graphql-stitching 1.2.1 → 1.2.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 4cb8c0d3b16cda5db8b82d0d0f9eba93a4537c3faa584b1b301dc84b34b5fd0e
-  data.tar.gz: 83f4b436479307ac8575f718ace5a82288024a6e5741f786151e55920d7ef61f
+  metadata.gz: 6753e205aac88d3dd3be1c906b743cc49ae0de84ff8ab56c7e9be6e889429aa3
+  data.tar.gz: 59440f9d70cb365ef124604c42012bf655934555dfb76905328368353f5e310b
 SHA512:
-  metadata.gz: e247e539e223a1fbefcd398c1aeb3940fa91320b81c45caeaa21b3ddca50631ffb6ddbd3963cde07f11609cbaacc1f814a167fefebab8f888979afe929bcb93f
-  data.tar.gz: 66de3bacb73749bd90b4d8e7ade4bf35cc144e0ae60b3ff1c81a5f548b20d74e204454b6727875984da293b4b3853aa9bf6215283666f2b037a7dc148e8aa64b
+  metadata.gz: ac218d2218c14b8debff552ffb11e83e00293dfebb404d25296ed9e408232e488c44575c701ee3fd933331d84755062517f2c2da8df2db98ac7c5e8594595846
+  data.tar.gz: ec4f5051ebd511e72536c8e3246302079b2eb65ab5edab37c7c7d7de5488cdc7581318c69a0ea54047bd79ff42bf9e9c5af3434a27fc5667cf029340527c8117

data/README.md

@@ -78,9 +78,7 @@ While the `Client` constructor is an easy quick start, the library also has seve
 
 - [Composer](./docs/composer.md) - merges and validates many schemas into one supergraph.
 - [Supergraph](./docs/supergraph.md) - manages the combined schema, location routing maps, and executable resources. Can be exported, cached, and rehydrated.
-- [Request](./docs/request.md) - prepares a requested GraphQL document and variables for stitching.
-- [Planner](./docs/planner.md) - builds a cacheable query plan for a request document.
-- [Executor](./docs/executor.md) - executes a query plan with given request variables.
+- [Request](./docs/request.md) - manages the lifecycle of a stitched GraphQL request.
 - [HttpExecutable](./docs/http_executable.md) - proxies requests to remotes with multipart file upload support.
 
 ## Merged types
@@ -134,7 +132,7 @@ client = GraphQL::Stitching::Client.new(locations: {
     executable:  GraphQL::Stitching::HttpExecutable.new(url: "http://localhost:3001"),
   },
   catalog: {
-    schema: GraphQL::Schema.from_definition(shipping_schema),
+    schema: GraphQL::Schema.from_definition(catalog_schema),
     executable:  GraphQL::Stitching::HttpExecutable.new(url: "http://localhost:3002"),
   },
 })
@@ -155,7 +153,7 @@ type Query {
 * The `@stitch` directive is applied to a root query where the merged type may be accessed. The merged type identity is inferred from the field return.
 * The `key: "id"` parameter indicates that an `{ id }` must be selected from prior locations so it may be submitted as an argument to this query. The query argument used to send the key is inferred when possible ([more on arguments](#multiple-query-arguments) later).
 
-Each location that provides a unique variant of a type must provide at least one stitching query. The exception to this requirement are types that contain only a single key field:
+Each location that provides a unique variant of a type must provide at least one stitching query for the type. The exception to this requirement are [foreign key types](./docs/mechanics.md##modeling-foreign-keys-for-stitching) that contain only a single key field:
 
 ```graphql
 type Product {
@@ -163,7 +161,7 @@ type Product {
 }
 ```
 
-The above representation of a `Product` type provides no unique data beyond a key that is available in other locations. Thus, this representation will never require an inbound request to fetch it, and its stitching query may be omitted. This pattern of providing key-only types is very common in stitching: it allows a foreign key to be represented as an object stub that may be enriched by data collected from other locations.
+The above representation of a `Product` type provides no unique data beyond a key that is available in other locations. Thus, this representation will never require an inbound request to fetch it, and its stitching query may be omitted.
 
