elasticgraph-graphql 0.19.1.1 → 0.19.2.0

data/lib/elastic_graph/graphql/resolvers/nested_relationships_source.rb

@@ -0,0 +1,324 @@
+# Copyright 2024 - 2025 Block, Inc.
+#
+# Use of this source code is governed by an MIT-style
+# license that can be found in the LICENSE file or at
+# https://opensource.org/licenses/MIT.
+#
+# frozen_string_literal: true
+
+require "elastic_graph/graphql/resolvers/query_source"
+
+module ElasticGraph
+  class GraphQL
+    module Resolvers
+      # A GraphQL dataloader responsible for solving a thorny N+1 query problem related to our `NestedRelationships` resolver.
+      # The `QuerySource` dataloader implements a basic batching optimization: multiple datastore queries are batched up into
+      # a single `msearch` call against the dataastore. This is significantly better than submitting a separate request per
+      # query, but is still not optimal--the datastore still must execute N different queries, which could cause significant load.
+      #
+      # A significantly improved optimization is possible in one particular situation from our `NestedRelationships` resolver.
+      # Here's an example of that situation:
+      #
+      #   - `Part` documents are indexed in a `parts` index and `Manufacturer` documents are indexed in a `manufacturers` index.
+      #   - `Part.manufacturer` is defined as: `t.relates_to_one "manufacturer", "Manufacturer", via: "manufacturer_id", dir: :out`.
+      #   - We are processing a GraphQL query like this: `parts(first: 10) { nodes { manufacturer { name } } }`.
+      #   - For each of the 10 parts, the `NestedRelationships` resolver has to resolve its related `Part.manufacturer`.
+      #   - Without the optimization provided by this class, `NestedRelationships` would have to execute 10 different queries,
+      #     each of which is identical except for a different filter: `{id: {equal_to_any_of: [part.manufacturer_id]}}`.
+      #   - Instead of executing this as 10 different queries, we can instead execute it as one query with this combined filter:
+      #     `{id: {equal_to_any_of: [part1.manufacturer_id, ..., part10.manufacturer_id]}}`
+      #   - When we do this, we get a single response, but `NestedRelationships` expects a separate response for each one.
+      #   - To satisfy that, we can split the single response into 10 different responses (one per filter).
+      #
+      # This optimization, when we can apply it, results in much less load on the datastore. In addition, it also helps to reduce
+      # the amount of overhead imposed by ElasticGraph. Profiling has shown that significant overhead is incurred when we repeatedly
+      # merge filters into a query (e.g. `query.merge_with(filters: [{id: {equal_to_any_of: [part.manufacturer_id]}}])` 10 times to
+      # produce 10 different queries). This optimization also avoids that overhead.
+      #
+      # Note: while the comments discuss the examples in terms of _parent objects_, in the implementation, we deal with id sets.
+      # A set of ids is contributed by each parent object.
+      class NestedRelationshipsSource < ::GraphQL::Dataloader::Source
+        # The optimization implemented by this class is not guaranteed to get all expected results in a single query for cases where
+        # the sorted search results are not well-distributed among each of the parent objects while we're resolving a `relates_to_many`
+        # field. (See the comments on `fetch_via_single_query_with_merged_filters` for a detailed description of when this occurs).
+        #
+        # To deal with this situation, we retry the query for just the parent objects which may have incomplete results. However,
+        # each attempt is run in serial, and we want to put a strict upper bound on how many attempts are made. This constant defines
+        # the maximum number of optimized attempts we allow.
+        #
+        # When exceeded, we fall back to building and executing a separate query (via a single `msearch` request) for each parent object.
+        MAX_OPTIMIZED_ATTEMPTS = 3
+
+        # Reattempts are less likely to be needed when we execute the query with a larger `size`, because we are more likely to get back
+        # complete results for each parent object. This multiplier is applied to the requested size to achieve that.
+        #
+        # 4 was chosen somewhat arbitrarily, but should make reattempts needed much less often while avoiding asking for an unreasonably
+        # large number of results.
+        #
+        # Note: asking the datastore for a larger `size` is quite a bit more efficient than needing to execute more queries.
+        # Once the datastore has gone to the spot in its inverted index with the matching documents, asking for more results
+        # isn't particularly expensive, compared to needing to re-run an extra query.
