ruby_tree_sitter 1.2.0-x86_64-linux → 1.3.0-x86_64-linux

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: cfbfdfc0c297585fe24e04c4b9822ff404baa795182b7a7b144f56a3941db672
-  data.tar.gz: 010faa21be194e40cdf7f5f00986c16ee5d51d64f377097bfb307a4245ba61b6
+  metadata.gz: a0f7c51b289e769ad5701be586b28d61f41f0979ee6fd3c08f87481ae307f146
+  data.tar.gz: 988052e5b5b5c41e9a9d974398c6c6cae401f2068a2ad4013d80198f3437c029
 SHA512:
-  metadata.gz: b98898e1a8b801de3599e6daebd2adaa27b2ea729a46ebe730095d31e053dfb84c48a83712364a09fd58bac236f1b499f16ced6dcf4640902d0d4c12ac4ac7be
-  data.tar.gz: 4dd45cf49030a553819febb05cc8a0755ed549c413d44ea1c88836d4e2fad3c55138db3a3079776f2e14a828c38fd7d7b11664366ea8e84083098910528950b3
+  metadata.gz: 9ef62c5ea3bd4451b03b43e6efc2c3501191ffcd9ea96dfe12d2e66b249d1f16f58e82d6127d833a6c8105e3f910dda2ba898a138d5e6d2756615bf69fcb2e05
+  data.tar.gz: 5bd46cc66e845c9091005a5eadeccb3db2984e0bbc5bf6d12e9fc4e3fd126f9ab6f1d37970fbe6760eee096ec2c011d4d328d5610a9e873ed313df3842caceb3

data/ext/tree_sitter/query.c

@@ -140,6 +140,13 @@ static VALUE query_initialize(VALUE self, VALUE language, VALUE source) {
     SELF = res;
   }
 
+  rb_iv_set(self, "@text_predicates", rb_ary_new());
+  rb_iv_set(self, "@property_predicates", rb_ary_new());
+  rb_iv_set(self, "@property_settings", rb_ary_new());
+  rb_iv_set(self, "@general_predicates", rb_ary_new());
+
+  rb_funcall(self, rb_intern("process"), 1, source);
+
   return self;
 }
 

data/ext/tree_sitter/query_cursor.c

@@ -49,7 +49,7 @@ DATA_FROM_VALUE(TSQueryCursor *, query_cursor)
  *
  * @return [QueryCursor]
  */
-static VALUE query_cursor_exec(VALUE self, VALUE query, VALUE node) {
+static VALUE query_cursor_exec_static(VALUE self, VALUE query, VALUE node) {
   VALUE res = query_cursor_allocate(cQueryCursor);
   query_cursor_t *query_cursor = unwrap(res);
   ts_query_cursor_exec(query_cursor->data, value_to_query(query),
@@ -57,6 +57,21 @@ static VALUE query_cursor_exec(VALUE self, VALUE query, VALUE node) {
   return res;
 }
 
+/**
+ * Start running a given query on a given node.
+ *
+ * @param query [Query]
+ * @param node  [Node]
+ *
+ * @return [QueryCursor]
+ */
+static VALUE query_cursor_exec(VALUE self, VALUE query, VALUE node) {
+  query_cursor_t *query_cursor = unwrap(self);
+  ts_query_cursor_exec(query_cursor->data, value_to_query(query),
+                       value_to_node(node));
+  return self;
+}
+
 /**
  * Manage the maximum number of in-progress matches allowed by this query
  * cursor.
@@ -190,13 +205,14 @@ void init_query_cursor(void) {
   rb_define_alloc_func(cQueryCursor, query_cursor_allocate);
 
   /* Module methods */
-  rb_define_module_function(cQueryCursor, "exec", query_cursor_exec, 2);
+  rb_define_module_function(cQueryCursor, "exec", query_cursor_exec_static, 2);
 
   /* Class methods */
   // Accessors
   DECLARE_ACCESSOR(cQueryCursor, query_cursor, match_limit)
 
