graphql-java使用手册:part7 拦截器Instrumentation

原文:http://blog.mygraphql.com/wordpress/?p=112

拦截器Instrumentation

通过实现 graphql.execution.instrumentation.Instrumentation
接口,你可以在执行查询的过程中注入定制代码。并可以修改运行期的行为。

它的主要用途是性能监控和定制日志,但也可以完成其它任务。

当创建 `Graphql 对象时,可以绑定相关的 Instrumentation

GraphQL.newGraphQL(schema)
        .instrumentation(new TracingInstrumentation())
        .build();

定制拦截器(Custom Instrumentation)

要实现 Instrumentation ,需要实现多个 “begin”
开头的方法。这方法会在查询执行过程中,每一步骤开始前被调用。

所有回调方法,都应该返回
graphql.execution.instrumentation.InstrumentationContext
对象,这个对象会在本步骤完成时被回调用,回调用时会告知数据的获取结果,如果出错,可以获取
Throwable 对象。.

下面是一个定制的 Instrumentation 。作用是测量执行时间。

class CustomInstrumentationState implements InstrumentationState {
    private Map<String, Object> anyStateYouLike = new HashMap<>();

    void recordTiming(String key, long time) {
        anyStateYouLike.put(key, time);
    }
}

class CustomInstrumentation implements Instrumentation {
    @Override
    public InstrumentationState createState() {
        //
        // instrumentation state is passed during each invocation of an Instrumentation method
        // and allows you to put stateful data away and reference it during the query execution
        //
        return new CustomInstrumentationState();
    }

    @Override
    public InstrumentationContext<ExecutionResult> beginExecution(InstrumentationExecutionParameters parameters) {
        long startNanos = System.nanoTime();
        return (result, throwable) -> {

            CustomInstrumentationState state = parameters.getInstrumentationState();
            state.recordTiming(parameters.getQuery(), System.nanoTime() - startNanos);
        };
    }

    @Override
    public InstrumentationContext<Document> beginParse(InstrumentationExecutionParameters parameters) {
        //
        // You MUST return a non null object but it does not have to do anything and hence
        // you use this class to return a no-op object
        //
        return new NoOpInstrumentation.NoOpInstrumentationContext<>();
    }

    @Override
    public InstrumentationContext<List<ValidationError>> beginValidation(InstrumentationValidationParameters parameters) {
        return new NoOpInstrumentation.NoOpInstrumentationContext<>();
    }

    @Override
    public InstrumentationContext<ExecutionResult> beginDataFetch(InstrumentationDataFetchParameters parameters) {
        return new NoOpInstrumentation.NoOpInstrumentationContext<>();
    }

    @Override
    public InstrumentationContext<CompletableFuture<ExecutionResult>> beginExecutionStrategy(InstrumentationExecutionStrategyParameters parameters) {
        return new NoOpInstrumentation.NoOpInstrumentationContext<>();
    }

    @Override
    public InstrumentationContext<ExecutionResult> beginField(InstrumentationFieldParameters parameters) {
        return new NoOpInstrumentation.NoOpInstrumentationContext<>();
    }

    @Override
    public InstrumentationContext<Object> beginFieldFetch(InstrumentationFieldFetchParameters parameters) {
        return new NoOpInstrumentation.NoOpInstrumentationContext<>();
    }

    @Override
    public DataFetcher<?> instrumentDataFetcher(DataFetcher<?> dataFetcher, InstrumentationFieldFetchParameters parameters) {
        //
        // this allows you to intercept the data fetcher used ot fetch a field and provide another one, perhaps
        // that enforces certain behaviours or has certain side effects on the data
        //
        return dataFetcher;
    }

    @Override
    public CompletableFuture<ExecutionResult> instrumentExecutionResult(ExecutionResult executionResult, InstrumentationExecutionParameters parameters) {
        //
        // this allows you to instrument the execution result some how.  For example the Tracing support uses this to put
        // the `extensions` map of data in place
        //
        return CompletableFuture.completedFuture(executionResult);
    }
}

链式拦截(Chaining Instrumentation)

