Add tenant ID concept into web app and backend (#160)

* hacked together a example of using zoekt grpc api

* provide tenant id to zoekt git indexer

* update zoekt version to point to multitenant branch

* pipe tenant id through header to zoekt

* remove incorrect submodule reference and settings typo

* update zoekt commit

* remove unused yarn script

* remove unused grpc client in web server

* remove unneeded deps and improve tenant id log

* pass tenant id when creating repo in db

* add mt yarn script

* add nocheckin comment to tenant id in v2 schema

---------

Co-authored-by: bkellam <bshizzle1234@gmail.com>

package.json

@@ -7,7 +7,9 @@
         "build": "yarn workspaces run build",
         "test": "yarn workspaces run test",
         "dev": "npm-run-all --print-label --parallel dev:zoekt dev:backend dev:web",
-        "dev:zoekt": "export PATH=\"$PWD/bin:$PATH\" && zoekt-webserver -index .sourcebot/index -rpc",
+        "dev:mt": "npm-run-all --print-label --parallel dev:zoekt:mt dev:backend dev:web",
+        "dev:zoekt": "export PATH=\"$PWD/bin:$PATH\" && export SRC_TENANT_ENFORCEMENT_MODE=none && zoekt-webserver -index .sourcebot/index -rpc",
+        "dev:zoekt:mt": "export PATH=\"$PWD/bin:$PATH\" && export SRC_TENANT_ENFORCEMENT_MODE=strict && zoekt-webserver -index .sourcebot/index -rpc",
         "dev:backend": "yarn workspace @sourcebot/backend dev:watch",
         "dev:web": "yarn workspace @sourcebot/web dev"
     },

packages/backend/src/config.ts

@@ -35,6 +35,7 @@ export const syncConfig = async (configPath: string, db: PrismaClient, signal: A
                 const gitHubRepos = await getGitHubReposFromConfig(repoConfig, signal, ctx);
                 const hostUrl = repoConfig.url ?? 'https://github.com';
                 const hostname = repoConfig.url ? new URL(repoConfig.url).hostname : 'github.com';
+                const tenantId = repoConfig.tenantId ?? 0;
 
                 await Promise.all(gitHubRepos.map((repo) => {
                     const repoName = `${hostname}/${repo.full_name}`;
@@ -51,6 +52,7 @@ export const syncConfig = async (configPath: string, db: PrismaClient, signal: A
                         name: repoName,
                         isFork: repo.fork,
                         isArchived: !!repo.archived,
+                        tenantId: tenantId,
                         metadata: {
                             'zoekt.web-url-type': 'github',
                             'zoekt.web-url': repo.html_url,
@@ -101,6 +103,7 @@ export const syncConfig = async (configPath: string, db: PrismaClient, signal: A
                             external_codeHostUrl: hostUrl,
                             cloneUrl: cloneUrl.toString(),
                             name: repoName,
+                            tenantId: 0, // TODO: add support for tenantId in GitLab config
                             isFork,
                             isArchived: project.archived,
                             metadata: {

packages/backend/src/schemas/v2.ts

@@ -72,6 +72,10 @@ export interface GitHubConfig {
    * @minItems 1
    */
   topics?: string[];
+  /**
+   * @nocheckin
+   */
+  tenantId?: number;
   exclude?: {
     /**
      * Exclude forked repositories from syncing.

packages/backend/src/types.ts

@@ -13,6 +13,7 @@ interface BaseRepository {
     codeHost?: string;
     topics?: string[];
     sizeInBytes?: number;
+    tenantId?: number;
 }
 
 /**

packages/backend/src/zoekt.ts

@@ -10,9 +10,12 @@ export const indexGitRepository = async (repo: Repo, ctx: AppContext) => {
     const revisions = [
         'HEAD'
     ];
+    
+    const tenantId = repo.tenantId ?? 0;
+    const shardPrefix = `${tenantId}_${repo.id}`;
 
     const repoPath = getRepoPath(repo, ctx);
-    const command = `zoekt-git-index -allow_missing_branches -index ${ctx.indexPath} -file_limit ${DEFAULT_SETTINGS.maxFileSize} -branches ${revisions.join(',')} -shard_prefix ${repo.id} ${repoPath}`;
+    const command = `zoekt-git-index -allow_missing_branches -index ${ctx.indexPath} -file_limit ${DEFAULT_SETTINGS.maxFileSize} -branches ${revisions.join(',')} -tenant_id ${tenantId} -shard_prefix ${shardPrefix} ${repoPath}`;
 
     return new Promise<{ stdout: string, stderr: string }>((resolve, reject) => {
         exec(command, (error, stdout, stderr) => {

