class Asciidoctor::Table::ParserContext
finally, a new buffer is allocated to track the next cell.
instantiated, the row is closed if the cell satisifies the column count and,
is located, the previous cell is closed, an instance of Table::Cell is
moves through the lines of the table using tail recursion. When a cell boundary
class are primarily responsible for tracking the buffer of a cell as the parser
Public: Methods for managing the parsing of an AsciiDoc table. Instances of this
def activate_rowspan(rowspan, colspan)
determining the effective number of cells in the current row.
Internal: Activate a rowspan. The rowspan Array is consulted when
def activate_rowspan(rowspan, colspan) 1.upto(rowspan - 1) {|i| @active_rowspans[i] = (@active_rowspans[i] || 0) + colspan } nil end
def advance
Internal: Advance to the next line (which may come after the parser begins processing
def advance @linenum += 1 end
def buffer_has_unclosed_quotes? append = nil
returns true if the buffer has unclosed quotes, false if it doesn't or it
Public: Determines whether the buffer has unclosed quotes. Used for CSV data.
def buffer_has_unclosed_quotes? append = nil if (record = append ? (@buffer + append).strip : @buffer.strip) == '"' true elsif record.start_with? '"' if ((trailing_quote = record.end_with? '"') && (record.end_with? '""')) || (record.start_with? '""') ((record = record.gsub '""', '').start_with? '"') && !(record.end_with? '"') else !trailing_quote end else false end end
def cell_closed?
Public: Checks whether the current cell has been marked as closed
def cell_closed? !@cell_open end
def cell_open?
Public: Checks whether the current cell is still open
def cell_open? @cell_open end
def close_cell(eol = false)
row has been met, close the row and begin a new one.
the current row and, if the number of expected columns for the current
Public: Close the current cell, instantiate a new Table::Cell, add it to
def close_cell(eol = false) if @format == 'psv' cell_text = @buffer @buffer = '' if (cellspec = take_cellspec) repeat = cellspec.delete('repeatcol') || 1 else logger.error message_with_context 'table missing leading separator; recovering automatically', source_location: Reader::Cursor.new(*@start_cursor_data) cellspec = {} repeat = 1 end else cell_text = @buffer.strip @buffer = '' cellspec = nil repeat = 1 if @format == 'csv' && !cell_text.empty? && cell_text.include?('"') # this may not be perfect logic, but it hits the 99% if cell_text.start_with?('"') && cell_text.end_with?('"') # unquote if (cell_text = cell_text.slice(1, cell_text.length - 2)) # trim whitespace and collapse escaped quotes cell_text = cell_text.strip.squeeze('"') else logger.error message_with_context 'unclosed quote in CSV data; setting cell to empty', source_location: @reader.cursor_at_prev_line cell_text = '' end else # collapse escaped quotes cell_text = cell_text.squeeze('"') end end end 1.upto(repeat) do |i| # TODO make column resolving an operation if @colcount == -1 @table.columns << (column = Table::Column.new(@table, @table.columns.size + i - 1)) if cellspec && (cellspec.key? 'colspan') && (extra_cols = cellspec['colspan'].to_i - 1) > 0 offset = @table.columns.size extra_cols.times do |j| @table.columns << Table::Column.new(@table, offset + j) end end else # QUESTION is this right for cells that span columns? unless (column = @table.columns[@current_row.size]) logger.error message_with_context 'dropping cell because it exceeds specified number of columns', source_location: @reader.cursor_before_mark return end end cell = Table::Cell.new(column, cell_text, cellspec, cursor: @reader.cursor_before_mark) @reader.mark unless !cell.rowspan || cell.rowspan == 1 activate_rowspan(cell.rowspan, (cell.colspan || 1)) end @column_visits += (cell.colspan || 1) @current_row << cell # don't close the row if we're on the first line and the column count has not been set explicitly # TODO perhaps the colcount/linenum logic should be in end_of_row? (or a should_end_row? method) close_row if end_of_row? && (@colcount != -1 || @linenum > 0 || (eol && i == repeat)) end @cell_open = false nil end
def close_open_cell(next_cellspec = {})
by the next cell.
cell spec captured from the end of this cell onto the stack for use
Public: If the current cell is open, close it. In additional, push the
def close_open_cell(next_cellspec = {}) push_cellspec next_cellspec close_cell(true) if cell_open? advance nil end
def close_row
Array and counter variables.
