| 1 | import * as EMRContainers from "@/AWS/EMRContainers"; |
| 2 | import { Role } from "@/AWS/IAM"; |
| 3 | import * as Lambda from "@/AWS/Lambda"; |
| 4 | import * as Duration from "effect/Duration"; |
| 5 | import * as Effect from "effect/Effect"; |
| 6 | import * as Layer from "effect/Layer"; |
| 7 | import * as Schedule from "effect/Schedule"; |
| 8 | import * as Stream from "effect/Stream"; |
| 9 | import { HttpServerRequest } from "effect/http/HttpServerRequest"; |
| 10 | import * as HttpServerResponse from "effect/http/HttpServerResponse"; |
| 11 | import path from "pathe"; |
| 12 | |
| 13 | const main = path.resolve(import.meta.dirname, "handler.ts"); |
| 14 | |
| 15 | export class EmrcTestFunction extends Lambda.Function<Lambda.Function>()( |
| 16 | "EmrcTestFunction", |
| 17 | ) {} |
| 18 | |
| 19 | export default EmrcTestFunction.make( |
| 20 | { |
| 21 | main, |
| 22 | functionUrl: true, |
| 23 | timeout: Duration.seconds(30), |
| 24 | }, |
| 25 | Effect.gen(function* () { |
| 26 | // The job template the template-scoped binding is bound to. Templates |
| 27 | // are free, account-level, and never run anything by themselves. |
| 28 | const jobRole = yield* Role("JobRole", { |
| 29 | assumeRolePolicyDocument: { |
| 30 | Version: "2012-10-17", |
| 31 | Statement: [ |
| 32 | { |
| 33 | Effect: "Allow", |
| 34 | Principal: { Service: "pods.eks.amazonaws.com" }, |
| 35 | Action: ["sts:AssumeRole", "sts:TagSession"], |
| 36 | }, |
| 37 | ], |
| 38 | }, |
| 39 | }); |
| 40 | const template = yield* EMRContainers.JobTemplate("BindingTemplate", { |
| 41 | jobTemplateData: { |
| 42 | executionRoleArn: jobRole.roleArn, |
| 43 | releaseLabel: "emr-7.5.0-latest", |
| 44 | jobDriver: { |
| 45 | sparkSubmitJobDriver: { |
| 46 | entryPoint: "s3://alchemy-test-emrc/scripts/etl.py", |
| 47 | }, |
| 48 | }, |
| 49 | }, |
| 50 | }); |
| 51 | |
| 52 | // Event source: subscribe the host to EMR on EKS job run state changes. |
| 53 | // The deploy proves the EventBridge rule + invoke permission wiring. |
| 54 | yield* EMRContainers.consumeJobRunEvents( |
| 55 | { states: ["FAILED", "COMPLETED"] }, |
| 56 | (events) => |
| 57 | Stream.runForEach(events, (event) => |
| 58 | Effect.log(`job run ${event.detail.id} -> ${event.detail.state}`), |
| 59 | ), |
| 60 | ); |
| 61 | |
| 62 | const describeJobTemplate = |
| 63 | yield* EMRContainers.DescribeJobTemplate(template); |
| 64 | const listJobTemplates = yield* EMRContainers.ListJobTemplates(); |
| 65 | const listVirtualClusters = yield* EMRContainers.ListVirtualClusters(); |
| 66 | const templateId = yield* template.jobTemplateId; |
| 67 | |
| 68 | const bound = { |
| 69 | describeJobTemplate, |
| 70 | listJobTemplates, |
| 71 | listVirtualClusters, |
| 72 | }; |
| 73 | |
| 74 | // Drain a bounded number of pages on every observation. AWS list APIs may |
| 75 | // return a short (or empty) page with a continuation token while the |
| 76 | // account is being mutated by other test files, so repeatedly reading only |
| 77 | // page one makes the full-concurrency sweep flaky. |
| 78 | const listAllJobTemplates = Effect.fn(function* () { |
| 79 | const templates: Array<{ readonly id?: string }> = []; |
| 80 | let nextToken: string | undefined; |
| 81 | for (let page = 0; page < 10; page++) { |
| 82 | const response = yield* listJobTemplates({ |
| 83 | maxResults: 100, |
| 84 | nextToken, |
| 85 | }); |
| 86 | templates.push(...(response.templates ?? [])); |
| 87 | nextToken = response.nextToken; |
| 88 | if (nextToken === undefined) break; |
| 89 | } |
| 90 | return templates; |
| 91 | }); |
| 92 | |
| 93 | return { |
| 94 | fetch: Effect.gen(function* () { |
| 95 | const request = yield* HttpServerRequest; |
| 96 | const url = new URL(request.originalUrl); |
| 97 | const pathname = url.pathname; |
| 98 | |
| 99 | if (request.method === "GET" && pathname === "/bindings") { |
| 100 | return yield* HttpServerResponse.json({ |
| 101 | bound: Object.keys(bound), |
| 102 | }); |
| 103 | } |
| 104 | |
| 105 | // Template-scoped read: the template id is injected by the binding. |
| 106 | if (request.method === "GET" && pathname === "/template") { |
| 107 | const { jobTemplate } = yield* describeJobTemplate(); |
| 108 | return yield* HttpServerResponse.json({ |
| 109 | id: jobTemplate?.id, |
| 110 | name: jobTemplate?.name, |
| 111 | releaseLabel: jobTemplate?.jobTemplateData.releaseLabel, |
| 112 | }); |
| 113 | } |
| 114 | |
| 115 | // Account-level list. |
| 116 | if (request.method === "GET" && pathname === "/templates") { |
| 117 | const expectedTemplateId = yield* templateId; |
| 118 | const templates = yield* Effect.repeat(listAllJobTemplates(), { |
| 119 | schedule: Schedule.fixed("2 seconds"), |
| 120 | until: (items) => |
| 121 | items.some((item) => item.id === expectedTemplateId), |
| 122 | times: 10, |
| 123 | }); |
| 124 | return yield* HttpServerResponse.json({ |
| 125 | ids: templates.map((template) => template.id), |
| 126 | }); |
| 127 | } |
| 128 | |
| 129 | // Account-level virtual cluster list (proves IAM grant + call path |
| 130 | // without needing an EKS-backed virtual cluster to exist). |
| 131 | if (request.method === "GET" && pathname === "/virtual-clusters") { |
| 132 | const { virtualClusters } = yield* listVirtualClusters({ |
| 133 | states: ["RUNNING"], |
| 134 | }); |
| 135 | return yield* HttpServerResponse.json({ |
| 136 | ids: (virtualClusters ?? []).map((vc) => vc.id), |
| 137 | }); |
| 138 | } |
| 139 | |
| 140 | return yield* HttpServerResponse.json( |
| 141 | { error: "Not found", method: request.method, pathname }, |
| 142 | { status: 404 }, |
| 143 | ); |
| 144 | }).pipe(Effect.orDie), |
| 145 | }; |
| 146 | }).pipe( |
| 147 | Effect.provide( |
| 148 | Layer.mergeAll( |
| 149 | Lambda.EventSource, |
| 150 | EMRContainers.DescribeJobTemplateHttp, |
| 151 | EMRContainers.ListJobTemplatesHttp, |
| 152 | EMRContainers.ListVirtualClustersHttp, |
| 153 | ), |
| 154 | ), |
| 155 | ), |
| 156 | ); |