   // Other
+  rb_define_method(cQueryCursor, "exec", query_cursor_exec, 2);
   rb_define_method(cQueryCursor, "exceed_match_limit?",
                    query_cursor_did_exceed_match_limit, 0);
   rb_define_method(cQueryCursor, "match_limit", query_cursor_get_match_limit,

data/lib/tree_sitter/helpers.rb

@@ -0,0 +1,23 @@
+# frozen_string_literal: true
+
+# splits an array like [rust](https://doc.rust-lang.org/std/primitive.slice.html#method.split)
+def array_split_like_rust(array, &block)
+  return enum_for(__method__, array) if !block_given?
+
+  return [] if array.empty?
+
+  result = []
+  current_slice = []
+
+  array.each do |element|
+    if yield(element)
+      result << current_slice
+      current_slice = []
+    else
+      current_slice << element
+    end
+  end
+
+  result << current_slice
+  result
+end

data/lib/tree_sitter/query.rb

@@ -0,0 +1,191 @@
+# frozen_string_literal: true
+
+require_relative 'helpers'
+
+module TreeSitter
+  # Query is a wrapper around a tree-sitter query.
+  class Query
+    attr_reader :capture_names
+    attr_reader :capture_quantifiers
+    attr_reader :text_predicates
+    attr_reader :property_predicates
+    attr_reader :property_settings
+    attr_reader :general_predicates
+
+    private
+
+    # Called from query.c on initialize.
+    #
+    # Prepares all the predicates so we could process them in places like
+    # {QueryMatch#satisfies_text_predicate?}.
+    #
+    # This is translation from the [rust bindings](https://github.com/tree-sitter/tree-sitter/blob/e553578696fe86071846ed612ee476d0167369c1/lib/binding_rust/lib.rs#L1860)
+    # Because it's a direct translation, it's way too long and we need to shut up rubocop.
+    # TODO: refactor + simplify when stable.
+    def process(source) # rubocop:disable Metrics/AbcSize,Metrics/CyclomaticComplexity,Metrics/MethodLength,Metrics/PerceivedComplexity
+      string_count = self.string_count
+      capture_count = self.capture_count
+      pattern_count = self.pattern_count
+
+      # Build a vector of strings to store the capture names.
+      capture_names = capture_count.times.map { |i| capture_name_for_id(i) }
+
+      # Build a vector to store capture qunatifiers.
+      capture_quantifiers =
+        pattern_count.times.map do |i|
+          capture_count.times.map do |j|
+            capture_quantifier_for_id(i, j)
+          end
+        end
+
+      # Build a vector of strings to represent literal values used in predicates.
+      string_values = string_count.times.map { |i| string_value_for_id(i) }
+
+      # Build a vector of predicates for each pattern.
+      pattern_count.times do |i| # rubocop:disable Metrics/BlockLength
+        predicate_steps = predicates_for_pattern(i)
+        byte_offset = start_byte_for_pattern(i)
+        row =
+          source.chars.map.with_index
+            .take_while { |_, i| i < byte_offset } # rubocop:disable Lint/ShadowingOuterLocalVariable
+            .filter { |c, _| c == "\n" }
+            .size
+        text_predicates = []
+        property_predicates = []
+        property_settings = []
+        general_predicates = []
+
+        array_split_like_rust(predicate_steps) { |s| s.type == QueryPredicateStep::DONE } # rubocop:disable Metrics/BlockLength
+          .each do |p|
+            next if p.empty?
+
+            if p[0] == QueryPredicateStep::STRING
+              cap = capture_names[p[0].value_id]
+              raise ArgumentError, <<~MSG.chomp
+                L#{row}: Expected predicate to start with a function name. Got @#{cap}.
+              MSG
+            end
+
+            # Build a predicate for each of the known predicate function names.
+            operator_name = string_values[p[0].value_id]
+
+            case operator_name
+            in 'any-eq?' | 'any-not-eq?' | 'eq?' | 'not-eq?'
+              if p.size != 3
+                raise ArgumentError, <<~MSG.chomp
+                  L#{row}: Wrong number of arguments to ##{operator_name} predicate. Expected 2, got #{p.size - 1}.
+                MSG
+              end
+
+              if p[1].type != QueryPredicateStep::CAPTURE
+                lit = string_values[p[1].value_id]
+                raise ArgumentError, <<~MSG.chomp
+                  L#{row}: First argument to ##{operator_name} predicate must be a capture name. Got literal "#{lit}".
+                MSG
+              end
+
+              is_positive = %w[eq? any-eq?].include?(operator_name)
+              match_all = %w[eq? not-eq?].include?(operator_name)
+              # NOTE: in the rust impl, match_all can hit an unreachable! but I am simplifying
+              # for readability. Same applies for the other `in` branches.
+              text_predicates <<
+                if p[2].type == QueryPredicateStep::CAPTURE
+                  TextPredicateCapture.eq_capture(p[1].value_id, p[2].value_id, is_positive, match_all)
+                else
+                  TextPredicateCapture.eq_string(p[1].value_id, string_values[p[2].value_id], is_positive, match_all)
+                end
+
+            in 'match?' | 'not-match?' | 'any-match?' | 'any-not-match?'
+              if p.size != 3
+                raise ArgumentError, <<~MSG.chomp
+                  L#{row}: Wrong number of arguments to ##{operator_name} predicate. Expected 2, got #{p.size - 1}.
+                MSG
+              end
+
+              if p[1].type != QueryPredicateStep::CAPTURE
+                lit = string_values[p[1].value_id]
+                raise ArgumentError, <<~MSG.chomp
+                  L#{row}: First argument to ##{operator_name} predicate must be a capture name. Got literal "#{lit}".
+                MSG
+              end
+
+              if p[2].type == QueryPredicateStep::CAPTURE
+                cap = capture_names[p[2].value_id]
+                raise ArgumentError, <<~MSG.chomp
+                  L#{row}: First argument to ##{operator_name} predicate must be a literal. Got capture @#{cap}".
+                MSG
+              end
+
+              is_positive = %w[match? any-match?].include?(operator_name)
+              match_all = %w[match? not-match?].include?(operator_name)
+              regex = /#{string_values[p[2].value_id]}/
+
+              text_predicates << TextPredicateCapture.match_string(p[1].value_id, regex, is_positive, match_all)
+
+            in 'set!'
+              property_settings << 'todo!'
+
+            in 'is?' | 'is-not?'
+              property_predicates << 'todo!'
+
+            in 'any-of?' | 'not-any-of?'
+              if p.size < 2
+                raise ArgumentError, <<~MSG.chomp
+                  L#{row}: Wrong number of arguments to ##{operator_name} predicate. Expected at least 1, got #{p.size - 1}.
+                MSG
+              end
+
+              if p[1].type != QueryPredicateStep::CAPTURE
+                lit = string_values[p[1].value_id]
+                raise ArgumentError, <<~MSG.chomp
+                  L#{row}: First argument to ##{operator_name} predicate must be a capture name. Got literal "#{lit}".
+                MSG
+              end
+
+              is_positive = operator_name == 'any_of'
+              values = []
+
+              p[2..].each do |arg|
+                if arg.type == QueryPredicateStep::CAPTURE
+                  lit = string_values[arg.value_id]
+                  raise ArgumentError, <<~MSG.chomp
+                    L#{row}: First argument to ##{operator_name} predicate must be a capture name. Got literal "#{lit}".
+                  MSG
+                end
+                values << string_values[arg.value_id]
+              end
+
+              # TODO: is the map to to_s necessary in ruby?
+              text_predicates <<
+                TextPredicateCapture.any_string(p[1].value_id, values.map(&:to_s), is_positive, match_all)
+            else
+              general_predicates <<
+                QueryPredicate.new(
+                  operator_name,
+                  p[1..].map do |a|
+                    if a.type == QueryPredicateStep::CAPTURE
+                      { capture: a.value_id }
+                    else
+                      { string: string_values[a.value_id] }
+                    end
+                  end,
+                )
+            end
+
+            @text_predicates << text_predicates
+            @property_predicates << property_predicates
+            @property_settings << property_settings
+            @general_predicates << general_predicates
+          end
+
+        @capture_names = capture_names
+        @capture_quantifiers = capture_quantifiers
+      end
+    end
+
+    # TODO
+    def parse_property
+      # todo
+    end
+  end
+end

