# $Id: tables.py 9363 2023-04-24 16:37:37Z aa-turner $
# Authors: David Goodger <goodger@python.org>; David Priest
# Copyright: This module has been placed in the public domain.

"""
Directives for table elements.
"""

__docformat__ = 'reStructuredText'


import csv
from pathlib import Path
import warnings

from docutils import io, nodes, statemachine, utils
from docutils.utils import SystemMessagePropagation
from docutils.parsers.rst import Directive
from docutils.parsers.rst import directives
from urllib.request import urlopen
from urllib.error import URLError


def align(argument):
    return directives.choice(argument, ('left', 'center', 'right'))


class Table(Directive):

    """
    Generic table base class.
    """

    optional_arguments = 1
    final_argument_whitespace = True
    option_spec = {'class': directives.class_option,
                   'name': directives.unchanged,
                   'align': align,
                   'width': directives.length_or_percentage_or_unitless,
                   'widths': directives.value_or(('auto', 'grid'),
                                                 directives.positive_int_list)}
    has_content = True

    def make_title(self):
        if self.arguments:
            title_text = self.arguments[0]
            text_nodes, messages = self.state.inline_text(title_text,
                                                          self.lineno)
            title = nodes.title(title_text, '', *text_nodes)
            (title.source,
             title.line) = self.state_machine.get_source_and_line(self.lineno)
        else:
            title = None
            messages = []
        return title, messages

    def process_header_option(self):
        # Provisional
        # * Will move to CSVTable in Docutils 0.21
        #   as it calls `self.HeaderDialect()` only defined in CSVTable.
        # * Will change to use the same CSV dialect as the body to get in line
        #   with the specification in ref/rst/directives.txt in Docutils 0.21.
        source = self.state_machine.get_source(self.lineno - 1)
        table_head = []
        max_header_cols = 0
        if 'header' in self.options:   # separate table header in option
            with warnings.catch_warnings():
                warnings.simplefilter('ignore')
                header_dialect = self.HeaderDialect()
            rows, max_header_cols = self.parse_csv_data_into_rows(
                self.options['header'].split('\n'), header_dialect,
                source)
            table_head.extend(rows)
        return table_head, max_header_cols

    def check_table_dimensions(self, rows, header_rows, stub_columns):
        if len(rows) < header_rows:
            error = self.reporter.error('%s header row(s) specified but '
                'only %s row(s) of data supplied ("%s" directive).'
                % (header_rows, len(rows), self.name),
                nodes.literal_block(self.block_text, self.block_text),
                line=self.lineno)
            raise SystemMessagePropagation(error)
        if len(rows) == header_rows > 0:
            error = self.reporter.error(
                f'Insufficient data supplied ({len(rows)} row(s)); '
                'no data remaining for table body, '
                f'required by "{self.name}" directive.',
                nodes.literal_block(self.block_text, self.block_text),
                line=self.lineno)
            raise SystemMessagePropagation(error)
        for row in rows:
            if len(row) < stub_columns:
                error = self.reporter.error(
                    f'{stub_columns} stub column(s) specified '
                    f'but only {len(row)} columns(s) of data supplied '
                    f'("{self.name}" directive).',
                    nodes.literal_block(self.block_text, self.block_text),
                    line=self.lineno)
                raise SystemMessagePropagation(error)
            if len(row) == stub_columns > 0:
                error = self.reporter.error(
                    'Insufficient data supplied (%s columns(s)); '
                    'no data remaining for table body, required '
                    'by "%s" directive.' % (len(row), self.name),
                    nodes.literal_block(self.block_text, self.block_text),
                    line=self.lineno)
                raise SystemMessagePropagation(error)

    def set_table_width(self, table_node):
        if 'width' in self.options:
            table_node['width'] = self.options.get('width')