+        EXTRA_SIZE_MULTIPLIER = 4
+
+        def initialize(query:, join:, context:, monotonic_clock:, mode:)
+          @query = query
+          @join = join
+          @filter_id_field_name_path = @join.filter_id_field_name.split(".")
+          @context = context
+          @schema_element_names = @context.fetch(:schema_element_names)
+          @logger = context.fetch(:logger)
+          @monotonic_clock = monotonic_clock
+          @mode = mode
+        end
+
+        def fetch(id_sets)
+          return fetch_original(id_sets) unless can_merge_filters?
+
+          case @mode
+          when :original
+            fetch_original(id_sets)
+          when :comparison
+            fetch_comparison(id_sets)
+          else
+            fetch_optimized(id_sets)
+          end
+        end
+
+        def self.execute_one(ids, query:, join:, context:, monotonic_clock:, mode:)
+          context.dataloader.with(self, query:, join:, context:, monotonic_clock:, mode:).load(ids)
+        end
+
+        private
+
+        def fetch_optimized(id_sets)
+          attempt_count = 0
+          duration_ms, responses_by_id_set = time_duration do
+            fetch_via_single_query_with_merged_filters(id_sets) { attempt_count += 1 }
+          end
+
+          if id_sets.size > 1
+            @logger.info({
+              "message_type" => "NestedRelationshipsMergedQueries",
+              "field" => @join.field.description,
+              "optimized_attempt_count" => [attempt_count, MAX_OPTIMIZED_ATTEMPTS].min,
+              "degraded_to_separate_queries" => (attempt_count > MAX_OPTIMIZED_ATTEMPTS),
+              "id_set_count" => id_sets.size,
+              "total_id_count" => id_sets.reduce(:union).size,
+              "duration_ms" => duration_ms
+            })
+          end
+
+          id_sets.map { |id_set| responses_by_id_set.fetch(id_set) }
+        end
+
+        def fetch_original(id_sets, requested_fields: [])
+          fetch_via_separate_queries(id_sets, requested_fields: requested_fields)
+        end
+
+        def fetch_comparison(id_sets)
+          # Note: we'd ideally run both versions of the logic in parallel, but our attempts to do that resulted in errors
+          # because of the fiber context in which dataloaders run.
+          original_duration_ms, original_results = time_duration do
+            # In the `fetch_optimized` implementation, we request this extra field. We don't need it for
+            # the original implementation (so `fetch_original` doesn't also request that field...) but for
+            # the purposes of comparison we need to request it so that the document payloads will have the
+            # same fields.
+            #
+            # Note: we don't add the requested field if we have only a single id set, in order to align with
+            # the short-circuiting logic in `fetch_via_single_query_with_merged_filters`. Otherwise, any time
+            # we have a single id set we always get reported differences which are not actually real!
+            requested_fields = (id_sets.size > 1) ? [@join.filter_id_field_name] : [] # : ::Array[::String]
+            fetch_original(id_sets, requested_fields: requested_fields)
+          end
+
+          optimized_duration_ms, optimized_results = time_duration do
+            fetch_optimized(id_sets)
+          end
+
+          # To see if we got the same results we only look at the documents, because we expect differences outside
+          # of the documents--for example, the `SearchResponse#metadata` will report different `took` values.
+          got_same_results = original_results.map(&:documents) == optimized_results.map(&:documents)
+          message = {
+            "message_type" => "NestedRelationshipsComparisonResults",
+            "field" => @join.field.description,
+            "original_duration_ms" => original_duration_ms,
+            "optimized_duration_ms" => optimized_duration_ms,
+            "optimized_faster" => (optimized_duration_ms < original_duration_ms),
+            "id_set_count" => id_sets.size,
+            "total_id_count" => id_sets.reduce(:union).size,
+            "got_same_results" => got_same_results
+          }
+
+          if got_same_results
+            @logger.info(message)
+          else
+            @logger.error(message.merge({
+              "original_documents" => loggable_results(original_results),
+              "optimized_documents" => loggable_results(optimized_results)
+            }))
+          end
+
+          original_results
+        end
+
+        # For "simple", document-based queries, we can safely merge filters. However, this cannot be done safely when the response
+        # cannot safely be "pulled part" into the bits that apply to a particular set of ids for a parent object. Specifically:
