tinygql 0.2.0 → 0.3.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 114da94b6827645fb92f0f3e85b9e4f34b64aaee8449c143217cc75fc903ab3c
-  data.tar.gz: 27031f9423d30f0453a6737be1184cb8c724a6290bc432b7e35dea255fa549a5
+  metadata.gz: e6f2ac9d1c4531277689c6eec21fd29cd1e6fe9715284813b57c741e3e9bace2
+  data.tar.gz: b9cd18ac29f1aeea347296f0408deadddec7a902d25495e3232b472180718695
 SHA512:
-  metadata.gz: a4362f8d3c3d638c715c4686ec9714d59f59d411be390cf115d0b7e8bc4f710fcbfe935ca058092f24935964fd29905643e6f1f4edeb84e238e1af0d5d2eff4e
-  data.tar.gz: 5ec4f82ff306221b2c8e008b3bb76feb1c55917ab26334e1531f7eb8e58f03e44a29901fe92765585cfa195c41a107fdd13337b3bb844f13bb9985f1394f8ee0
+  metadata.gz: 63c931df357b6091070a7cd0adab859f7a7489f8fa1220ffa2cfb7fc6a7af6660ede87af6c3a8b804ef1a2114594aa3086a3792b5d3ef40e20ac1a4546d13120
+  data.tar.gz: 96e0a9164e835cb382391c9ea761cdf2405ac6f38e268034f810f8ce369d4e968ec5a2c174f22f03f50a99bd1f155449e5d9e7b46c65fc771263cbc03a8f2f8b

data/README.md

@@ -66,6 +66,29 @@ ast = TinyGQL.parse "{ neat { cool } }"
 p ast.fold(Fold, []) # => ["neat", "cool"]
 ```
 
+Nodes store their position in the source GraphQL document.
+If you'd like to extract the line number of the node, you'll need to keep a reference to the document and pass it to the `line` method on the node:
+
+```ruby
+doc = <<-eod
+mutation {
+  likeStory(sturyID: 12345) {
+    story {
+      likeCount
+    }
+  }
+}
+
+eod
+
+parser = TinyGQL::Parser.new doc
+ast = parser.parse
+
+ast.find_all(&:field?).each { |node|
+  p node.name => node.line(doc)
+}
+```
+
 ## LICENSE:
 
 I've licensed this code as Apache 2.0, but the lexer is from [GraphQL-Ruby](https://github.com/rmosolgo/graphql-ruby/blob/772734dfcc7aa0513c867259912474ef0ba799c3/lib/graphql/language/lexer.rb) and is under the MIT license.

data/bin/make_hash.rb

@@ -0,0 +1,51 @@
+require "tinygql"
+
+# Calculate a perfect hash for GraphQL keywords
+
+def bits x
+  count = 0
+  while x > 0
+    count += 1
+    x >>= 1
+  end
+  count
+end
+
+# on is too short, and subscription is the longest.
+# The lexer can easily detect them by length, so lets calculate a perfect
+# hash for the rest.
+kws = TinyGQL::Lexer::KEYWORDS - ["on", "subscription"]
+MASK = (1 << bits(kws.length)) - 1
+
+prefixes = kws.map { |word| word[1, 2] }
+
+# make sure they're unique
+raise "Not unique" unless prefixes.uniq == prefixes
+
+keys = prefixes.map { |prefix|
+  prefix.bytes.reverse.inject(0) { |c, byte|
+    c << 8 | byte
+  }
+}
+
+shift = 32 - bits(kws.length) # use the top bits
+
+c = 13
+loop do
+  z = keys.map { |k| ((k * c) >> shift) & MASK }
+  break if z.uniq.length == z.length
+  c += 1
+end
+
+table = kws.zip(keys).each_with_object([]) { |(word, k), o|
+  hash = ((k * c) >> shift) & MASK
+  o[hash] = word.upcase.to_sym
+}
+
+print "KW_LUT ="
+pp table
+puts <<-eomethod
+def hash key
+  (key * #{c}) >> #{shift} & #{sprintf("%#0x", MASK)}
+end
+eomethod

data/lib/tinygql/lexer.rb

@@ -75,8 +75,6 @@ module TinyGQL
       o[Literals.const_get(n).ord] = n
     }
 
-    LIT = Regexp.union(Literals.constants.map { |n| Literals.const_get(n) })
-
     QUOTED_STRING = %r{#{QUOTE} ((?:#{STRING_CHAR})*) #{QUOTE}}x
     BLOCK_STRING = %r{
         #{BLOCK_QUOTE}
@@ -95,11 +93,10 @@ module TinyGQL
 
       @string = string
       @scan = StringScanner.new string
+      @start = nil
     end
 
-    def pos
-      @scan.pos
-    end
+    attr_reader :start
 
     def line
       @scan.string[0, @scan.pos].count("\n") + 1
@@ -109,60 +106,27 @@ module TinyGQL
       @scan.eos?