data/lib/tree_sitter/query_captures.rb

@@ -0,0 +1,28 @@
+# frozen_string_literal: true
+
+module TreeSitter
+  # A sequence of {TreeSitter::QueryCapture} associated with a given {TreeSitter::QueryCursor}.
+  class QueryCaptures
+    def initialize(cursor, query, src)
+      @cursor = cursor
+      @query = query
+      @src = src
+    end
+
+    # Iterator over captures.
+    #
+    # @yieldparam match [TreeSitter::QueryMatch]
+    # @yieldparam capture_index [Integer]
+    def each
+      return enum_for __method__ if !block_given?
+
+      while (capture_index, match = @cursor.next_capture)
+        next if !match.is_a?(TreeSitter::QueryMatch)
+
+        if match.satisfies_text_predicate?(@query, @src)
+          yield [match, capture_index]
+        end
+      end
+    end
+  end
+end

data/lib/tree_sitter/query_cursor.rb

@@ -0,0 +1,27 @@
+# frozen_string_literal: true
+
+module TreeSitter
+  # A Cursor for {Query}.
+  class QueryCursor
+    # Iterate over all of the matches in the order that they were found.
+    #
+    # Each match contains the index of the pattern that matched, and a list of
+    # captures. Because multiple patterns can match the same set of nodes,
+    # one match may contain captures that appear *before* some of the
+    # captures from a previous match.
+    def matches(query, node, src)
+      self.exec(query, node)
+      QueryMatches.new(self, query, src)
+    end
+
+    # Iterate over all of the individual captures in the order that they
+    # appear.
+    #
+    # This is useful if you don't care about which pattern matched, and just
+    # want a single, ordered sequence of captures.
+    def captures(query, node, src)
+      self.exec(query, node)
+      QueryCaptures.new(self, query, src)
+    end
+  end
+end