    @property
    def widths(self):
        return self.options.get('widths', '')

    def get_column_widths(self, n_cols):
        if isinstance(self.widths, list):
            if len(self.widths) != n_cols:
                # TODO: use last value for missing columns?
                error = self.reporter.error('"%s" widths do not match the '
                    'number of columns in table (%s).' % (self.name, n_cols),
                    nodes.literal_block(self.block_text, self.block_text),
                    line=self.lineno)
                raise SystemMessagePropagation(error)
            col_widths = self.widths
        elif n_cols:
            col_widths = [100 // n_cols] * n_cols
        else:
            error = self.reporter.error('No table data detected in CSV file.',
                nodes.literal_block(self.block_text, self.block_text),
                line=self.lineno)
            raise SystemMessagePropagation(error)
        return col_widths

    def extend_short_rows_with_empty_cells(self, columns, parts):
        for part in parts:
            for row in part:
                if len(row) < columns:
                    row.extend([(0, 0, 0, [])] * (columns - len(row)))


class RSTTable(Table):
    """
    Class for the `"table" directive`__ for formal tables using rST syntax.

    __ https://docutils.sourceforge.io/docs/ref/rst/directives.html
    """

    def run(self):
        if not self.content:
            warning = self.reporter.warning('Content block expected '
                'for the "%s" directive; none found.' % self.name,
                nodes.literal_block(self.block_text, self.block_text),
                line=self.lineno)
            return [warning]
        title, messages = self.make_title()
        node = nodes.Element()          # anonymous container for parsing
        self.state.nested_parse(self.content, self.content_offset, node)
        if len(node) != 1 or not isinstance(node[0], nodes.table):
            error = self.reporter.error('Error parsing content block for the '
                '"%s" directive: exactly one table expected.' % self.name,
                nodes.literal_block(self.block_text, self.block_text),
                line=self.lineno)
            return [error]
        table_node = node[0]
        table_node['classes'] += self.options.get('class', [])
        self.set_table_width(table_node)
        if 'align' in self.options:
            table_node['align'] = self.options.get('align')
        if isinstance(self.widths, list):
            tgroup = table_node[0]
            try:
                col_widths = self.get_column_widths(tgroup["cols"])
            except SystemMessagePropagation as detail:
                return [detail.args[0]]
            colspecs = [child for child in tgroup.children
                        if child.tagname == 'colspec']
            for colspec, col_width in zip(colspecs, col_widths):
                colspec['colwidth'] = col_width
        if self.widths == 'auto':
            table_node['classes'] += ['colwidths-auto']
        elif self.widths:  # "grid" or list of integers
            table_node['classes'] += ['colwidths-given']
        self.add_name(table_node)
        if title:
            table_node.insert(0, title)
        return [table_node] + messages


class CSVTable(Table):

    option_spec = {'header-rows': directives.nonnegative_int,
                   'stub-columns': directives.nonnegative_int,
                   'header': directives.unchanged,
                   'width': directives.length_or_percentage_or_unitless,
                   'widths': directives.value_or(('auto', ),
                                                 directives.positive_int_list),
                   'file': directives.path,
                   'url': directives.uri,
                   'encoding': directives.encoding,
                   'class': directives.class_option,
                   'name': directives.unchanged,
                   'align': align,
                   # field delimiter char
                   'delim': directives.single_char_or_whitespace_or_unicode,
                   # treat whitespace after delimiter as significant
                   'keepspace': directives.flag,
                   # text field quote/unquote char:
                   'quote': directives.single_char_or_unicode,
                   # char used to escape delim & quote as-needed:
                   'escape': directives.single_char_or_unicode}

    class DocutilsDialect(csv.Dialect):

        """CSV dialect for `csv_table` directive."""

        delimiter = ','
        quotechar = '"'
        doublequote = True
        skipinitialspace = True
        strict = True
        lineterminator = '\n'
        quoting = csv.QUOTE_MINIMAL

        def __init__(self, options):
            if 'delim' in options:
                self.delimiter = options['delim']
            if 'keepspace' in options:
                self.skipinitialspace = False
            if 'quote' in options:
                self.quotechar = options['quote']
            if 'escape' in options:
                self.doublequote = False
                self.escapechar = options['escape']
            super().__init__()

    class HeaderDialect(csv.Dialect):
        """
        CSV dialect used for the "header" option data.