     end
 
-    KW_LUT = [:FRAGMENT,
-              :INTERFACE,
-              :MUTATION,
-              :EXTEND,
-              :FALSE,
-              :ENUM,
-              :TRUE,
-              :NULL,
-              nil,
-              nil,
-              nil,
-              nil,
-              nil,
-              nil,
-              :QUERY,
-              nil,
-              nil,
-              nil,
-              :REPEATABLE,
-              :IMPLEMENTS,
-              :INPUT,
-              :TYPE,
-              :SCHEMA,
-              nil,
-              nil,
-              nil,
-              :DIRECTIVE,
-              :UNION,
-              nil,
-              nil,
-              :SCALAR]
-
     def advance
       @scan.skip(IGNORE)
 
+      @start = @scan.pos
+
       case
       when @scan.eos?                      then false
       when @scan.skip(ELLIPSIS)            then :ELLIPSIS
-      when tok = LIT_NAME_LUT[@string.getbyte(@scan.pos)] then
+      when tok = LIT_NAME_LUT[@string.getbyte(@start)] then
         @scan.pos += 1
         tok
-      when @scan.skip(KW_RE) then
-        len = @scan.matched_size
+      when len = @scan.skip(KW_RE) then
         return :ON if len == 2
         return :SUBSCRIPTION if len == 12
 
-        pos = @scan.pos - len
+        pos = @start
 
-        # First 3 bytes are unique, so we'll hash on those
-        key = (@string.getbyte(pos + 2) << 16) |
-          (@string.getbyte(pos + 1) << 8) |
-          @string.getbyte(pos + 0)
+        # Second 2 bytes are unique, so we'll hash on those
+        key = (@string.getbyte(pos + 2) << 8) | @string.getbyte(pos + 1)
 
-        KW_LUT[hash(key)]
+        KW_LUT[_hash(key)]
       when @scan.skip(IDENTIFIER)    then :IDENTIFIER
       when @scan.skip(BLOCK_STRING)  then :STRING
       when @scan.skip(QUOTED_STRING) then :STRING
@@ -315,10 +279,41 @@ module TinyGQL
       lines.size > 1 ? lines.join("\n") : (lines.first || "".dup)
     end
 
-    def hash key
-      # Constant came from perfect hash generation
-      m = key * 145291
-      (m >> 28) & 0x1f
+    KW_LUT =[:INTERFACE,
+             :MUTATION,
+             :EXTEND,
+             :FALSE,
+             :ENUM,
+             :TRUE,
+             :NULL,
+             nil,
+             nil,
+             nil,
+             nil,
+             nil,
+             nil,
+             nil,
+             :QUERY,
+             nil,
+             nil,
+             :REPEATABLE,
+             :IMPLEMENTS,
+             :INPUT,
+             :TYPE,
+             :SCHEMA,
+             nil,
+             nil,
+             nil,
+             :DIRECTIVE,
+             :UNION,
+             nil,
+             nil,
+             :SCALAR,
+             nil,
+             :FRAGMENT]
+
+    def _hash key
+      (key * 18592990) >> 27 & 0x1f
     end
   end
 end

data/lib/tinygql/nodes.rb

@@ -3,16 +3,16 @@ module TinyGQL
     class Node
       include Enumerable
 
-      # `pos` is the position of the _end_ of a token
-      attr_reader :pos
+      # `start` is the start position of this node in the source document
+      attr_reader :start
 
-      def initialize pos
-        @pos = pos
+      def initialize start
+        @start = start
       end
 
       # Return the line of this node given `doc`
       def line doc
-        doc[0, @pos].count("\n") + 1
+        doc[0, @start].count("\n") + 1
       end
 
       def document?; false; end

data/lib/tinygql/nodes.rb.erb

@@ -3,16 +3,16 @@ module TinyGQL
     class Node
       include Enumerable
 
-      # `pos` is the position of the _end_ of a token
-      attr_reader :pos
+      # `start` is the start position of this node in the source document
+      attr_reader :start
 
-      def initialize pos
-        @pos = pos
+      def initialize start
+        @start = start
       end
 
       # Return the line of this node given `doc`
       def line doc
-        doc[0, @pos].count("\n") + 1
+        doc[0, @start].count("\n") + 1
       end
 
 <%- nodes.each do |node| -%>

data/lib/tinygql/parser.rb

@@ -13,7 +13,7 @@ module TinyGQL
 
     def initialize doc
       @lexer = Lexer.new doc
-      @token_name = @lexer.advance
