Merge branch 'main' into dw/hashdev

.github/workflows/codeql-analysis.yml

@@ -36,7 +36,7 @@ jobs:
 
       # Initializes the CodeQL tools for scanning.
       - name: Initialize CodeQL
-        uses: github/codeql-action/init@05963f47d870e2cb19a537396c1f668a348c7d8f # v3.24.8
+        uses: github/codeql-action/init@1b1aada464948af03b950897e5eb522f92603cc2 # v3.24.9
         with:
           languages: ${{ matrix.language }}
           # If you wish to specify custom queries, you can do so here or in a config file.
@@ -47,7 +47,7 @@ jobs:
       # Autobuild attempts to build any compiled languages  (C/C++, C#, or Java).
       # If this step fails, then you should remove it and run the build manually (see below)
       - name: Autobuild
-        uses: github/codeql-action/autobuild@05963f47d870e2cb19a537396c1f668a348c7d8f # v3.24.8
+        uses: github/codeql-action/autobuild@1b1aada464948af03b950897e5eb522f92603cc2 # v3.24.9
 
       # ℹ️ Command-line programs to run using the OS shell.
       # 📚 https://git.io/JvXDl
@@ -61,4 +61,4 @@ jobs:
       #   make release
 
       - name: Perform CodeQL Analysis
-        uses: github/codeql-action/analyze@05963f47d870e2cb19a537396c1f668a348c7d8f # v3.24.8
+        uses: github/codeql-action/analyze@1b1aada464948af03b950897e5eb522f92603cc2 # v3.24.9

.github/workflows/dependency-review.yml

@@ -17,4 +17,4 @@ jobs:
       - name: "Checkout Repository"
         uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11 # v4.1.1
       - name: "Dependency Review"
-        uses: actions/dependency-review-action@0fa40c3c10055986a88de3baa0d6ec17c5a894b3 # v4.2.3
+        uses: actions/dependency-review-action@733dd5d4a5203f238c33806593ec0f5fc5343d8c # v4.2.4

.github/workflows/hash-backend-cd.yml

@@ -522,7 +522,7 @@ jobs:
     if: ${{ failure() }}
     steps:
       - name: Slack Notification
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@72c987ec9674843e8a882c6beb4fb043c1ea926e
         env:
           SLACK_LINK_NAMES: true
           SLACK_MESSAGE: "Error deploying the HASH backend <@U0143NL4GMP> <@U027NPY8Y3X> <@U02NLJY0FGX>"

.github/workflows/lint.yml

@@ -188,7 +188,7 @@ jobs:
           mv "$tmp" clippy.sarif
 
       - name: Upload SARIF file
-        uses: github/codeql-action/upload-sarif@05963f47d870e2cb19a537396c1f668a348c7d8f # v3.24.8
+        uses: github/codeql-action/upload-sarif@1b1aada464948af03b950897e5eb522f92603cc2 # v3.24.9
         if: always() && steps.lints.outputs.has-clippy == 'true'
         with:
           sarif_file: ${{ matrix.directory }}/clippy.sarif

.github/workflows/release.yml

@@ -27,7 +27,7 @@ jobs:
           NPM_TOKEN: ${{ secrets.NPM_TOKEN }}
 
       - name: Notify Slack on failure
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@72c987ec9674843e8a882c6beb4fb043c1ea926e
         if: ${{ failure() }}
         env:
           SLACK_LINK_NAMES: true

.github/workflows/semgrep.yml

@@ -43,7 +43,7 @@ jobs:
           files: "semgrep.sarif"
 
       - name: Upload SARIF file for GitHub Advanced Security Dashboard
-        uses: github/codeql-action/upload-sarif@05963f47d870e2cb19a537396c1f668a348c7d8f # v3.24.8
+        uses: github/codeql-action/upload-sarif@1b1aada464948af03b950897e5eb522f92603cc2 # v3.24.9
         if: steps.sarif_file_check.outputs.files_exists == 'true'
         with:
           sarif_file: semgrep.sarif

.github/workflows/tf-apply-hash.yml

@@ -63,7 +63,7 @@ jobs:
           env: ${{ matrix.env }}
 
       - name: Notify Slack on failure
-        uses: rtCamp/action-slack-notify@4e5fb42d249be6a45a298f3c9543b111b02f7907
+        uses: rtCamp/action-slack-notify@72c987ec9674843e8a882c6beb4fb043c1ea926e
         if: ${{ failure() }}
         env:
           SLACK_LINK_NAMES: true

Cargo.toml

@@ -53,7 +53,7 @@ type-system.path = "libs/@blockprotocol/type-system/rust"
 error-stack = { version = "0.4.1", default-features = false }
 