你可以用 graphql.execution.instrumentation.ChainedInstrumentation
把多个 Instrumentation 连接起来。这些 Instrumentation
对象会按顺序被调用。

List<Instrumentation> chainedList = new ArrayList<>();
chainedList.add(new FooInstrumentation());
chainedList.add(new BarInstrumentation());
ChainedInstrumentation chainedInstrumentation = new ChainedInstrumentation(chainedList);

GraphQL.newGraphQL(schema)
        .instrumentation(chainedInstrumentation)
        .build();

Apollo跟踪与拦截( Tracing Instrumentation)

graphql.execution.instrumentation.tracing.TracingInstrumentation
是一个可以收集跟踪信息的拦截器。

它按照 Apollo 跟踪格式 https://github.com/apollograp...
来收集跟踪信息。

详细的跟踪信息( tracing map)会放在查询结果的 extensions(扩展)
部分。

如以下的查询:

query {
  hero {
    name
    friends {
      name
    }
  }
}

会返回如下的结果:

{
  "data": {
    "hero": {
      "name": "R2-D2",
      "friends": [
        {
          "name": "Luke Skywalker"
        },
        {
          "name": "Han Solo"
        },
        {
          "name": "Leia Organa"
        }
      ]
    }
  },
  "extensions": {
    "tracing": {
      "version": 1,
      "startTime": "2017-08-14T23:13:39.362Z",
      "endTime": "2017-08-14T23:13:39.497Z",
      "duration": 135589186,
      "execution": {
        "resolvers": [
          {
            "path": [
              "hero"
            ],
            "parentType": "Query",
            "returnType": "Character",
            "fieldName": "hero",
            "startOffset": 105697585,
            "duration": 79111240
          },
          {
            "path": [
              "hero",
              "name"
            ],
            "parentType": "Droid",
            "returnType": "String",
            "fieldName": "name",
            "startOffset": 125010028,
            "duration": 20213
          },
          {
            "path": [
              "hero",
              "friends"
            ],
            "parentType": "Droid",
            "returnType": "[Character]",
            "fieldName": "friends",
            "startOffset": 133352819,
            "duration": 7927560
          },
          {
            "path": [
              "hero",
              "friends",
              0,
              "name"
            ],
            "parentType": "Human",
            "returnType": "String",
            "fieldName": "name",
            "startOffset": 134105887,
            "duration": 6783
          },
          {
            "path": [
              "hero",
              "friends",
              1,
              "name"
            ],
            "parentType": "Human",
            "returnType": "String",
            "fieldName": "name",
            "startOffset": 134725922,
            "duration": 7016
          },
          {
            "path": [
              "hero",
              "friends",
              2,
              "name"
            ],
            "parentType": "Human",
            "returnType": "String",
            "fieldName": "name",
            "startOffset": 134875089,
            "duration": 6342
          }
        ]
      }
    }
  }
}

字段校验拦截器(Field Validation Instrumentation)

graphql.execution.instrumentation.fieldvalidation.FieldValidationInstrumentation
拦截器,可以在执行查询前校验字段和字段参数。如果校验失败,查询将停止,并返回错误信息。

你可以编写自己的FieldValidation 实现,或者直接用
SimpleFieldValidation 去为每个field定义校验逻辑。

ExecutionPath fieldPath = ExecutionPath.parse("/user");
FieldValidation fieldValidation = new SimpleFieldValidation()
        .addRule(fieldPath, new BiFunction<FieldAndArguments, FieldValidationEnvironment, Optional<GraphQLError>>() {
            @Override
            public Optional<GraphQLError> apply(FieldAndArguments fieldAndArguments, FieldValidationEnvironment environment) {
                String nameArg = fieldAndArguments.getFieldArgument("name");
                if (nameArg.length() > 255) {
                    return Optional.of(environment.mkError("Invalid user name", fieldAndArguments));
                }
                return Optional.empty();
            }
        });

FieldValidationInstrumentation instrumentation = new FieldValidationInstrumentation(
        fieldValidation
);

GraphQL.newGraphQL(schema)
        .instrumentation(instrumentation)
        .build();



评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值