        Deprecated. Will be removed in Docutils 0.22.
        """
        # The separate HeaderDialect was introduced in revision 2294
        # (2004-06-17) in the sandbox before the "csv-table" directive moved
        # to the trunk in r2309. Discussion in docutils-devel around this time
        # did not mention a rationale (part of the discussion was in private
        # mail).
        # This is in conflict with the documentation, which always said:
        # ""
        # and did not change in this aspect.
        #
        # Maybe it was intended to have similar escape rules for rST and CSV,
        # however with the current implementation this means we need
        # `\\` for rST markup and ``\\\\`` for a literal backslash
        # in the "option" header but ``\`` and ``\\`` in the header-lines and
        # table cells of the main CSV data.
        delimiter = ','
        quotechar = '"'
        escapechar = '\\'
        doublequote = False
        skipinitialspace = True
        strict = True
        lineterminator = '\n'
        quoting = csv.QUOTE_MINIMAL

        def __init__(self):
            warnings.warn('CSVTable.HeaderDialect will be removed '
                          'in Docutils 0.22.',
                          PendingDeprecationWarning, stacklevel=2)
            super().__init__()

    @staticmethod
    def check_requirements():
        warnings.warn('CSVTable.check_requirements()'
                      ' is not required with Python 3'
                      ' and will be removed in Docutils 0.22.',
                      DeprecationWarning, stacklevel=2)

    def run(self):
        try:
            if (not self.state.document.settings.file_insertion_enabled
                and ('file' in self.options
                     or 'url' in self.options)):
                warning = self.reporter.warning('File and URL access '
                    'deactivated; ignoring "%s" directive.' % self.name,
                    nodes.literal_block(self.block_text, self.block_text),
                    line=self.lineno)
                return [warning]
            title, messages = self.make_title()
            csv_data, source = self.get_csv_data()
            table_head, max_header_cols = self.process_header_option()
            rows, max_cols = self.parse_csv_data_into_rows(
                csv_data, self.DocutilsDialect(self.options), source)
            max_cols = max(max_cols, max_header_cols)
            header_rows = self.options.get('header-rows', 0)
            stub_columns = self.options.get('stub-columns', 0)
            self.check_table_dimensions(rows, header_rows, stub_columns)
            table_head.extend(rows[:header_rows])
            table_body = rows[header_rows:]
            col_widths = self.get_column_widths(max_cols)
            self.extend_short_rows_with_empty_cells(max_cols,
                                                    (table_head, table_body))
        except SystemMessagePropagation as detail:
            return [detail.args[0]]
        except csv.Error as detail:
            message = str(detail)
            error = self.reporter.error('Error with CSV data'
                ' in "%s" directive:\n%s' % (self.name, message),
                nodes.literal_block(self.block_text, self.block_text),
                line=self.lineno)
            return [error]
        table = (col_widths, table_head, table_body)
        table_node = self.state.build_table(table, self.content_offset,
                                            stub_columns, widths=self.widths)
        table_node['classes'] += self.options.get('class', [])
        if 'align' in self.options:
            table_node['align'] = self.options.get('align')
        self.set_table_width(table_node)
        self.add_name(table_node)
        if title:
            table_node.insert(0, title)
        return [table_node] + messages