 # Shared third-party dependencies
-bytes = "1.5.0"
+bytes = "1.6.0"
 clap = { version = "4.5.3", default-features = false, features = ["std"] }
 email_address = { version = "0.2.4", default-features = false }
 futures = { version = "0.3.30", default-features = false }

apps/hash-ai-worker-ts/package.json

@@ -4,6 +4,7 @@
   "private": true,
   "description": "A TypeScript 'AI' worker for HASH",
   "license": "AGPL-3.0",
+  "type": "module",
   "exports": {
     ".": "./src/main.ts",
     "./workflows": "./src/workflows.ts",

apps/hash-ai-worker-ts/scripts/bundle-workflow-code.ts

@@ -1,8 +1,16 @@
 import { writeFile } from "node:fs/promises";
+import { createRequire } from "node:module";
 import * as path from "node:path";
+import { dirname } from "node:path";
+import { fileURLToPath } from "node:url";
 
 import { bundleWorkflowCode } from "@temporalio/worker";
 
+const __filename = fileURLToPath(import.meta.url);
+const __dirname = dirname(__filename);
+
+const require = createRequire(import.meta.url);
+
 async function bundle() {
   const { code } = await bundleWorkflowCode({
     workflowsPath: require.resolve("../src/workflows"),

apps/hash-ai-worker-ts/src/activities/infer-entities/log.ts

@@ -1,8 +1,12 @@
 import fs from "node:fs";
-import path from "node:path";
+import path, { dirname } from "node:path";
+import { fileURLToPath } from "node:url";
 
 import { Context } from "@temporalio/activity";
 
+const __filename = fileURLToPath(import.meta.url);
+const __dirname = dirname(__filename);
+
 export const log = (message: string) => {
   const requestId = Context.current().info.workflowExecution.workflowId;
 

apps/hash-ai-worker-ts/src/activities/infer-entities/persist-entities/create-entities.ts

@@ -406,7 +406,17 @@ export const createEntities = async ({
                 proposedEntity,
                 operation: "create",
                 status: "failure",
-                failureReason: `Link entities must have both a sourceEntityId and a targetEntityId.${originalProposal ? `You originally proposed that entityId ${proposedEntity.entityId} should have sourceEntityId ${originalProposal.sourceEntityId?.toString() ?? ""} and targetEntityId ${originalProposal.targetEntityId?.toString() ?? ""}.` : ""}`,
+                failureReason: `Link entities must have both a sourceEntityId and a targetEntityId.${
+                  originalProposal
+                    ? `You originally proposed that entityId ${
+                        proposedEntity.entityId
+                      } should have sourceEntityId ${
+                        originalProposal.sourceEntityId?.toString() ?? ""
+                      } and targetEntityId ${
+                        originalProposal.targetEntityId?.toString() ?? ""
+                      }.`
+                    : ""
+                }`,
               };
             return;
           }

apps/hash-ai-worker-ts/src/activities/shared/embeddings.ts

@@ -34,7 +34,11 @@ const createPropertyEmbeddingInput = (params: {
   propertyTypeSchema: Pick<PropertyType, "title">;
   propertyValue: EntityPropertyValue;
 }): string => {
-  return `${params.propertyTypeSchema.title}: ${typeof params.propertyValue === "string" ? params.propertyValue : JSON.stringify(params.propertyValue)}`;
+  return `${params.propertyTypeSchema.title}: ${
+    typeof params.propertyValue === "string"
+      ? params.propertyValue
+      : JSON.stringify(params.propertyValue)
+  }`;
 };
 
 /**

apps/hash-ai-worker-ts/src/main.ts

@@ -1,5 +1,8 @@
 import * as http from "node:http";
+import { createRequire } from "node:module";
 import * as path from "node:path";
+import { dirname } from "node:path";
+import { fileURLToPath } from "node:url";
 
 import { createGraphClient } from "@local/hash-backend-utils/create-graph-client";
 import { getRequiredEnv } from "@local/hash-backend-utils/environment";
@@ -9,6 +12,11 @@ import { config } from "dotenv-flow";
 
 import { createAiActivities, createGraphActivities } from "./activities";
 
+const __filename = fileURLToPath(import.meta.url);
+const __dirname = dirname(__filename);
+
+const require = createRequire(import.meta.url);
+
 export const monorepoRootDir = path.resolve(__dirname, "../../..");
 
 config({ silent: true, path: monorepoRootDir });

apps/hash-api/package.json

@@ -4,6 +4,7 @@
   "private": true,
   "description": "API and data store for HASH",
   "license": "AGPL-3.0",
+  "type": "module",
   "scripts": {
     "build:docker": "docker buildx build --tag hash-api --file ../../infra/docker/api/prod/Dockerfile ../../ --load",
     "codegen": "rimraf './src/**/*.gen.*'; graphql-codegen --config codegen.config.ts",
@@ -24,7 +25,7 @@
     "@aws-sdk/credential-provider-node": "3.436.0",
     "@aws-sdk/s3-presigned-post": "3.436.0",
     "@aws-sdk/s3-request-presigner": "3.436.0",
-    "@blockprotocol/core": "0.1.2",
+    "@blockprotocol/core": "0.1.3",
     "@blockprotocol/type-system": "0.1.1",
     "@graphql-tools/schema": "8.5.0",
     "@linear/sdk": "6.0.0",
@@ -52,6 +53,7 @@
     "@sentry/node": "7.93.0",
     "@snowplow/node-tracker": "3.3.1",
     "@temporalio/client": "1.8.1",
+    "@temporalio/proto": "1.8.1",
     "@types/ws": "8.5.10",
     "agentkeepalive": "4.2.1",
     "apollo-datasource": "3.3.2",
@@ -96,7 +98,7 @@
     "ws": "8.15.1"
   },
   "devDependencies": {
-    "@blockprotocol/graph": "0.3.4-canary-20240313191438",
+    "@blockprotocol/graph": "0.3.4",
     "@graphql-codegen/cli": "5.0.0",
     "@graphql-codegen/introspection": "2.2.3",
     "@graphql-codegen/typescript": "2.8.7",