 #### List queries
 
@@ -427,6 +425,7 @@ The [Executor](./docs/executor.md) component builds atop the Ruby fiber-based im
 
 ## Additional topics
 
+- [Modeling foreign keys for stitching](./docs/mechanics.md##modeling-foreign-keys-for-stitching)
 - [Deploying a stitched schema](./docs/mechanics.md#deploying-a-stitched-schema)
 - [Field selection routing](./docs/mechanics.md#field-selection-routing)
 - [Root selection routing](./docs/mechanics.md#root-selection-routing)

data/docs/README.md

@@ -10,8 +10,6 @@ Major components include:
 - [Composer](./composer.md) - merges and validates many schemas into one graph.
 - [Supergraph](./supergraph.md) - manages the combined schema and location routing maps. Can be exported, cached, and rehydrated.
 - [Request](./request.md) - prepares a requested GraphQL document and variables for stitching.
-- [Planner](./planner.md) - builds a cacheable query plan for a request document.
-- [Executor](./executor.md) - executes a query plan with given request variables.
 - [HttpExecutable](./http_executable.md) - proxies requests to remotes with multipart file upload support.
 
 Additional topics:

data/docs/mechanics.md

@@ -1,5 +1,47 @@
 ## Schema Stitching, mechanics
 
+### Modeling foreign keys for stitching
+
+Foreign keys in a GraphQL schema typically look like the `Product.imageId` field here:
+
+```graphql
+# -- Products schema:
+
+type Product {
+  id: ID!
+  imageId: ID!
+}
+
+# -- Images schema:
+
+type Image {
+  id: ID!
+  url: String!
+}
+```
+
+However, this design does not lend itself to stitching where types need to _merge_ across locations. A simple schema refactor makes this foreign key more expressive as an entity type, and turns the key into an _object_ that will merge with analogous object types in other locations:
+
+```graphql
+# -- Products schema:
+
+type Product {
+  id: ID!
+  image: Image!
+}
+
+type Image {
+  id: ID!
+}
+
+# -- Images schema:
+
+type Image {
+  id: ID!
+  url: String!
+}
+```
+
 ### Deploying a stitched schema
 
 Among the simplest and most effective ways to manage a stitched schema is to compose it locally, write the composed SDL as a `.graphql` file in your repo, and then load the composed schema into a stitching client at runtime. For example, setup a `rake` task that loads/fetches subgraph schemas, composes them, and then writes the composed schema definition as a file committed to the repo:

data/docs/request.md

@@ -25,37 +25,12 @@ A `Request` provides the following information:
 - `req.variable_definitions`: a mapping of variable names to their type definitions
 - `req.fragment_definitions`: a mapping of fragment names to their fragment definitions
 
-### Preparing requests
+### Request lifecycle
 
-A request should be prepared for stitching using the `prepare!` method _after_ validations have been run:
+A request manages the flow of stitching behaviors. These are sequenced by the `Client`
+component, or you may invoke them manually:
 
-```ruby
-document = <<~GRAPHQL
-  query FetchMovie($id: ID!, $lang: String = "en", $withShowtimes: Boolean = true) {
-    movie(id:$id) {
-      id
-      title(lang: $lang)
-      showtimes @include(if: $withShowtimes) {
-        time
-      }
-    }
-  }
-GRAPHQL
-
-request = GraphQL::Stitching::Request.new(
-  supergraph,
-  document,
-  variables: { "id" => "1" },
-  operation_name: "FetchMovie",
-)
-
-errors = MySchema.validate(request.document)
-# return early with any static validation errors...
-
-request.prepare!
-```
-
-Preparing a request will apply several destructive transformations:
-
-- Default values from variable definitions will be added to request variables.
-- The document will be pre-shaped based on `@skip` and `@include` directives.
+1. `request.validate`: runs static validations on the request using the combined schema.
+2. `request.prepare!`: inserts variable defaults and pre-renders skip/include conditional shaping.
+3. `request.plan`: builds a plan for the request. May act as a setting for plans pulled from cache.
+4. `request.execute`: executes the request, and returns the resulting data.

data/lib/graphql/stitching/composer.rb

@@ -263,7 +263,6 @@ module GraphQL
         enum_values_by_name_location = types_by_location.each_with_object({}) do |(location, type_candidate), memo|
           type_candidate.enum_values.each do |enum_value_candidate|
             memo[enum_value_candidate.graphql_name] ||= {}
-            memo[enum_value_candidate.graphql_name][location] ||= {}
             memo[enum_value_candidate.graphql_name][location] = enum_value_candidate
           end
         end
@@ -376,7 +375,6 @@ module GraphQL
             @field_map[type_name][field_candidate.name] << location
 