+        #
+        #   - If `total_document_count_needed` is true, we can't merge filters, because there's no way to get a separate count
+        #     for each parent object unless we execute separate queries (or combine them into a grouped aggregation count query,
+        #     but that requires a much more challenging transformation of the query and response).
+        #   - If the query has any `aggregations`, we likewise can't merge the filters, because we have no way to "pull apart"
+        #     the aggregations response.
+        def can_merge_filters?
+          !@query.total_document_count_needed && @query.aggregations.empty?
+        end
+
+        # Executes a single query that contains a merged filter from the set union of the given `id_sets`.
+        # This merged query is (theoretically) capable of getting all the results we're looking for in a
+        # single query, which is much more efficient than building and performing a separate query for each
+        # id set. We can use `search_response.filter_results(id_set)` with each id set to get a
+        # response with the documents filtered down to just the ones that match the id set. (Essentially,
+        # this is the response we would have gotten if we had executed a separate query for the id set).
+        #
+        # However, it is not guaranteed that we will get back complete results with this approach. Consider this example:
+        #
+        #   - The datastore has 50 documents that match `id_set_1`, and 50 that match `id_set_2`.
+        #   - The requested size of `@query` is 10 (meaning the client expects the first 10 results matching `id_set_1` and
+        #     the first 10 results matching `id_set_2).
+        #   - All 50 documents that match `id_set_1` sort before all 50 documents that match `id_set_2`.
+        #   - When we execute our merged query filtering on the `union(id_set_1, id_set_2)` set, we ask for
+        #     20 documents (since we want 10 for `id_set_1` and 10 for `id_set_2`).
+        #   - ...but we get back 20 documents for `id_set_1` and 0 documents for `id_set_2`.
+        #
+        # There is no way to guarantee that we get back the desired number of results for each id set unless we build and
+        # execute a separate query per id set, which is inefficient (in some situations, it causes one GraphQL query to
+        # execute hundreds of queries against the datastore!).
+        #
+        # To deal with this possibility, this method takes an iterative approach:
+        #
+        #   - It builds and executes an initial optimized merged query, with a large `size_multiplier` which gives us a good bit of
+        #     "headroom" for this kind of situation. In the example above, if we requested 60 results from the datastore, we'd be
+        #     able to get the 10 results for both id sets we are looking for--50 for `id_set_1` nad 10 for `id_set_2`.
+        #   - It then inspects the response. If the datastore returned fewer results than we asked for, then there are no missing
+        #     results and we can trust that we got all the results we would have gotten if we had executed a separate query per
+        #     id set.
+        #   - If we got back the number of results we asked for, then it's possible that we've run into this situation. We need
+        #     to inspect each filtered response produced for each id set to see if more results were expected.
+        #     - Note: the fact that more results were expected doesn't necessarily mean there are more results. But we have no way
+        #       to tell for sure without querying the datastore again, so we err on the side of safety and treat this kind of response
+        #       as being incomplete.
+        #   - For each id set that appears to be incomplete, we try again. But on the next attempt, we exclude the id sets
+        #     which got a complete set of results.
+        #   - This may cause us to iterate a couple of times (which could make the single GraphQL query we are processing slower than
+        #     it would have been without this optimization, particularly if the datastore was not under any other load...) but we expect
+        #     it to make a big difference in the amount of load we put on the datastore, and that helps _all_ query traffic to be more
+        #     performant overall.
+        def fetch_via_single_query_with_merged_filters(id_sets, remaining_attempts: MAX_OPTIMIZED_ATTEMPTS)
+          yield # yield to signal an attempt
+
+          # Fallback to executing separate queries when one of the following occurs:
+          #
+          #   - We lack multiple sets of ids.
+          #   - We have exhausted our MAX_OPTIMIZED_ATTEMPTS.
+          if id_sets.size < 2 || remaining_attempts < 1
