csv_plus_plus 0.1.1 → 0.1.3

data/lib/csv_plus_plus/language/runtime.rb

@@ -1,205 +0,0 @@
-# frozen_string_literal: true
-
-require_relative 'entities'
-require_relative 'syntax_error'
-require 'tempfile'
-
-module CSVPlusPlus
-  module Language
-    # The runtime state of the compiler (the current +line_number+/+row_index+, +cell+ being processed, etc).  We take
-    # multiple runs through the input file for parsing so it's really convenient to have a central place for these
-    # things to be managed.
-    #
-    # @attr_reader filename [String, nil] The filename that the input came from (mostly used for debugging since
-    #   +filename+ can be +nil+ if it's read from stdin.
-    # @attr_reader length_of_code_section [Integer] The length (count of lines) of the code section part of the original
-    #   input.
-    # @attr_reader length_of_csv_section [Integer] The length (count of lines) of the CSV part of the original csvpp
-    #   input.
-    # @attr_reader length_of_original_file [Integer] The length (count of lines) of the original csvpp input.
-    #
-    # @attr cell [Cell] The current cell being processed
-    # @attr cell_index [Integer] The index of the current cell being processed (starts at 0)
-    # @attr row_index [Integer] The index of the current row being processed (starts at 0)
-    # @attr line_number [Integer] The line number of the original csvpp template (starts at 1)
-    class Runtime
-      attr_reader :filename, :length_of_code_section, :length_of_csv_section, :length_of_original_file
-
-      attr_accessor :cell, :cell_index, :row_index, :line_number
-
-      # @param input [String] The input to be parsed
-      # @param filename [String, nil] The filename that the input came from (mostly used for debugging since +filename+
-      #   can be +nil+ if it's read from stdin
-      def initialize(input:, filename:)
-        @filename = filename || 'stdin'
-
-        init_input!(input)
-        start!
-      end
-
-      # Map over an a csvpp file and keep track of line_number and row_index
-      #
-      # @param lines [Array]
-      #
-      # @return [Array]
-      def map_lines(lines, &block)
-        @line_number = 1
-        lines.map do |line|
-          block.call(line).tap { next_line! }
-        end
-      end
-
-      # Map over a single row and keep track of the cell and it's index
-      #
-      # @param row [Array<Cell>] The row to map each cell over
-      #
-      # @return [Array]
-      def map_row(row, &block)
-        @cell_index = 0
-        row.map.with_index do |cell, index|
-          set_cell!(cell, index)
-          block.call(cell, index)
-        end
-      end
-
-      # Map over all rows and keep track of row and line numbers
-      #
-      # @param rows [Array<Row>] The rows to map over (and keep track of indexes)
-      # @param cells_too [boolean] If the cells of each +row+ should be iterated over also.
-      #
-      # @return [Array]
-      def map_rows(rows, cells_too: false, &block)
-        @row_index = 0
-        map_lines(rows) do |row|
-          if cells_too
-            # it's either CSV or a Row object
-            map_row((row.is_a?(::CSVPlusPlus::Row) ? row.cells : row), &block)
-          else
-            block.call(row)
-          end
-        end
-      end
-
-      # Increment state to the next line
-      #
-      # @return [Integer]
-      def next_line!
-        @row_index += 1 unless @row_index.nil?
-        @line_number += 1
-      end
-
-      # Return the current spreadsheet row number.  It parallels +@row_index+ but starts at 1.
-      #
-      # @return [Integer, nil]
-      def rownum
-        return if @row_index.nil?
-
-        @row_index + 1
-      end
-
-      # Set the current cell and index
-      #
-      # @param cell [Cell] The current cell
-      # @param cell_index [Integer] The index of the cell
-      def set_cell!(cell, cell_index)
-        @cell = cell
-        @cell_index = cell_index
-      end
-
-      # Each time we run a parse on the input, reset the runtime state starting at the beginning of the file
-      def start!
-        @row_index = @cell_index = nil
-        @line_number = 1
-      end
-
-      # Reset the runtime state starting at the CSV section
-      def start_at_csv!