apps/hash-api/src/ai/gpt/README.md

@@ -2,6 +2,18 @@
 
 This folder contains functions which support the use of HASH as an external datastore for a custom GPT called HashGPT.
 
+## GPT Instructions
+
+These instructions are set in the ChatGPT interface. They are repeated here to allow them to be version controlled.
+
+Update them if you find that amending the GPT's instructions improves its performance.
+
+> HashGPT is designed to give users a way of exploring the contents of the HASH knowledge graph through natural language queries. It helps users understand the contents of entities in their 'web' (an area of the HASH graph that belongs to them), or webs of the organizations they belong to. It can suggest follow-up questions that the user might like to ask about any data, based on interesting characteristics within it. HashGPT presents its findings in an informative and user-friendly manner, which may include formatted tables, graphs, and other visual aids to enhance understanding – it uses the code interpreter for these visual aids wherever possible, rather than generating images.
+>
+> The GPT ensures clarity and depth in responses. It never invents entities which were not included in the API response. The user values accuracy above all else – the GPT is careful to review the data thoroughly before providing an answer which takes full account of all relevant information. The GPT also biases heavily towards only using the facts in the data it gets from the HASH graph to offer answers, avoiding speculation about things not confirmed by the data unless absolutely necessary or unless the user requests it. Again, the most important thing for users is accuracy – they depend on data in the graph being represented accurately, including which entities link to which other entities. Literal, accurate and comprehensive use of data in the graph is key! Links and entities are never invented or misrepresented by HashGPT.
+>
+> Whenever HashGPT mentions an entity, it includes a link for the user to view more about the entity, as provided by the HASH API as entityHref.
+
 ## Development
 
 ### Fresh local instance setup

apps/hash-api/src/ai/gpt/generate-hashgpt-schema.ts

@@ -1,9 +1,13 @@
 import { writeFileSync } from "node:fs";
-import { resolve } from "node:path";
+import { dirname, resolve } from "node:path";
+import { fileURLToPath } from "node:url";
 
 import { apiOrigin } from "@local/hash-isomorphic-utils/environment";
 import * as generator from "ts-json-schema-generator";
 
+const __filename = fileURLToPath(import.meta.url);
+const __dirname = dirname(__filename);
+
 const config = {
   diagnostics: false,
   noTopRef: true,
@@ -22,6 +26,22 @@ const {
   .createGenerator(config)
   .createSchema("GptQueryEntitiesResponseBody");
 
+const { $ref: queryTypesRequestRef, definitions: queryTypesRequestDefs } =
+  generator
+    .createGenerator({
+      ...config,
+      path: resolve(__dirname, "gpt-query-types.ts"),
+    })
+    .createSchema("GptQueryTypesRequestBody");
+
+const { $ref: queryTypesResponseRef, definitions: queryTypesResponseDefs } =
+  generator
+    .createGenerator({
+      ...config,
+      path: resolve(__dirname, "gpt-query-types.ts"),
+    })
+    .createSchema("GptQueryTypesResponseBody");
+
 const { $ref: getUserWebsResponseRef, definitions: getUserWebsResponseDefs } =
   generator
     .createGenerator({
@@ -34,6 +54,8 @@ const components = {
   schemas: {
     ...queryEntitiesRequestDefs,
     ...queryEntitiesResponseDefs,
+    ...queryTypesRequestDefs,
+    ...queryTypesResponseDefs,
     ...getUserWebsResponseDefs,
   },
 };
@@ -97,6 +119,31 @@ const openApiSchema = {
         },
       },
     },
+    "/gpt/entities/query-types": {
+      post: {
+        description:
+          "Retrieve entity types which match a semantic query, or which belong to a specific web",
+        operationId: "queryTypes",
+        "x-openai-isConsequential:": false,
+        requestBody: {
+          content: {
+            "application/json": {
+              schema: { $ref: queryTypesRequestRef },
+            },
+          },
+        },
+        responses: {
+          "200": {
+            description: "response",
+            content: {
+              "application/json": {
+                schema: { $ref: queryTypesResponseRef },
+              },
+            },
+          },
+        },
+      },
+    },
   },
   components,
 };