    def get_csv_data(self):
        """
        Get CSV data from the directive content, from an external
        file, or from a URL reference.
        """
        encoding = self.options.get(
            'encoding', self.state.document.settings.input_encoding)
        error_handler = self.state.document.settings.input_encoding_error_handler  # noqa:E501
        if self.content:
            # CSV data is from directive content.
            if 'file' in self.options or 'url' in self.options:
                error = self.reporter.error('"%s" directive may not both '
                    'specify an external file and have content.' % self.name,
                    nodes.literal_block(self.block_text, self.block_text),
                    line=self.lineno)
                raise SystemMessagePropagation(error)
            source = self.content.source(0)
            csv_data = self.content
        elif 'file' in self.options:
            # CSV data is from an external file.
            if 'url' in self.options:
                error = self.reporter.error('The "file" and "url" options '
                    'may not be simultaneously specified '
                    'for the "%s" directive.' % self.name,
                    nodes.literal_block(self.block_text, self.block_text),
                    line=self.lineno)
                raise SystemMessagePropagation(error)
            source = self.options['file']
            # resolve path to external file
            _base = Path(self.state.document.current_source).parent
            source = utils.relative_path(None, _base/source)
            try:
                csv_file = io.FileInput(source_path=source,
                                        encoding=encoding,
                                        error_handler=error_handler)
                csv_data = csv_file.read().splitlines()
            except OSError as error:
                severe = self.reporter.severe(
                    'Problems with "%s" directive path:\n%s.'
                    % (self.name, error),
                    nodes.literal_block(self.block_text, self.block_text),
                    line=self.lineno)
                raise SystemMessagePropagation(severe)
            else:
                self.state.document.settings.record_dependencies.add(source)
        elif 'url' in self.options:
            source = self.options['url']
            try:
                with urlopen(source) as response:
                    csv_text = response.read()
            except (URLError, OSError, ValueError) as error:
                severe = self.reporter.severe(
                      'Problems with "%s" directive URL "%s":\n%s.'
                      % (self.name, self.options['url'], error),
                      nodes.literal_block(self.block_text, self.block_text),
                      line=self.lineno)
                raise SystemMessagePropagation(severe)
            csv_file = io.StringInput(
                source=csv_text, source_path=source, encoding=encoding,
                error_handler=(self.state.document.settings.
                               input_encoding_error_handler))
            csv_data = csv_file.read().splitlines()
        else:
            error = self.reporter.warning(
                'The "%s" directive requires content; none supplied.'
                % self.name,
                nodes.literal_block(self.block_text, self.block_text),
                line=self.lineno)
            raise SystemMessagePropagation(error)
        return csv_data, source

    @staticmethod
    def decode_from_csv(s):
        warnings.warn('CSVTable.decode_from_csv()'
                  ' is not required with Python 3'
                  ' and will be removed in Docutils 0.21 or later.',
                  DeprecationWarning, stacklevel=2)
        return s

    @staticmethod
    def encode_for_csv(s):
        warnings.warn('CSVTable.encode_from_csv()'
                  ' is not required with Python 3'
                  ' and will be removed in Docutils 0.21 or later.',
                  DeprecationWarning, stacklevel=2)
        return s

    def parse_csv_data_into_rows(self, csv_data, dialect, source):
        csv_reader = csv.reader((line + '\n' for line in csv_data),
                                dialect=dialect)
        rows = []
        max_cols = 0
        for row in csv_reader:
            row_data = []
            for cell in row:
                cell_data = (0, 0, 0, statemachine.StringList(
                    cell.splitlines(), source=source))
                row_data.append(cell_data)
            rows.append(row_data)
            max_cols = max(max_cols, len(row))
        return rows, max_cols


class ListTable(Table):