             memo[field_name] ||= {}
-            memo[field_name][location] ||= {}
             memo[field_name][location] = field_candidate
           end
         end
@@ -406,7 +404,6 @@ module GraphQL
         args_by_name_location = members_by_location.each_with_object({}) do |(location, member_candidate), memo|
           member_candidate.arguments.each do |argument_name, argument|
             memo[argument_name] ||= {}
-            memo[argument_name][location] ||= {}
             memo[argument_name][location] = argument
           end
         end
@@ -461,7 +458,6 @@ module GraphQL
         directives_by_name_location = members_by_location.each_with_object({}) do |(location, member_candidate), memo|
           member_candidate.directives.each do |directive|
             memo[directive.graphql_name] ||= {}
-            memo[directive.graphql_name][location] ||= {}
             memo[directive.graphql_name][location] = directive
           end
         end
@@ -652,22 +648,22 @@ module GraphQL
 
         schemas.each do |schema|
           introspection_types = schema.introspection_system.types.keys
-          schema.types.values.each do |type|
+          schema.types.each_value do |type|
             next if introspection_types.include?(type.graphql_name)
 
             if type.kind.object? || type.kind.interface?
-              type.fields.values.each do |field|
+              type.fields.each_value do |field|
                 field_type = field.type.unwrap
                 reads << field_type.graphql_name if field_type.kind.enum?
 
-                field.arguments.values.each do |argument|
+                field.arguments.each_value do |argument|
                   argument_type = argument.type.unwrap
                   writes << argument_type.graphql_name if argument_type.kind.enum?
                 end
               end
 
             elsif type.kind.input_object?
-              type.arguments.values.each do |argument|
+              type.arguments.each_value do |argument|
                 argument_type = argument.type.unwrap
                 writes << argument_type.graphql_name if argument_type.kind.enum?
               end

data/lib/graphql/stitching/executor/root_source.rb

@@ -20,10 +20,22 @@ module GraphQL::Stitching
         result = @executor.request.supergraph.execute_at_location(op.location, query_document, query_variables, @executor.request)
         @executor.query_count += 1
 
-        @executor.data.merge!(result["data"]) if result["data"]
+        if result["data"]
+          if op.path.any?
+            # Nested root scopes must expand their pathed origin set
+            origin_set = op.path.reduce([@executor.data]) do |set, ns|
+              set.flat_map { |obj| obj && obj[ns] }.tap(&:compact!)
+            end
+
+            origin_set.each { _1.merge!(result["data"]) }
+          else
+            # Actual root scopes merge directly into results data
+            @executor.data.merge!(result["data"])
+          end
+        end
+
         if result["errors"]&.any?
-          result["errors"].each { _1.delete("locations") }
-          @executor.errors.concat(result["errors"])
+          @executor.errors.concat(format_errors!(result["errors"], op.path))
         end
 
         ops.map(&:step)
@@ -51,6 +63,16 @@ module GraphQL::Stitching
         doc << op.selections
         doc
       end
+
+      # Format response errors without a document location (because it won't match the request doc),
+      # and prepend any insertion path for the scope into error paths.
+      def format_errors!(errors, path)
+        errors.each do |err|
+          err.delete("locations")
+          err["path"].unshift(*path) if err["path"] && path.any?
+        end
+        errors
+      end
     end
   end
 end

data/lib/graphql/stitching/planner.rb

@@ -276,19 +276,21 @@ module GraphQL
           routes.each_value do |route|
             route.reduce(locale_selections) do |parent_selections, boundary|
               # E.1) Add the key of each boundary query into the prior location's selection set.
-              foreign_key = ExportSelection.key(boundary.key)
-              has_key = false
-              has_typename = false
-
-              parent_selections.each do |node|
-                next unless node.is_a?(GraphQL::Language::Nodes::Field)
-                has_key ||= node.alias == foreign_key
-                has_typename ||= node.alias == ExportSelection.typename_node.alias
+              if boundary.key
+                foreign_key = ExportSelection.key(boundary.key)
+                has_key = false
+                has_typename = false
+
+                parent_selections.each do |node|
+                  next unless node.is_a?(GraphQL::Language::Nodes::Field)
+                  has_key ||= node.alias == foreign_key
+                  has_typename ||= node.alias == ExportSelection.typename_node.alias
+                end
+
+                parent_selections << ExportSelection.key_node(boundary.key) unless has_key
+                parent_selections << ExportSelection.typename_node unless has_typename
               end
 