Internal: Close the row by adding it to the Table and resetting the row
def close_row @table.rows.body << @current_row # don't have to account for active rowspans here # since we know this is first row @colcount = @column_visits if @colcount == -1 @column_visits = 0 @current_row = [] @active_rowspans.shift @active_rowspans[0] ||= 0 nil end
def effective_column_visits
Internal: Calculate the effective column visits, which consists of the number of
def effective_column_visits @column_visits + @active_rowspans[0] end
def end_of_row?
def end_of_row? @colcount == -1 || effective_column_visits == @colcount end
def initialize reader, table, attributes = {}
def initialize reader, table, attributes = {} @start_cursor_data = (@reader = reader).mark @table = table if attributes.key? 'format' if FORMATS.include?(xsv = attributes['format']) if xsv == 'tsv' # NOTE tsv is just an alias for csv with a tab separator @format = 'csv' elsif (@format = xsv) == 'psv' && table.document.nested? xsv = '!sv' end else logger.error message_with_context %(illegal table format: #{xsv}), source_location: reader.cursor_at_prev_line @format, xsv = 'psv', (table.document.nested? ? '!sv' : 'psv') end else @format, xsv = 'psv', (table.document.nested? ? '!sv' : 'psv') end if attributes.key? 'separator' if (sep = attributes['separator']).nil_or_empty? @delimiter, @delimiter_rx = DELIMITERS[xsv] # QUESTION should we support any other escape codes or multiple tabs? elsif sep == '\t' @delimiter, @delimiter_rx = DELIMITERS['tsv'] else @delimiter, @delimiter_rx = sep, /#{::Regexp.escape sep}/ end else @delimiter, @delimiter_rx = DELIMITERS[xsv] end @colcount = table.columns.empty? ? -1 : table.columns.size @buffer = '' @cellspecs = [] @cell_open = false @active_rowspans = [0] @column_visits = 0 @current_row = [] @linenum = -1 end
def keep_cell_open
reached and the cell may contain additional text.
Public: Marks that the cell should be kept open. Used when the end of the line is
def keep_cell_open @cell_open = true nil end
def mark_cell_closed
instance and add it to the current row.
Public: Marks the cell as closed so that the parser knows to instantiate a new cell
def mark_cell_closed @cell_open = false nil end
def match_delimiter(line)
used by this table.
Public: Checks whether the line provided contains the cell delimiter
def match_delimiter(line) @delimiter_rx.match(line) end
def push_cellspec(cellspec = {})
stack is used to carry over the spec to the next cell.
Public: Puts a cell spec onto the stack. Cell specs precede the delimiter, so a
def push_cellspec(cellspec = {}) # this shouldn't be nil, but we check anyway @cellspecs << (cellspec || {}) nil end
def skip_past_delimiter(pre)
Public: Skip past the matched delimiter because it's inside quoted text.
def skip_past_delimiter(pre) @buffer = %(#{@buffer}#{pre}#{@delimiter}) nil end
def skip_past_escaped_delimiter(pre)
Public: Skip past the matched delimiter because it's escaped.
def skip_past_escaped_delimiter(pre) @buffer = %(#{@buffer}#{pre.chop}#{@delimiter}) nil end
def starts_with_delimiter?(line)
used by this table.
Public: Checks whether the line provided starts with the cell delimiter
def starts_with_delimiter?(line) line.start_with? @delimiter end
def take_cellspec
when the cell is being closed.
stack is used to carry over the spec from the previous cell to the current cell
Public: Takes a cell spec from the stack. Cell specs precede the delimiter, so a
def take_cellspec @cellspecs.shift end