-        # TODO: isn't the input re-written anyway without the code section? why do we need this?
-        start!
-        @line_number = @length_of_code_section || 1
-      end
-
-      # @return [String]
-      def to_s
-        "Runtime(cell: #{@cell}, row_index: #{@row_index}, cell_index: #{@cell_index})"
-      end
-
-      # Get the current (entity) value of a runtime value
-      #
-      # @param var_id [String, Symbol] The Variable#id  of the variable being resolved.
-      #
-      # @return [Entity]
-      def runtime_value(var_id)
-        if runtime_variable?(var_id)
-          ::CSVPlusPlus::Language::Builtins::VARIABLES[var_id.to_sym].resolve_fn.call(self)
-        else
-          raise_syntax_error('Undefined variable', var_id)
-        end
-      end
-
-      # Is +var_id+ a runtime variable?  (it's a static variable otherwise)
-      #
-      # @param var_id [String, Symbol] The Variable#id to check if it's a runtime variable
-      #
-      # @return [boolean]
-      def runtime_variable?(var_id)
-        ::CSVPlusPlus::Language::Builtins::VARIABLES.key?(var_id.to_sym)
-      end
-
-      # Called when an error is encoutered during parsing.  It will construct a useful
-      # error with the current +@row/@cell_index+, +@line_number+ and +@filename+
-      #
-      # @param message [String] A message relevant to why this error is being raised.
-      # @param bad_input [String] The offending input that caused this error to be thrown.
-      # @param wrapped_error [StandardError, nil] The underlying error that was raised (if it's not from our own logic)
-      def raise_syntax_error(message, bad_input, wrapped_error: nil)
-        raise(::CSVPlusPlus::Language::SyntaxError.new(message, bad_input, self, wrapped_error:))
-      end
-
-      # The currently available input for parsing.  The tmp state will be re-written
-      # between parsing the code section and the CSV section
-      #
-      # @return [String]
-      def input
-        @tmp
-      end
-
-      # We mutate the input over and over. It's ok because it's just a Tempfile
-      #
-      # @param data [String] The data to rewrite our input file to
-      def rewrite_input!(data)
-        @tmp.truncate(0)
-        @tmp.write(data)
-        @tmp.rewind
-      end
-
-      # Clean up the Tempfile we're using for parsing
-      def cleanup!
-        return unless @tmp
-
-        @tmp.close
-        @tmp.unlink
-        @tmp = nil
-      end
-
-      private
-
-      def count_code_section_lines(lines)
-        eoc = ::CSVPlusPlus::Lexer::END_OF_CODE_SECTION
-        lines.include?(eoc) ? (lines.take_while { |l| l != eoc }).length + 1 : 0
-      end
-
-      def init_input!(input)
-        lines = (input || '').split(/\s*\n\s*/)
-        @length_of_original_file = lines.length
-        @length_of_code_section = count_code_section_lines(lines)
-        @length_of_csv_section = @length_of_original_file - @length_of_code_section
-
-        # we're gonna take our input file, write it to a tmp file then each
-        # step is gonna mutate that tmp file
-        @tmp = ::Tempfile.new
-        rewrite_input!(input)
-      end
-    end
-  end
-end

data/lib/csv_plus_plus/language/scope.rb

@@ -1,192 +0,0 @@
-# frozen_string_literal: true
-
-require_relative '../code_section'
-require_relative '../graph'
-require_relative './entities'
-require_relative './references'
-require_relative './syntax_error'
-
-module CSVPlusPlus
-  module Language
-    # A class representing the scope of the current Template and responsible for resolving variables
-    #
-    # @attr_reader code_section [CodeSection] The CodeSection containing variables and functions to be resolved
-    # @attr_reader runtime [Runtime] The compiler's current runtime
-    #
-    # rubocop:disable Metrics/ClassLength
-    class Scope
-      attr_reader :code_section, :runtime
-
-      # initialize with a +Runtime+ and optional +CodeSection+
-      #
-      # @param runtime [Runtime]
-      # @param code_section [Runtime, nil]
-      def initialize(runtime:, code_section: nil)
-        @code_section = code_section if code_section
-        @runtime = runtime
-      end
-
-      # Resolve all values in the ast of the current cell being processed
-      #
-      # @return [Entity]
-      def resolve_cell_value
-        return unless (ast = @runtime.cell&.ast)
-
-        last_round = nil
-        loop do
-          refs = ::CSVPlusPlus::Language::References.extract(ast, @code_section)
-          return ast if refs.empty?