+            return id_sets.zip(fetch_via_separate_queries(id_sets)).to_h
+          end
+
+          # First, we build a combined query with filters that account for all ids we are filtering on from all `id_sets`.
+          filtered_query = @query.merge_with(
+            filters: filters_for(id_sets.reduce(:union)),
+            requested_fields: [@join.filter_id_field_name],
+            # We need to request a larger size than `@query` originally had. If the original size was `10` and we have
+            # 5 sets of ids, then, at a minimum, we need to request 50 results (10 results for each id set).
+            #
+            # In addition, we apply `EXTRA_SIZE_MULTIPLIER` to increase the size further and make it less likely that
+            # we we get incomplete results and have to retry.
+            size_multiplier: id_sets.size * EXTRA_SIZE_MULTIPLIER
+          )
+
+          # Then we execute that combined query.
+          response = QuerySource.execute_one(filtered_query, for_context: @context)
+
+          # Next, we produce a separate response for each id set by filtering the results to the ones that match the ids in the set.
+          filtered_responses_by_id_set = id_sets.to_h do |id_set|
+            filtered_response = response.filter_results(@filter_id_field_name_path, id_set, @query.effective_size)
+            [id_set, filtered_response]
+          end
+
+          # If our merged/filtered query got back fewer results than we requested, then no matching results are missing,
+          # and we know that we've gotten complete results for all id sets.
+          if response.size < filtered_query.effective_size
+            return filtered_responses_by_id_set
+          end
+
+          # Since our `filtered_query` got back as many results as we asked for, there may be additional matching results that
+          # were not returned, and some id sets may have gotten fewer results than requested by the client.
+          # Here we determine which id sets that applies to.
+          id_sets_with_apparently_incomplete_results = filtered_responses_by_id_set.filter_map do |id_set, filtered_response|
+            id_set if filtered_response.size < @query.effective_size
+          end
+
+          # Then we try again, excluding the id sets which have already gotten complete results.
+          another_attempt_results = fetch_via_single_query_with_merged_filters(
+            id_sets_with_apparently_incomplete_results,
+            remaining_attempts: remaining_attempts - 1
+          ) { yield }
+
+          # Finally, we merge the results.
+          filtered_responses_by_id_set.merge(another_attempt_results)
+        end
+
+        def fetch_via_separate_queries(id_sets, requested_fields: [])
+          queries = id_sets.map do |ids|
+            @query.merge_with(filters: filters_for(ids), requested_fields: requested_fields)
+          end
+
+          results = QuerySource.execute_many(queries, for_context: @context)
+          queries.map { |q| results.fetch(q) }
+        end
+
+        def filters_for(ids)
+          join_filter = build_filter(@join.filter_id_field_name, nil, @join.foreign_key_nested_paths, ids.to_a)
+
+          if @join.additional_filter.empty?
+            [join_filter]
+          else
+            [join_filter, @join.additional_filter]
+          end
+        end
+
+        def build_filter(path, previous_nested_path, nested_paths, ids)
+          next_nested_path, *rest_nested_paths = nested_paths
+
+          if next_nested_path.nil?
+            path = path.delete_prefix("#{previous_nested_path}.") if previous_nested_path
+            {path => {@schema_element_names.equal_to_any_of => ids}}
+          else
+            sub_filter = build_filter(path, next_nested_path, rest_nested_paths, ids)
+            next_nested_path = next_nested_path.delete_prefix("#{previous_nested_path}.") if previous_nested_path
+            {next_nested_path => {@schema_element_names.any_satisfy => sub_filter}}
+          end
+        end
+
+        def time_duration
+          start_time = @monotonic_clock.now_in_ms
+          result = yield
+          stop_time = @monotonic_clock.now_in_ms
+          [stop_time - start_time, result]
+        end
+
+        # Converts the given list of responses into a format we can safely log when we are logging
+        # response differences. We include the `id` (to identify the document) and the `hash` (so
+        # we can tell if the payload of a document differed, without logging the contents of that
+        # payload).
+        def loggable_results(responses)
+          responses.map do |response|
+            response.documents.map do |doc|
+              "#{doc.id} (hash: #{doc.hash})"
+            end
+          end
+        end
+      end
+    end
+  end
+end

