380         except SystemMessagePropagation, detail:
 381             return [detail.args[0]]
 382         table_node = self.build_table_from_list(table_data, col_widths,
 383                                                 header_rows, stub_columns)
 384         table_node['classes'] += self.options.get('class', [])
 385         self.add_name(table_node)
 386         if title:
 387             table_node.insert(0, title)
 388         return [table_node] + messages
 389
 390     def check_list_content(self, node):
 391         if len(node) != 1 or not isinstance(node[0], nodes.bullet_list):
 392             error = self.state_machine.reporter.error(
 393                 'Error parsing content block for the "%s" directive: '
 394                 'exactly one bullet list expected.' % self.name,
 395                 nodes.literal_block(self.block_text, self.block_text),
 396                 line=self.lineno)
 397             raise SystemMessagePropagation(error)
 398         list_node = node[0]
 399         # Check for a uniform two-level bullet list:
 400         for item_index in range(len(list_node)):
 401             item = list_node[item_index]
 402             if len(item) != 1 or not isinstance(item[0], nodes.bullet_list):
 403                 error = self.state_machine.reporter.error(
 404                     'Error parsing content block for the "%s" directive: '
 405                     'two-level bullet list expected, but row %s does not '
 406                     'contain a second-level bullet list.'
 407                     % (self.name, item_index + 1), nodes.literal_block(
 408                     self.block_text, self.block_text), line=self.lineno)
 409                 raise SystemMessagePropagation(error)
 410             elif item_index:
 411                 # ATTN pychecker users: num_cols is guaranteed to be set in the
 412                 # "else" clause below for item_index==0, before this branch is
 413                 # triggered.
 414                 if len(item[0]) != num_cols:
 415                     error = self.state_machine.reporter.error(
 416                         'Error parsing content block for the "%s" directive: '
 417                         'uniform two-level bullet list expected, but row %s '
 418                         'does not contain the same number of items as row 1 '
 419                         '(%s vs %s).'
 420                         % (self.name, item_index + 1, len(item[0]), num_cols),
 421                         nodes.literal_block(self.block_text, self.block_text),
 422                         line=self.lineno)
 423                     raise SystemMessagePropagation(error)
 424             else:
 425                 num_cols = len(item[0])
 426         col_widths = self.get_column_widths(num_cols)
 427         return num_cols, col_widths
 428
 429     def build_table_from_list(self, table_data, col_widths, header_rows, stub_columns):
 430         table = nodes.table()
 431         tgroup = nodes.tgroup(cols=len(col_widths))
 432         table += tgroup
 433         for col_width in col_widths:
 434             colspec = nodes.colspec(colwidth=col_width)
 435             if stub_columns:
 436                 colspec.attributes['stub'] = 1
 437                 stub_columns -= 1
 438             tgroup += colspec
 439         rows = []
 440         for row in table_data:
 441             row_node = nodes.row()
 442             for cell in row:
 443                 entry = nodes.entry()
 444                 entry += cell
 445                 row_node += entry
 446             rows.append(row_node)
 447         if header_rows:
 448             thead = nodes.thead()
 449             thead.extend(rows[:header_rows])
 450             tgroup += thead
 451         tbody = nodes.tbody()
 452         tbody.extend(rows[header_rows:])
 453         tgroup += tbody
 454         return table