Class: GraphQL::Execution::Multiplex Private

Inherits:
Object
  • Object
show all
Includes:
Tracing::Traceable
Defined in:
lib/graphql/execution/multiplex.rb

Overview

This class is part of a private API. You should avoid using this class if possible, as it may be removed or be changed in the future.

Execute multiple queries under the same multiplex "umbrella". They can share a batching context and reduce redundant database hits.

The flow is:

  • Multiplex instrumentation setup
  • Query instrumentation setup
  • Analyze the multiplex + each query
  • Begin each query
  • Resolve lazy values, breadth-first across all queries
  • Finish each query (eg, get errors)
  • Query instrumentation teardown
  • Multiplex instrumentation teardown

If one query raises an application error, all queries will be in undefined states.

Validation errors and GraphQL::ExecutionErrors are handled in isolation: one of these errors in one query will not affect the other queries.

See Also:

  • for public API

Constant Summary collapse

NO_OPERATION =

This constant is part of a private API. You should avoid using this constant if possible, as it may be removed or be changed in the future.

Used internally to signal that the query shouldn't be executed

{}.freeze

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Methods included from Tracing::Traceable

#trace

Constructor Details

#initialize(schema:, queries:, context:, max_complexity:) ⇒ Multiplex

This method is part of a private API. You should avoid using this method if possible, as it may be removed or be changed in the future.

Returns a new instance of Multiplex.


33
34
35
36
37
38
39
40
41
42
43
44
45
# File 'lib/graphql/execution/multiplex.rb', line 33

def initialize(schema:, queries:, context:, max_complexity:)
  @schema = schema
  @queries = queries
  @queries.each { |q| q.multiplex = self }
  @context = context
  @dataloader = @context[:dataloader] ||= @schema.dataloader_class.new
  @tracers = schema.tracers + (context[:tracers] || [])
  # Support `context: {backtrace: true}`
  if context[:backtrace] && !@tracers.include?(GraphQL::Backtrace::Tracer)
    @tracers << GraphQL::Backtrace::Tracer
  end
  @max_complexity = max_complexity
end

Instance Attribute Details

#contextObject (readonly)

This method is part of a private API. You should avoid using this method if possible, as it may be removed or be changed in the future.


32
33
34
# File 'lib/graphql/execution/multiplex.rb', line 32

def context
  @context
end

#dataloaderObject (readonly)

This method is part of a private API. You should avoid using this method if possible, as it may be removed or be changed in the future.


32
33
34
# File 'lib/graphql/execution/multiplex.rb', line 32

def dataloader
  @dataloader
end

#max_complexityObject (readonly)

This method is part of a private API. You should avoid using this method if possible, as it may be removed or be changed in the future.


32
33
34
# File 'lib/graphql/execution/multiplex.rb', line 32

def max_complexity
  @max_complexity
end

#queriesObject (readonly)

This method is part of a private API. You should avoid using this method if possible, as it may be removed or be changed in the future.


32
33
34
# File 'lib/graphql/execution/multiplex.rb', line 32

def queries
  @queries
end

#schemaObject (readonly)

This method is part of a private API. You should avoid using this method if possible, as it may be removed or be changed in the future.


32
33
34
# File 'lib/graphql/execution/multiplex.rb', line 32

def schema
  @schema
end

Class Method Details

.begin_query(results, idx, query, multiplex) ⇒ Object

This method is part of a private API. You should avoid using this method if possible, as it may be removed or be changed in the future.

Parameters:


112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
# File 'lib/graphql/execution/multiplex.rb', line 112

def begin_query(results, idx, query, multiplex)
  operation = query.selected_operation
  result = if operation.nil? || !query.valid? || query.context.errors.any?
    NO_OPERATION
  else
    begin
      query.schema.query_execution_strategy.begin_query(query, multiplex)
    rescue GraphQL::ExecutionError => err
      query.context.errors << err
      NO_OPERATION
    end
  end
  results[idx] = result
  nil
end

.run_all(schema, query_options, context: {}, max_complexity: schema.max_complexity) ⇒ Array<Hash>

This method is part of a private API. You should avoid using this method if possible, as it may be removed or be changed in the future.

Returns One result per query.

Parameters:

  • schema (GraphQL::Schema)
  • queries (Array<GraphQL::Query, Hash>)
  • context (Hash) (defaults to: {})
  • max_complexity (Integer, nil) (defaults to: schema.max_complexity)

Returns:

  • (Array<Hash>)

    One result per query


53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
# File 'lib/graphql/execution/multiplex.rb', line 53

def run_all(schema, query_options, context: {}, max_complexity: schema.max_complexity)
  queries = query_options.map do |opts|
    case opts
    when Hash
      GraphQL::Query.new(schema, nil, **opts)
    when GraphQL::Query
      opts
    else
      raise "Expected Hash or GraphQL::Query, not #{opts.class} (#{opts.inspect})"
    end
  end

  multiplex = self.new(schema: schema, queries: queries, context: context, max_complexity: max_complexity)
  multiplex.trace("execute_multiplex", { multiplex: multiplex }) do
    GraphQL::Execution::Instrumentation.apply_instrumenters(multiplex) do
      schema = multiplex.schema
      multiplex_analyzers = schema.multiplex_analyzers
      if multiplex.max_complexity
        multiplex_analyzers += [GraphQL::Analysis::AST::MaxQueryComplexity]
      end

      schema.analysis_engine.analyze_multiplex(multiplex, multiplex_analyzers)

      begin
        multiplex.schema.query_execution_strategy.begin_multiplex(multiplex)
        # Do as much eager evaluation of the query as possible
        results = []
        queries.each_with_index do |query, idx|
          multiplex.dataloader.append_job { begin_query(results, idx, query, multiplex) }
        end

        multiplex.dataloader.run

        # Then, work through lazy results in a breadth-first way
        multiplex.dataloader.append_job {
          multiplex.schema.query_execution_strategy.finish_multiplex(results, multiplex)
        }
        multiplex.dataloader.run

        # Then, find all errors and assign the result to the query object
        results.each_with_index do |data_result, idx|
          query = queries[idx]
          finish_query(data_result, query, multiplex)
          # Get the Query::Result, not the Hash
          results[idx] = query.result
        end

        results
      rescue Exception
        # TODO rescue at a higher level so it will catch errors in analysis, too
        # Assign values here so that the query's `@executed` becomes true
        queries.map { |q| q.result_values ||= {} }
        raise
      end
    end
  end
end