data/lib/elastic_graph/graphql/resolvers/object.rb

@@ -0,0 +1,36 @@
+# Copyright 2024 - 2025 Block, Inc.
+#
+# Use of this source code is governed by an MIT-style
+# license that can be found in the LICENSE file or at
+# https://opensource.org/licenses/MIT.
+#
+# frozen_string_literal: true
+
+module ElasticGraph
+  class GraphQL
+    module Resolvers
+      # Resolvers which just delegates to `object` for resolving.
+      module Object
+        class WithLookahead
+          def initialize(elasticgraph_graphql:, config:)
+            # Nothing to initialize, but needs to be defined to satisfy the resolver interface.
+          end
+
+          def resolve(field:, object:, args:, context:, lookahead:)
+            object.resolve(field: field, object: object, args: args, context: context, lookahead: lookahead)
+          end
+        end
+
+        class WithoutLookahead
+          def initialize(elasticgraph_graphql:, config:)
+            # Nothing to initialize, but needs to be defined to satisfy the resolver interface.
+          end
+
+          def resolve(field:, object:, args:, context:)
+            object.resolve(field: field, object: object, args: args, context: context)
+          end
+        end
+      end
+    end
+  end
+end

data/lib/elastic_graph/graphql/resolvers/query_adapter.rb

@@ -1,4 +1,4 @@
-# Copyright 2024 Block, Inc.
+# Copyright 2024 - 2025 Block, Inc.
 #
 # Use of this source code is governed by an MIT-style
 # license that can be found in the LICENSE file or at
@@ -17,7 +17,7 @@ module ElasticGraph
           @datastore_query_adapters = datastore_query_adapters
         end
 
-        def build_query_from(field:, args:, lookahead:, context: {})
+        def build_query_from(field:, args:, lookahead:, context:)
           monotonic_clock_deadline = context[:monotonic_clock_deadline]
 
           # Building an `DatastoreQuery` is not cheap; we do a lot of work to:

data/lib/elastic_graph/graphql/resolvers/query_source.rb

@@ -1,4 +1,4 @@
-# Copyright 2024 Block, Inc.
+# Copyright 2024 - 2025 Block, Inc.
 #
 # Use of this source code is governed by an MIT-style
 # license that can be found in the LICENSE file or at
@@ -11,11 +11,14 @@ require "graphql"
 module ElasticGraph
   class GraphQL
     module Resolvers
-      # Provides a way to avoid N+1 query problems by batching up multiple
-      # datastore queries into one `msearch` call. In general, it is recommended
+      # Provides a way to avoid N+1 request problems by batching up multiple
+      # datastore queries into one `msearch` request. In general, it is recommended
       # that you use this from any resolver that needs to query the datastore, to
       # maximize our ability to combine multiple datastore requests. Importantly,
       # this should never be instantiated directly; instead use the `execute` method from below.
+      #
+      # Note: `NestedRelationshipsSource` implements further optimizations on top of this, and should
+      # be used rather than this class when applicable.
       class QuerySource < ::GraphQL::Dataloader::Source
         def initialize(datastore_router, query_tracker)
           @datastore_router = datastore_router

data/lib/elastic_graph/graphql/resolvers/relay_connection/array_adapter.rb

@@ -1,4 +1,4 @@
-# Copyright 2024 Block, Inc.
+# Copyright 2024 - 2025 Block, Inc.
 #
 # Use of this source code is governed by an MIT-style
 # license that can be found in the LICENSE file or at

data/lib/elastic_graph/graphql/resolvers/relay_connection/generic_adapter.rb

@@ -1,4 +1,4 @@
-# Copyright 2024 Block, Inc.
+# Copyright 2024 - 2025 Block, Inc.
 #
 # Use of this source code is governed by an MIT-style
 # license that can be found in the LICENSE file or at

data/lib/elastic_graph/graphql/resolvers/relay_connection/page_info.rb

@@ -1,4 +1,4 @@
-# Copyright 2024 Block, Inc.
+# Copyright 2024 - 2025 Block, Inc.
 #
 # Use of this source code is governed by an MIT-style
 # license that can be found in the LICENSE file or at

data/lib/elastic_graph/graphql/resolvers/relay_connection/search_response_adapter_builder.rb