+      accept_token
     end
 
     def parse
@@ -25,7 +25,7 @@ module TinyGQL
     attr_reader :token_name
 
     def pos
-      @lexer.pos
+      @lexer.start
     end
 
     def document
@@ -477,7 +477,17 @@ module TinyGQL
     end
 
     def operation_type
-      expect_tokens([:QUERY, :MUTATION, :SUBSCRIPTION])
+      val = if at?(:QUERY)
+        "query"
+      elsif at?(:MUTATION)
+        "mutation"
+      elsif at?(:SUBSCRIPTION)
+        "subscription"
+      else
+        expect_token(:QUERY)
+      end
+      accept_token
+      val
     end
 
     def directives
@@ -528,7 +538,7 @@ module TinyGQL
 
     def variable_definition
       loc = pos
-      var = variable
+      var = variable if at?(:VAR_SIGN)
       expect_token(:COLON)
       type = self.type
       default_value = if at?(:EQUALS)
@@ -608,7 +618,7 @@ module TinyGQL
         accept_token
         Nodes::BooleanValue.new(pos, "false")
       else
-        expect_tokens([:TRUE, :FALSE])
+        expect_token(:TRUE)
       end
     end
 
@@ -643,15 +653,14 @@ module TinyGQL
     end
 
     def variable
-      return unless at?(:VAR_SIGN)
       loc = pos
-      accept_token
+      expect_token(:VAR_SIGN)
       Nodes::Variable.new loc, name
     end
 
     def name
       case token_name
-      when :IDENTIFIER then accept_token_value
+      when :IDENTIFIER then expect_token_value(:IDENTIFIER)
       when :TYPE then
         accept_token
         "type"
@@ -662,7 +671,7 @@ module TinyGQL
         accept_token
         "input"
       else
-        expect_token_value(:IDENTIFIER)
+        expect_token(:IDENTIFIER)
       end
     end
 
@@ -670,13 +679,6 @@ module TinyGQL
       @token_name = @lexer.advance
     end
 
-    # Only use when we care about the accepted token's value
-    def accept_token_value
-      token_value = @lexer.token_value
-      accept_token
-      token_value
-    end
-
     def expect_token tok
       unless at?(tok)
         raise UnexpectedToken, "Expected token #{tok}, actual: #{token_name} #{@lexer.token_value} line: #{@lexer.line}"
@@ -697,15 +699,6 @@ module TinyGQL
       token_value
     end
 
-    def expect_tokens toks
-      token_value = @lexer.token_value
-      unless toks.any? { |tok| at?(tok) }
-        raise UnexpectedToken, "Expected token #{tok}, actual: #{token_name}"
-      end
-      accept_token
-      token_value
-    end
-
     def at? tok
       token_name == tok
     end

data/lib/tinygql/version.rb

@@ -1,3 +1,3 @@
 module TinyGQL
-  VERSION = '0.2.0'
+  VERSION = '0.3.0'
 end

data/test/lexer_test.rb

@@ -179,5 +179,28 @@ eod
 
       assert_equal words.map { |x| [x.upcase.to_sym, x] }, toks
     end
+
+    def test_looks_like_kw
+      words = ["fragment", "fragments"]
+      doc = words.join(" ")
+
+      lexer = Lexer.new doc
+      toks = []
+      while tok = lexer.advance
+        toks << tok
+      end
+
+      assert_equal [:FRAGMENT, :IDENTIFIER], toks
+    end
+
+    def test_num_with_dots
+      lexer = Lexer.new "1...2"
+      toks = []
+      while tok = lexer.advance
+        toks << tok
+      end
+
+      assert_equal [:INT, :ELLIPSIS, :INT], toks
+    end
   end
 end

data/test/parser_test.rb

@@ -58,8 +58,8 @@ mutation {
 eod
       parser = Parser.new doc
       ast = parser.parse
-      expected = ["likeStory", "story", "likeCount"].map { |str| doc.index(str) + str.bytesize }
-      assert_equal expected, ast.find_all(&:field?).map(&:pos)
+      expected = ["likeStory", "story", "likeCount"].map { |str| doc.index(str) }
+      assert_equal expected, ast.find_all(&:field?).map(&:start)
       assert_equal [2, 3, 4], ast.find_all(&:field?).map { |n| n.line(doc) }
     end
 

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: tinygql
 version: !ruby/object:Gem::Version
-  version: 0.2.0
+  version: 0.3.0
 platform: ruby
 authors:
 - Aaron Patterson
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2023-08-28 00:00:00.000000000 Z
+date: 2023-08-30 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake
@@ -53,6 +53,7 @@ files:
 - Rakefile
 - benchmark/fixtures/negotiate.gql
 - bin/bench.rb
+- bin/make_hash.rb
 - bin/profile.rb
 - lib/tinygql.rb
 - lib/tinygql/lexer.rb