-
-          # TODO: throw an error here instead I think - basically we did a round and didn't make progress
-          return ast if last_round == refs
-
-          ast = resolve_functions(resolve_variables(ast, refs.variables), refs.functions)
-        end
-      end
-
-      # Set the +code_section+ and resolve all inner dependencies in it's variables and functions.
-      #
-      # @param code_section [CodeSection] The code_section to be resolved
-      def code_section=(code_section)
-        @code_section = code_section
-        resolve_static_variables!
-      end
-
-      # @return [String]
-      def to_s
-        "Scope(code_section: #{@code_section}, runtime: #{@runtime})"
-      end
-
-      private
-
-      # Resolve all variable references defined statically in the code section
-      # TODO: experiment with getting rid of this - does it even play correctly with runtime vars?
-      def resolve_static_variables!
-        variables = @code_section.variables
-        last_var_dependencies = {}
-        loop do
-          var_dependencies, resolution_order = variable_resolution_order(only_static_vars(variables))
-          return if var_dependencies == last_var_dependencies
-
-          # TODO: make the contract better here where we're not seting the variables of another class
-          @code_section.variables = resolve_dependencies(var_dependencies, resolution_order, variables)
-          last_var_dependencies = var_dependencies.clone
-        end
-      end
-
-      def only_static_vars(var_dependencies)
-        var_dependencies.reject { |k| @runtime.runtime_variable?(k) }
-      end
-
-      def resolve_functions(ast, refs)
-        refs.reduce(ast.dup) do |acc, elem|
-          function_replace(acc, elem.id, resolve_function(elem.id))
-        end
-      end
-
-      def resolve_variables(ast, refs)
-        refs.reduce(ast.dup) do |acc, elem|
-          variable_replace(acc, elem.id, resolve_variable(elem.id))
-        end
-      end
-
-      # Make a copy of the AST represented by +node+ and replace +fn_id+ with +replacement+ throughout
-      # rubocop:disable Metrics/MethodLength
-      def function_replace(node, fn_id, replacement)
-        if node.function_call? && node.id == fn_id
-          call_function_or_runtime_value(replacement, node)
-        elsif node.function_call?
-          # not our function, but continue our depth first search on it
-          ::CSVPlusPlus::Language::Entities::FunctionCall.new(
-            node.id,
-            node.arguments.map { |n| function_replace(n, fn_id, replacement) },
-            infix: node.infix
-          )
-        else
-          node
-        end
-      end
-      # rubocop:enable Metrics/MethodLength
-
-      def resolve_function(fn_id)
-        id = fn_id.to_sym
-        return @code_section.functions[id] if @code_section.defined_function?(id)
-
-        ::CSVPlusPlus::Language::Builtins::FUNCTIONS[id]
-      end
-
-      def call_function_or_runtime_value(function_or_runtime_value, function_call)
-        if function_or_runtime_value.function?
-          call_function(function_or_runtime_value, function_call)
-        else
-          function_or_runtime_value.resolve_fn.call(@runtime, function_call.arguments)
-        end
-      end
-
-      def call_function(function, function_call)
-        i = 0
-        function.arguments.reduce(function.body.dup) do |ast, argument|
-          variable_replace(ast, argument, function_call.arguments[i]).tap do
-            i += 1
-          end
-        end
-      end
-
-      # Make a copy of the AST represented by +node+ and replace +var_id+ with +replacement+ throughout
-      def variable_replace(node, var_id, replacement)
-        if node.function_call?