@@ -1,4 +1,4 @@
-# Copyright 2024 Block, Inc.
+# Copyright 2024 - 2025 Block, Inc.
 #
 # Use of this source code is governed by an MIT-style
 # license that can be found in the LICENSE file or at

data/lib/elastic_graph/graphql/resolvers/relay_connection.rb

@@ -1,4 +1,4 @@
-# Copyright 2024 Block, Inc.
+# Copyright 2024 - 2025 Block, Inc.
 #
 # Use of this source code is governed by an MIT-style
 # license that can be found in the LICENSE file or at

data/lib/elastic_graph/graphql/resolvers/resolvable_value.rb

@@ -1,4 +1,4 @@
-# Copyright 2024 Block, Inc.
+# Copyright 2024 - 2025 Block, Inc.
 #
 # Use of this source code is governed by an MIT-style
 # license that can be found in the LICENSE file or at
@@ -27,16 +27,11 @@ module ElasticGraph
           end
         end
 
-        def resolve(field:, object:, context:, args:, lookahead:)
+        def resolve(field:, object:, context:, args:)
           method_name = canonical_name_for(field.name, "Field")
           public_send(method_name, **args_to_canonical_form(args))
         end
 
-        def can_resolve?(field:, object:)
-          method_name = schema_element_names.canonical_name_for(field.name)
-          !!method_name && respond_to?(method_name)
-        end
-
         private
 
         def args_to_canonical_form(args)

data/lib/elastic_graph/graphql/scalar_coercion_adapters/cursor.rb

@@ -1,4 +1,4 @@
-# Copyright 2024 Block, Inc.
+# Copyright 2024 - 2025 Block, Inc.
 #
 # Use of this source code is governed by an MIT-style
 # license that can be found in the LICENSE file or at

data/lib/elastic_graph/graphql/scalar_coercion_adapters/date.rb

@@ -1,4 +1,4 @@
-# Copyright 2024 Block, Inc.
+# Copyright 2024 - 2025 Block, Inc.
 #
 # Use of this source code is governed by an MIT-style
 # license that can be found in the LICENSE file or at

data/lib/elastic_graph/graphql/scalar_coercion_adapters/date_time.rb

@@ -1,4 +1,4 @@
-# Copyright 2024 Block, Inc.
+# Copyright 2024 - 2025 Block, Inc.
 #
 # Use of this source code is governed by an MIT-style
 # license that can be found in the LICENSE file or at

data/lib/elastic_graph/graphql/scalar_coercion_adapters/local_time.rb

@@ -1,4 +1,4 @@
-# Copyright 2024 Block, Inc.
+# Copyright 2024 - 2025 Block, Inc.
 #
 # Use of this source code is governed by an MIT-style
 # license that can be found in the LICENSE file or at

data/lib/elastic_graph/graphql/scalar_coercion_adapters/longs.rb

@@ -1,4 +1,4 @@
-# Copyright 2024 Block, Inc.
+# Copyright 2024 - 2025 Block, Inc.
 #
 # Use of this source code is governed by an MIT-style
 # license that can be found in the LICENSE file or at

data/lib/elastic_graph/graphql/scalar_coercion_adapters/no_op.rb

@@ -1,4 +1,4 @@
-# Copyright 2024 Block, Inc.
+# Copyright 2024 - 2025 Block, Inc.
 #
 # Use of this source code is governed by an MIT-style
 # license that can be found in the LICENSE file or at

data/lib/elastic_graph/graphql/scalar_coercion_adapters/time_zone.rb

@@ -1,4 +1,4 @@
-# Copyright 2024 Block, Inc.
+# Copyright 2024 - 2025 Block, Inc.
 #
 # Use of this source code is governed by an MIT-style
 # license that can be found in the LICENSE file or at

data/lib/elastic_graph/graphql/scalar_coercion_adapters/untyped.rb

@@ -1,4 +1,4 @@
-# Copyright 2024 Block, Inc.
+# Copyright 2024 - 2025 Block, Inc.
 #
 # Use of this source code is governed by an MIT-style
 # license that can be found in the LICENSE file or at