-              parent_selections << ExportSelection.key_node(boundary.key) unless has_key
-              parent_selections << ExportSelection.typename_node unless has_typename
-
               # E.2) Add a planner step for each new entrypoint location.
               add_step(
                 location: boundary.location,
@@ -296,7 +298,7 @@ module GraphQL
                 parent_type: parent_type,
                 selections: remote_selections_by_location[boundary.location] || [],
                 path: path.dup,
-                boundary: boundary,
+                boundary: boundary.key ? boundary : nil,
               ).selections
             end
           end

data/lib/graphql/stitching/request.rb

@@ -120,12 +120,13 @@ module GraphQL
       end
 
       # Validates the request using the combined supergraph schema.
+      # @return [Array<GraphQL::ExecutionError>] an array of static validation errors
       def validate
         result = @supergraph.static_validator.validate(@query)
         result[:errors]
       end
 
-      # Prepares the request for stitching by rendering variable defaults and applying @skip/@include conditionals.
+      # Prepares the request for stitching by inserting variable defaults and applying @skip/@include conditionals.
       def prepare!
         operation.variables.each do |v|
           @variables[v.name] = v.default_value if @variables[v.name].nil? && !v.default_value.nil?
@@ -143,7 +144,17 @@ module GraphQL
         self
       end
 
-      # Gets and sets the query plan for the request. Assigned query plans may pull from cache.
+      # Gets and sets the query plan for the request. Assigned query plans may pull from a cache,
+      # which is useful for redundant GraphQL documents (commonly sent by frontend clients).
+      # ```ruby
+      # if cached_plan = $cache.get(request.digest)
+      #   plan = GraphQL::Stitching::Plan.from_json(JSON.parse(cached_plan))
+      #   request.plan(plan)
+      # else
+      #   plan = request.plan
+      #   $cache.set(request.digest, JSON.generate(plan.as_json))
+      # end
+      # ```
       # @param new_plan [Plan, nil] a cached query plan for the request.
       # @return [Plan] query plan for the request.
       def plan(new_plan = nil)

data/lib/graphql/stitching/supergraph.rb

@@ -245,7 +245,11 @@ module GraphQL
       # ("Type") => ["id", ...]
       def possible_keys_for_type(type_name)
         @possible_keys_by_type[type_name] ||= begin
-          @boundaries[type_name].map(&:key).tap(&:uniq!)
+          if type_name == @schema.query.graphql_name
+            GraphQL::Stitching::EMPTY_ARRAY
+          else
+            @boundaries[type_name].map(&:key).tap(&:uniq!)
+          end
         end
       end
 
@@ -262,16 +266,25 @@ module GraphQL
       # For a given type, route from one origin location to one or more remote locations
       # used to connect a partial type across locations via boundary queries
       def route_type_to_locations(type_name, start_location, goal_locations)
-        if possible_keys_for_type(type_name).length > 1
+        key_count = possible_keys_for_type(type_name).length
+
+        if key_count.zero?
+          # nested root scopes have no boundary keys and just return a location
+          goal_locations.each_with_object({}) do |goal_location, memo|
+            memo[goal_location] = [Boundary.new(location: goal_location)]
+          end
+
+        elsif key_count > 1
           # multiple keys use an A* search to traverse intermediary locations
-          return route_type_to_locations_via_search(type_name, start_location, goal_locations)
-        end
+          route_type_to_locations_via_search(type_name, start_location, goal_locations)
 
-        # types with a single key attribute must all be within a single hop of each other,
-        # so can use a simple match to collect boundaries for the goal locations.
-        @boundaries[type_name].each_with_object({}) do |boundary, memo|
-          if goal_locations.include?(boundary.location)
-            memo[boundary.location] = [boundary]
+        else
+          # types with a single key attribute must all be within a single hop of each other,
+          # so can use a simple match to collect boundaries for the goal locations.
+          @boundaries[type_name].each_with_object({}) do |boundary, memo|
+            if goal_locations.include?(boundary.location)
+              memo[boundary.location] = [boundary]
+            end
           end
         end
       end

data/lib/graphql/stitching/util.rb

@@ -54,7 +54,7 @@ module GraphQL
           return parent_type.possible_types if parent_type.kind.union?
 