data/lib/tree_sitter/query_match.rb

@@ -0,0 +1,100 @@
+# frozen_string_literal: true
+
+require_relative 'query_captures'
+
+module TreeSitter
+  # A match for a {Query}.
+  class QueryMatch
+    # All nodes at a given capture index.
+    #
+    # @param index [Integer]
+    #
+    # @return [TreeSitter::Node]
+    def nodes_for_capture_index(index) = captures.filter_map { |capture| capture.node if capture.index == index }
+
+    # Whether the {QueryMatch} satisfies the text predicates in the query.
+    #
+    # This is a translation from the [rust bindings](https://github.com/tree-sitter/tree-sitter/blob/e553578696fe86071846ed612ee476d0167369c1/lib/binding_rust/lib.rs#L2502).
+    # Because it's a direct translation, it's way too long and we need to shut up rubocop.
+    # TODO: refactor + simplify when satable.
+    def satisfies_text_predicate?(query, src) # rubocop:disable Metrics/AbcSize,Metrics/CyclomaticComplexity,Metrics/MethodLength,Metrics/PerceivedComplexity
+      return true if query.text_predicates[pattern_index].nil?
+
+      query # rubocop:disable Metrics/BlockLength
+        .text_predicates[pattern_index]
+        .all? do |predicate|
+          case predicate.type
+          in TextPredicateCapture::EQ_CAPTURE
+            fst_nodes = nodes_for_capture_index(predicate.fst)
+            snd_nodes = nodes_for_capture_index(predicate.snd)
+            res = nil
+            consumed = 0
+            fst_nodes.zip(snd_nodes).each do |node1, node2|
+              text1 = node_text(node1, src)
+              text2 = node_text(node2, src)
+              if (text1 == text2) != predicate.positive? && predicate.match_all?
+                res = false
+                break
+              end
+              if (text1 == text2) == predicate.positive? && !predicate.match_all?
+                res = true
+                break
+              end
+              consumed += 1
+            end
+            (res.nil? && consumed == fst_nodes.length && consumed == snd_nodes.length) \
+              || res
+
+          in TextPredicateCapture::EQ_STRING
+            nodes = nodes_for_capture_index(predicate.fst)
+            res = true
+            nodes.each do |node|
+              text = node_text(node, src)
+              if (predicate.snd == text) != predicate.positive? && predicate.match_all?
+                res = false
+                break
+              end
+              if (predicate.snd == text) == predicate.positive? && !predicate.match_all?
+                res = true
+                break
+              end
+            end
+            res
+
+          in TextPredicateCapture::MATCH_STRING
+            nodes = nodes_for_capture_index(predicate.fst)
+            res = true
+            nodes.each do |node|
+              text = node_text(node, src)
+              if predicate.snd.match?(text) != predicate.positive? && predicate.match_all?
+                res = false
+                break
+              end
+              if predicate.snd.match?(text) == predicate.positive? && !predicate.match_all?
+                res = true
+                break
+              end
+            end
+            res
+
+          in TextPredicateCapture::ANY_STRING
+            nodes = nodes_for_capture_index(predicate.fst)
+            res = true
+            nodes.each do |node|
+              text = node_text(node, src)
+              if predicate.snd.any? { |v| v == text } != predicate.positive?
+                res = false
+                break
+              end
+            end
+            res
+
+          end
+        end
+    end
+
+    private
+
+    def node_text(node, text) = text.byteslice(node.start_byte...node.end_byte)
+  end
+end

data/lib/tree_sitter/query_matches.rb

@@ -0,0 +1,25 @@
+# frozen_string_literal: true
+
+module TreeSitter
+  # A sequence of {QueryMatch} associated with a given {QueryCursor}.
+  class QueryMatches
+    def initialize(cursor, query, src)
+      @cursor = cursor
+      @query = query
+      @src = src
+    end
+
+    # Iterator over matches.
+    #
+    # @yieldparam match [TreeSitter::QueryMatch]
+    def each
+      return enum_for __method__ if !block_given?
+
+      while match = @cursor.next_match
+        if match.satisfies_text_predicate?(@query, @src)
+          yield match
+        end
+      end
+    end
+  end
+end