-          arguments = node.arguments.map { |n| variable_replace(n, var_id, replacement) }
-          # TODO: refactor these places where we copy functions... it's brittle with the kwargs
-          ::CSVPlusPlus::Language::Entities::FunctionCall.new(node.id, arguments, infix: node.infix)
-        elsif node.variable? && node.id == var_id
-          replacement
-        else
-          node
-        end
-      end
-
-      def resolve_variable(var_id)
-        id = var_id.to_sym
-        return @code_section.variables[id] if @code_section.defined_variable?(id)
-
-        # this will throw a syntax error if it doesn't exist (which is what we want)
-        @runtime.runtime_value(id)
-      end
-
-      def check_unbound_vars(dependencies, variables)
-        unbound_vars = dependencies.values.flatten - variables.keys
-        return if unbound_vars.empty?
-
-        @runtime.raise_syntax_error('Undefined variables', unbound_vars.map(&:to_s).join(', '))
-      end
-
-      def variable_resolution_order(variables)
-        # we have a hash of variables => ASTs but they might have references to each other, so
-        # we need to interpolate them first (before interpolating the cell values)
-        var_dependencies = ::CSVPlusPlus::Graph.dependency_graph(variables, @runtime)
-        # are there any references that we don't have variables for? (undefined variable)
-        check_unbound_vars(var_dependencies, variables)
-
-        # a topological sort will give us the order of dependencies
-        [var_dependencies, ::CSVPlusPlus::Graph.topological_sort(var_dependencies)]
-        # TODO: don't expose this exception directly to the caller
-      rescue ::TSort::Cyclic
-        @runtime.raise_syntax_error('Cyclic variable dependency detected', var_refs.keys)
-      end
-
-      def resolve_dependencies(var_dependencies, resolution_order, variables)
-        {}.tap do |resolved_vars|
-          # for each var and each dependency it has, build up and mutate resolved_vars
-          resolution_order.each do |var|
-            resolved_vars[var] = variables[var].dup
-
-            var_dependencies[var].each do |dependency|
-              resolved_vars[var] = variable_replace(resolved_vars[var], dependency, variables[dependency])
-            end
-          end
-        end
-      end
-    end
-    # rubocop:enable Metrics/ClassLength
-  end
-end

data/lib/csv_plus_plus/language/syntax_error.rb

@@ -1,66 +0,0 @@
-# frozen_string_literal: true
-
-module CSVPlusPlus
-  module Language
-    # An error that can be thrown for various syntax errors
-    class SyntaxError < ::CSVPlusPlus::Error
-      # @param message [String] The primary message to be shown to the user
-      # @param bad_input [String] The offending input that caused the error to be thrown
-      # @param runtime [Runtime] The current runtime
-      # @param wrapped_error [StandardError] The underlying error that caused the syntax error.  For example a
-      #   Racc::ParseError that was thrown
-      def initialize(message, bad_input, runtime, wrapped_error: nil)
-        @bad_input = bad_input.to_s
-        @runtime = runtime
-        @wrapped_error = wrapped_error
-        @message = message
-
-        super(message)
-      end
-
-      # @return [String]
-      def to_s
-        to_trace
-      end
-
-      # Output a verbose user-helpful string that references the current runtime
-      def to_verbose_trace
-        warn(@wrapped_error.full_message) if @wrapped_error
-        warn(@wrapped_error.backtrace) if @wrapped_error
-        to_trace
-      end
-
-      # Output a user-helpful string that references the runtime state
-      #
-      # @return [String]
-      def to_trace
-        "#{message_prefix}#{cell_index} #{message_postfix}"
-      end
-
-      private
-
-      def cell_index
-        row_index = @runtime.row_index
-        if @runtime.cell_index
-          "[#{row_index},#{@runtime.cell_index}]"
-        elsif row_index
-          "[#{row_index}]"
-        else
-          ''
-        end
-      end
-
-      def message_prefix
-        line_number = @runtime.line_number
-        filename = @runtime.filename
-
-        line_str = line_number ? ":#{line_number}" : ''
-        "csv++ #{filename}#{line_str}"
-      end
-
-      def message_postfix
-        "#{@message}: \"#{@bad_input}\""
-      end
-    end
-  end
-end