    """
    Implement tables whose data is encoded as a uniform two-level bullet list.
    For further ideas, see
    https://docutils.sourceforge.io/docs/dev/rst/alternatives.html#list-driven-tables
    """

    option_spec = {'header-rows': directives.nonnegative_int,
                   'stub-columns': directives.nonnegative_int,
                   'width': directives.length_or_percentage_or_unitless,
                   'widths': directives.value_or(('auto', ),
                                                 directives.positive_int_list),
                   'class': directives.class_option,
                   'name': directives.unchanged,
                   'align': align}

    def run(self):
        if not self.content:
            error = self.reporter.error('The "%s" directive is empty; '
                'content required.' % self.name,
                nodes.literal_block(self.block_text, self.block_text),
                line=self.lineno)
            return [error]
        title, messages = self.make_title()
        node = nodes.Element()          # anonymous container for parsing
        self.state.nested_parse(self.content, self.content_offset, node)
        try:
            num_cols, col_widths = self.check_list_content(node)
            table_data = [[item.children for item in row_list[0]]
                          for row_list in node[0]]
            header_rows = self.options.get('header-rows', 0)
            stub_columns = self.options.get('stub-columns', 0)
            self.check_table_dimensions(table_data, header_rows, stub_columns)
        except SystemMessagePropagation as detail:
            return [detail.args[0]]
        table_node = self.build_table_from_list(table_data, col_widths,
                                                header_rows, stub_columns)
        if 'align' in self.options:
            table_node['align'] = self.options.get('align')
        table_node['classes'] += self.options.get('class', [])
        self.set_table_width(table_node)
        self.add_name(table_node)
        if title:
            table_node.insert(0, title)
        return [table_node] + messages

    def check_list_content(self, node):
        if len(node) != 1 or not isinstance(node[0], nodes.bullet_list):
            error = self.reporter.error(
                'Error parsing content block for the "%s" directive: '
                'exactly one bullet list expected.' % self.name,
                nodes.literal_block(self.block_text, self.block_text),
                line=self.lineno)
            raise SystemMessagePropagation(error)
        list_node = node[0]
        num_cols = 0
        # Check for a uniform two-level bullet list:
        for item_index in range(len(list_node)):
            item = list_node[item_index]
            if len(item) != 1 or not isinstance(item[0], nodes.bullet_list):
                error = self.reporter.error(
                    'Error parsing content block for the "%s" directive: '
                    'two-level bullet list expected, but row %s does not '
                    'contain a second-level bullet list.'
                    % (self.name, item_index + 1),
                    nodes.literal_block(self.block_text, self.block_text),
                    line=self.lineno)
                raise SystemMessagePropagation(error)
            elif item_index:
                if len(item[0]) != num_cols:
                    error = self.reporter.error(
                        'Error parsing content block for the "%s" directive: '
                        'uniform two-level bullet list expected, but row %s '
                        'does not contain the same number of items as row 1 '
                        '(%s vs %s).'
                        % (self.name, item_index + 1, len(item[0]), num_cols),
                        nodes.literal_block(self.block_text, self.block_text),
                        line=self.lineno)
                    raise SystemMessagePropagation(error)
            else:
                num_cols = len(item[0])
        col_widths = self.get_column_widths(num_cols)
        return num_cols, col_widths

    def build_table_from_list(self, table_data,
                              col_widths, header_rows, stub_columns):
        table = nodes.table()
        if self.widths == 'auto':
            table['classes'] += ['colwidths-auto']
        elif self.widths:  # explicitly set column widths
            table['classes'] += ['colwidths-given']
        tgroup = nodes.tgroup(cols=len(col_widths))
        table += tgroup
        for col_width in col_widths:
            colspec = nodes.colspec()
            if col_width is not None:
                colspec.attributes['colwidth'] = col_width
            if stub_columns:
                colspec.attributes['stub'] = 1
                stub_columns -= 1
            tgroup += colspec
        rows = []
        for row in table_data:
            row_node = nodes.row()
            for cell in row:
                entry = nodes.entry()
                entry += cell
                row_node += entry
            rows.append(row_node)
        if header_rows:
            thead = nodes.thead()
            thead.extend(rows[:header_rows])
            tgroup += thead
        tbody = nodes.tbody()
        tbody.extend(rows[header_rows:])
        tgroup += tbody
        return table