data/lib/tree_sitter/query_predicate.rb

@@ -0,0 +1,14 @@
+# frozen_string_literal: true
+
+module TreeSitter
+  # A {Query} predicate generic representation.
+  class QueryPredicate
+    attr_accessor :operator
+    attr_accessor :args
+
+    def initialize(operator, args)
+      @operator = operator
+      @args = args
+    end
+  end
+end

data/lib/tree_sitter/text_predicate_capture.rb

@@ -0,0 +1,37 @@
+# frozen_string_literal: true
+
+module TreeSitter
+  # A representation for text predicates.
+  class TextPredicateCapture
+    EQ_CAPTURE = 0   # Equality Capture
+    EQ_STRING = 1    # Equality String
+    MATCH_STRING = 2 # Match String
+    ANY_STRING = 3   # Any String
+
+    attr_reader :fst
+    attr_reader :snd
+    attr_reader :type
+
+    # Create a TextPredicateCapture for {EQ_CAPTURE}.
+    def self.eq_capture(...) = new(EQ_CAPTURE, ...)
+    # Create a TextPredicateCapture for {EQ_STRING}.
+    def self.eq_string(...) = new(EQ_STRING, ...)
+    # Create a TextPredicateCapture for {MATCH_STRING}.
+    def self.match_string(...) = new(MATCH_STRING, ...)
+    # Create a TextPredicateCapture for {ANY_STRING}.
+    def self.any_string(...) = new(ANY_STRING, ...)
+
+    def initialize(type, fst, snd, positive, match_all)
+      @type = type
+      @fst = fst
+      @snd = snd
+      @positive = positive
+      @match_all = match_all
+    end
+
+    # `#eq` is positive, `#not-eq` is not.
+    def positive? = @positive
+    # `#any-` means don't match all.
+    def match_all? = @match_all
+  end
+end

data/lib/tree_sitter/version.rb

@@ -4,5 +4,5 @@ module TreeSitter
   # The version of the tree-sitter library.
   TREESITTER_VERSION = '0.22.6'
   # The current version of the gem.
-  VERSION = '1.2.0'
+  VERSION = '1.3.0'
 end

data/lib/tree_sitter.rb

@@ -6,9 +6,16 @@ end
 
 require 'set'
 
+require 'tree_sitter/tree_sitter'
 require 'tree_sitter/version'
 
-require 'tree_sitter/tree_sitter'
 require 'tree_sitter/node'
+require 'tree_sitter/query'
+require 'tree_sitter/query_captures'
+require 'tree_sitter/query_cursor'
+require 'tree_sitter/query_match'
+require 'tree_sitter/query_matches'
+require 'tree_sitter/query_predicate'
+require 'tree_sitter/text_predicate_capture'
 
 ObjectSpace.define_finalizer(TreeSitter::Tree.class, proc { TreeSitter::Tree.finalizer })

data/lib/tree_stand/parser.rb

@@ -43,7 +43,7 @@ module TreeStand
         .freeze
 
     # The default paths we use to lookup parsers when no specific
-    # {TreeStand::Config#parser_path} is {nil}.
+    # {TreeStand::Config#parser_path} is `nil`.
     # Order matters.
     LIBDIRS = [
       'tree-sitter-parsers',
@@ -180,7 +180,7 @@ module TreeStand
 
       # We know that the bindings will accept `lib`, but I don't know how to tell sorbet
       # the types in ext/tree_sitter where `load` is defined.
-      TreeSitter::Language.load(name.gsub(/-/, '_'), T.unsafe(lib))
+      TreeSitter::Language.load(name.tr('-', '_'), T.unsafe(lib))
     end
 
     # @param language [String]

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: ruby_tree_sitter
 version: !ruby/object:Gem::Version
-  version: 1.2.0
+  version: 1.3.0
 platform: x86_64-linux
 authors:
 - Firas al-Khalil
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2024-05-27 00:00:00.000000000 Z
+date: 2024-06-26 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: sorbet-runtime
@@ -74,7 +74,15 @@ files:
 - ext/tree_sitter/tree_sitter.c
 - ext/tree_sitter/tree_sitter.h
 - lib/tree_sitter.rb
+- lib/tree_sitter/helpers.rb
 - lib/tree_sitter/node.rb
+- lib/tree_sitter/query.rb
+- lib/tree_sitter/query_captures.rb
+- lib/tree_sitter/query_cursor.rb
+- lib/tree_sitter/query_match.rb
+- lib/tree_sitter/query_matches.rb
+- lib/tree_sitter/query_predicate.rb
+- lib/tree_sitter/text_predicate_capture.rb
 - lib/tree_sitter/tree_sitter.so
 - lib/tree_sitter/version.rb
 - lib/tree_stand.rb