packages/db/prisma/migrations/20250114222109_add_tenant_id_to_repo/migration.sql

@@ -0,0 +1,30 @@
+/*
+  Warnings:
+
+  - Added the required column `tenantId` to the `Repo` table without a default value. This is not possible if the table is not empty.
+
+*/
+-- RedefineTables
+PRAGMA defer_foreign_keys=ON;
+PRAGMA foreign_keys=OFF;
+CREATE TABLE "new_Repo" (
+    "id" INTEGER NOT NULL PRIMARY KEY AUTOINCREMENT,
+    "name" TEXT NOT NULL,
+    "createdAt" DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
+    "updatedAt" DATETIME NOT NULL,
+    "indexedAt" DATETIME,
+    "isFork" BOOLEAN NOT NULL,
+    "isArchived" BOOLEAN NOT NULL,
+    "metadata" JSONB NOT NULL,
+    "cloneUrl" TEXT NOT NULL,
+    "tenantId" INTEGER NOT NULL,
+    "external_id" TEXT NOT NULL,
+    "external_codeHostType" TEXT NOT NULL,
+    "external_codeHostUrl" TEXT NOT NULL
+);
+INSERT INTO "new_Repo" ("cloneUrl", "createdAt", "external_codeHostType", "external_codeHostUrl", "external_id", "id", "indexedAt", "isArchived", "isFork", "metadata", "name", "updatedAt") SELECT "cloneUrl", "createdAt", "external_codeHostType", "external_codeHostUrl", "external_id", "id", "indexedAt", "isArchived", "isFork", "metadata", "name", "updatedAt" FROM "Repo";
+DROP TABLE "Repo";
+ALTER TABLE "new_Repo" RENAME TO "Repo";
+CREATE UNIQUE INDEX "Repo_external_id_external_codeHostUrl_key" ON "Repo"("external_id", "external_codeHostUrl");
+PRAGMA foreign_keys=ON;
+PRAGMA defer_foreign_keys=OFF;

packages/db/prisma/schema.prisma

@@ -20,6 +20,7 @@ model Repo {
   isArchived Boolean
   metadata   Json
   cloneUrl   String
+  tenantId   Int
 
   // The id of the repo in the external service
   external_id           String

packages/web/src/app/api/(server)/search/route.ts

@@ -8,13 +8,21 @@ import { NextRequest } from "next/server";
 
 export const POST = async (request: NextRequest) => {
     const body = await request.json();
-    const parsed = await searchRequestSchema.safeParseAsync(body);
+    const tenantId = await request.headers.get("X-Tenant-ID");
+
+    console.log(`Search request received. Tenant ID: ${tenantId}`);
+
+    const parsed = await searchRequestSchema.safeParseAsync({
+        ...body,
+        ...(tenantId && { tenantId: parseInt(tenantId) }),
+    });
     if (!parsed.success) {
         return serviceErrorResponse(
             schemaValidationError(parsed.error)
         );
     }
 
+
     const response = await search(parsed.data);
     if (isServiceError(response)) {
         return serviceErrorResponse(response);

packages/web/src/lib/schemas.ts

@@ -4,6 +4,7 @@ export const searchRequestSchema = z.object({
     query: z.string(),
     maxMatchDisplayCount: z.number(),
     whole: z.boolean().optional(),
+    tenantId: z.number().optional(),
 });
 
 

packages/web/src/lib/server/searchService.ts

@@ -34,7 +34,7 @@ const aliasPrefixMappings: Record<string, zoektPrefixes> = {
     "revision:": zoektPrefixes.branch,
 }
 
-export const search = async ({ query, maxMatchDisplayCount, whole }: SearchRequest): Promise<SearchResponse | ServiceError> => {
+export const search = async ({ query, maxMatchDisplayCount, whole, tenantId }: SearchRequest): Promise<SearchResponse | ServiceError> => {
     // Replace any alias prefixes with their corresponding zoekt prefixes.
     for (const [prefix, zoektPrefix] of Object.entries(aliasPrefixMappings)) {
         query = query.replaceAll(prefix, zoektPrefix);
@@ -53,9 +53,17 @@ export const search = async ({ query, maxMatchDisplayCount, whole }: SearchReque
         }
     });
 
+    let header: Record<string, string> = {};
+    if (tenantId) {
+        header = {
+            "X-Tenant-ID": tenantId.toString()
+        };
+    }
+
     const searchResponse = await zoektFetch({
         path: "/api/search",
         body,
+        header,
         method: "POST",
     });
 

packages/web/src/lib/server/zoektClient.ts

@@ -1,3 +1,4 @@
+import { headers } from "next/headers";
 import { ZOEKT_WEBSERVER_URL } from "../environment"
 
 
@@ -5,6 +6,7 @@ interface ZoektRequest {
     path: string,
     body: string,
     method: string,
+    header?: Record<string, string>,
     cache?: RequestCache,
 }
 
@@ -12,6 +14,7 @@ export const zoektFetch = async ({
     path,
     body,
     method,
+    header,
     cache,
 }: ZoektRequest) => {
     const response = await fetch(
@@ -19,6 +22,7 @@ export const zoektFetch = async ({
         {
             method,
             headers: {
+                ...header,
                 "Content-Type": "application/json",
             },
             body,

schemas/v2/index.json

@@ -141,6 +141,10 @@
                         ["docs", "core"]
                     ]
                 },
+                "tenantId": {
+                    "type": "number",
+                    "description": "@nocheckin"
+                },
                 "exclude": {
                     "type": "object",
                     "properties": {