           result = []
-          schema.types.values.each do |type|
+          schema.types.each_value do |type|
             next unless type <= GraphQL::Schema::Interface && type != parent_type
             next unless type.interfaces.include?(parent_type)
             result << type

data/lib/graphql/stitching/version.rb

@@ -2,6 +2,6 @@
 
 module GraphQL
   module Stitching
-    VERSION = "1.2.1"
+    VERSION = "1.2.2"
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: graphql-stitching
 version: !ruby/object:Gem::Version
-  version: 1.2.1
+  version: 1.2.2
 platform: ruby
 authors:
 - Greg MacWilliam
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2024-01-13 00:00:00.000000000 Z
+date: 2024-02-11 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: graphql
@@ -82,13 +82,11 @@ files:
 - docs/README.md
 - docs/client.md
 - docs/composer.md
-- docs/executor.md
 - docs/http_executable.md
 - docs/images/library.png
 - docs/images/merging.png
 - docs/images/stitching.png
 - docs/mechanics.md
-- docs/planner.md
 - docs/request.md
 - docs/supergraph.md
 - examples/file_uploads/Gemfile

data/docs/executor.md

@@ -1,68 +0,0 @@
-## GraphQL::Stitching::Executor
-
-An `Executor` accepts a [`Supergraph`](./supergraph.md), a [query plan hash](./planner.md), and optional request variables. It handles executing requests and merging results collected from across graph locations.
-
-```ruby
-query = <<~GRAPHQL
-  query MyQuery($id: ID!) {
-    product(id:$id) {
-      title
-      brands { name }
-    }
-  }
-GRAPHQL
-
-request = GraphQL::Stitching::Request.new(
-  supergraph,
-  query,
-  variables: { "id" => "123" },
-  operation_name: "MyQuery",
-  context: { ... },
-)
-
-# Via Request:
-result = request.execute
-
-# Via Executor:
-result = GraphQL::Stitching::Executor.new(request).perform
-```
-
-### Raw results
-
-By default, execution results are always returned with document shaping (stitching additions removed, missing fields added, null bubbling applied). You may access the raw execution result by calling the `perform` method with a `raw: true` argument:
-
-```ruby
-# get the raw result without shaping using either form:
-raw_result = request.execute(raw: true)
-raw_result = GraphQL::Stitching::Executor.new(request).perform(raw: true)
-```
-
-The raw result will contain many irregularities from the stitching process, however may be insightful when debugging inconsistencies in results:
-
-```ruby
-{
-  "data" => {
-    "product" => {
-      "upc" => "1",
-      "_export_upc" => "1",
-      "_export_typename" => "Product",
-      "name" => "iPhone",
-      "price" => nil,
-    }
-  }
-}
-```
-
-### Batching
-
-The Executor batches together as many requests as possible to a given location at a given time. Batched queries are written with the operation name suffixed by all operation keys in the batch, and root stitching fields are each prefixed by their batch index and collection index (for non-list fields):
-
-```graphql
-query MyOperation_2_3($lang:String!,$currency:Currency!){
-  _0_result: storefronts(ids:["7","8"]) { name(lang:$lang) }
-  _1_0_result: product(upc:"abc") { price(currency:$currency) }
-  _1_1_result: product(upc:"xyz") { price(currency:$currency) }
-}
-```
-
-All told, the executor will make one request per location per generation of data. Generations started on separate forks of the resolution tree will be resolved independently.

data/docs/planner.md

@@ -1,45 +0,0 @@
-## GraphQL::Stitching::Planner
-
-A `Planner` generates a query plan for a given [`Supergraph`](./supergraph.md) and [`Request`](./request.md). The generated plan breaks down all the discrete GraphQL operations that must be delegated across locations and their sequencing.
-
-```ruby
-document = <<~GRAPHQL
-  query MyQuery($id: ID!) {
-    product(id:$id) {
-      title
-      brands { name }
-    }
-  }
-GRAPHQL
-
-request = GraphQL::Stitching::Request.new(
-  supergraph,
-  document,
-  variables: { "id" => "1" },
-  operation_name: "MyQuery",
-).prepare!
-
-# Via Request:
-plan = request.plan
-
-# Via Planner:
-plan = GraphQL::Stitching::Planner.new(request).perform
-```
-
-### Caching
-
-Plans are designed to be cacheable. This is very useful for redundant GraphQL documents (commonly sent by frontend clients) where there's no sense in planning every request individually. It's far more efficient to generate a plan once and cache it, then simply retreive the plan and execute it for future requests.
-
-```ruby
-cached_plan = $cache.get(request.digest)
-
-if cached_plan
-  plan = GraphQL::Stitching::Plan.from_json(JSON.parse(cached_plan))
-  request.plan(plan)
-else
-  plan = request.plan
-  $cache.set(request.digest, JSON.generate(plan.as_json))
-end
-
-# execute the plan...
-```