data/lib/elastic_graph/graphql/scalar_coercion_adapters/valid_time_zones.rb

@@ -1,4 +1,4 @@
-# Copyright 2024 Block, Inc.
+# Copyright 2024 - 2025 Block, Inc.
 #
 # Use of this source code is governed by an MIT-style
 # license that can be found in the LICENSE file or at

data/lib/elastic_graph/graphql/schema/arguments.rb

@@ -1,4 +1,4 @@
-# Copyright 2024 Block, Inc.
+# Copyright 2024 - 2025 Block, Inc.
 #
 # Use of this source code is governed by an MIT-style
 # license that can be found in the LICENSE file or at

data/lib/elastic_graph/graphql/schema/enum_value.rb

@@ -1,4 +1,4 @@
-# Copyright 2024 Block, Inc.
+# Copyright 2024 - 2025 Block, Inc.
 #
 # Use of this source code is governed by an MIT-style
 # license that can be found in the LICENSE file or at

data/lib/elastic_graph/graphql/schema/field.rb

@@ -1,4 +1,4 @@
-# Copyright 2024 Block, Inc.
+# Copyright 2024 - 2025 Block, Inc.
 #
 # Use of this source code is governed by an MIT-style
 # license that can be found in the LICENSE file or at
@@ -18,37 +18,18 @@ module ElasticGraph
         # The type in which the field resides.
         attr_reader :parent_type
 
-        attr_reader :schema, :schema_element_names, :graphql_field, :name_in_index, :relation, :computation_detail
+        attr_reader :schema, :schema_element_names, :graphql_field, :name_in_index, :relation, :computation_detail, :resolver
 
-        def initialize(schema, parent_type, graphql_field, runtime_metadata)
+        def initialize(schema, parent_type, graphql_field, runtime_metadata, resolvers_needing_lookahead)
           @schema = schema
           @schema_element_names = schema.element_names
           @parent_type = parent_type
           @graphql_field = graphql_field
           @relation = runtime_metadata&.relation
           @computation_detail = runtime_metadata&.computation_detail
-          @name_in_index = runtime_metadata&.name_in_index&.to_sym || name
-
-          # Adds the :extras required by ElasticGraph. For now, this blindly adds `:lookahead`
-          # to each field so that we have access to what the child selections are, as described here:
-          #
-          # https://graphql-ruby.org/queries/lookahead
-          #
-          # Currently we only need this when building an `DatastoreQuery` (which is not done for all
-          # fields) so a future optimization may only add this to fields where we actually need it.
-          # For now we add it to all fields because it's simplest and it's not clear if there is
-          # any performance benefit to not adding it when we do not use it.
-          #
-          # Note: input fields do not respond to `extras`, which is why we do it conditionally here.
-          #
-          # Note: on GraphQL gem introspection types (e.g. `__Field`), the fields respond to `:extras`,
-          # but that later causes a weird error (`ArgumentError: unknown keyword: :lookahead`)
-          # when those types are accessed in a Query. We don't really want to mutate the fields on the
-          # built-in types by adding `:lookahead` so it's best to avoid setting that extra on the built
-          # in types.
-          if @graphql_field.respond_to?(:extras) && !BUILT_IN_TYPE_NAMES.include?(parent_type.name.to_s)
-            @graphql_field.extras([:lookahead])
-          end
+          @resolver = runtime_metadata&.resolver
+          @name_in_index = runtime_metadata&.name_in_index || name
+          @graphql_field.extras([:lookahead]) if resolvers_needing_lookahead.include?(@resolver)
         end
 
         def type
@@ -56,7 +37,11 @@ module ElasticGraph
         end
 
         def name
-          @name ||= @graphql_field.name.to_sym
+          @name ||= @graphql_field.name
+        end
+
+        def path_in_index
+          @path_in_index ||= name_in_index.split(".")
         end
 
         # Returns an object that knows how this field joins to its relation.
@@ -93,7 +78,7 @@ module ElasticGraph
           return [] if parent_type.relay_connection? || parent_type.relay_edge?
           return index_id_field_names_for_relation if relation_join
 
-          [name_in_index.to_s]
+          [name_in_index]
         end
 
         # Indicates this field should be hidden in the GraphQL schema so as to not be queryable.