docs: update AGENTS.md

AGENTS.md

@@ -2,10 +2,15 @@
 
 This file provides guidance to coding agents when working with code in this repository.
 
+## Prerequisites
+
+- **Node.js**: 22.16.0 (managed via Volta)
+- **Package Manager**: pnpm 9.14.4
+
 ## Essential Commands
 
 **Development:**
-- `pnpm run dev` - Start API server with hot reload on port 5000
+- `pnpm run dev` - Start API server with hot reload on port 3000
 - `pnpm run dev:background` - Start background processor
 - `pnpm run dev:temporal-worker` - Start Temporal worker
 - `pnpm run dev:temporal-server` - Start Temporal server for local development
@@ -35,15 +40,17 @@ This file provides guidance to coding agents when working with code in this repo
 - **Mercurius** - GraphQL server with caching, upload support, and subscriptions
 - **TypeORM** - Database ORM with entity-based modeling and migrations
 - **PostgreSQL** - Primary database with master/slave replication setup. Favor read replica when you're ok with eventually consistent data.
-- **Redis** - Caching and pub/sub via ioRedisPool
+- **Redis** - Caching and pub/sub via `@dailydotdev/ts-ioredis-pool`
 - **Temporal** - Workflow orchestration for background jobs
 - **ClickHouse** - Analytics and metrics storage
 
 **Application Entry Points:**
 - `src/index.ts` - Main Fastify server setup with GraphQL, auth, and middleware
-- `bin/cli.ts` - CLI dispatcher supporting api, background, temporal, cron modes
+- `bin/cli.ts` - CLI dispatcher supporting api, background, temporal, cron, personalized-digest modes
 - `src/background.ts` - Pub/Sub message handlers and background processing
 - `src/cron.ts` - Scheduled task execution
+- `src/temporal/` - Temporal workflow definitions and workers
+- `src/commands/` - Standalone command implementations (e.g., personalized digest)
 
 **GraphQL Schema Organization:**
 - `src/graphql.ts` - Combines all schema modules with transformers and directives
@@ -71,11 +78,12 @@ This file provides guidance to coding agents when working with code in this repo
 
 **Business Domains:**
 - **Content**: Posts, comments, bookmarks, feeds, sources
-- **Users**: Authentication, preferences, profiles, user experience
-- **Organizations**: Squad management, member roles, campaigns
+- **Users**: Authentication, preferences, profiles, user experience, streaks
+- **Squads**: Squad management, member roles, public requests
+- **Organizations**: Organization management, campaigns
 - **Notifications**: Push notifications, email digests, alerts
-- **Monetization**: Paddle subscription management, premium features
-- **Squads**: Squad management, member roles, campaigns
+- **Monetization**: Paddle subscription management, premium features, cores/transactions
+- **Opportunities**: Job matching, recruiter features, candidate profiles
 
 **Testing Strategy:**
 - Jest with supertest for integration testing
@@ -88,4 +96,10 @@ This file provides guidance to coding agents when working with code in this repo
 - GrowthBook for feature flags and A/B testing
 - OneSignal for push notifications
 - Temporal workflows for async job processing
-- Rate limiting and caching at multiple layers
+- Rate limiting and caching at multiple layers
+
+**Infrastructure as Code:**
+- `.infra/` - Pulumi configuration for deployment
+- `.infra/crons.ts` - Cron job schedules and resource limits
+- `.infra/common.ts` - Worker subscription definitions
+- `.infra/index.ts` - Main Pulumi deployment configuration

src/cron/AGENTS.md

@@ -56,7 +56,7 @@ Create a new file in `src/cron/` directory following this pattern:
 import { Cron } from './cron';
 import { YourEntity } from '../entity';
 
-const cron: Cron = {
+export const yourCronName: Cron = {
   name: 'your-cron-name', // Must match name in .infra/crons.ts
   handler: async (con, logger, pubsub) => {
     // Your cron logic here
@@ -75,20 +75,18 @@ const cron: Cron = {
     logger.info({ count: results.length }, 'Cron job completed');
   },
 };
-
-export default cron;
 ```
 
 ### Step 2: Register in Index
 
 Add your cron to `src/cron/index.ts`:
 
 ```typescript
-import yourCron from './yourCron';
+import { yourCronName } from './yourCronName';
 
 export const crons: Cron[] = [
   // ... existing crons
-  yourCron,
+  yourCronName,
 ];
 ```
 
@@ -202,21 +200,32 @@ handler: async (con, logger, pubsub) => {
 
 ### 6. Testing
 
-Create tests in `__tests__/cron/` directory:
+Create tests in `__tests__/cron/` directory. Tests use a real database connection (reset before each test run):
 
 ```typescript
-import cron from '../../src/cron/yourCron';
+import { yourCronName } from '../../src/cron/yourCronName';
+import { expectSuccessfulCron, saveFixtures } from '../helpers';
+import { YourEntity } from '../../src/entity';
+import { DataSource } from 'typeorm';
+import createOrGetConnection from '../../src/db';
 
-describe('yourCron', () => {
-  it('should execute successfully', async () => {
-    const mockCon = createMockConnection();
-    const mockLogger = createMockLogger();
-    const mockPubsub = createMockPubsub();
-
-    await cron.handler(mockCon, mockLogger, mockPubsub);
-
-    // Assertions
-  });
+let con: DataSource;
+
+beforeAll(async () => {
+  con = await createOrGetConnection();
+});
+
+beforeEach(async () => {
+  // Set up test fixtures
+  await saveFixtures(con, YourEntity, yourFixtures);
+});
+
+it('should execute successfully', async () => {
+  await expectSuccessfulCron(yourCronName);
+
+  // Verify database state after cron execution
+  const results = await con.getRepository(YourEntity).find();
+  expect(results).toHaveLength(expectedLength);
 });
 ```
 
@@ -225,7 +234,7 @@ describe('yourCron', () => {
 ### Incremental Processing with Checkpoints
 
 ```typescript
-const cron: Cron = {
+export const incrementalUpdate: Cron = {
   name: 'incremental-update',
   handler: async (con) => {
     const checkpointKey = 'last_incremental_update';
@@ -290,13 +299,13 @@ handler: async (con, logger) => {
 
 ### Kubernetes CronJobs
 
-Crons are deployed as Kubernetes CronJobs via Pulumi:
+Crons are deployed as Kubernetes CronJobs via Pulumi (see `.infra/index.ts` lines 577-591):
 
 - **Command**: `['dumb-init', 'node', 'bin/cli', 'cron', '<cron-name>']`
-- **Spot Instances**: Enabled by default (70% weight) for cost optimization
-- **Default Limits**: Memory and CPU limits can be overridden per cron
-- **Default Deadline**: 300 seconds (5 minutes), configurable per cron
-- **Adhoc Environments**: Crons are disabled in adhoc environments (see `.infra/index.ts`)
+- **Spot Instances**: Enabled by default for all crons
+- **Default Limits**: Uses background worker limits (`512Mi` memory) unless overridden per cron
+- **Default Deadline**: 300 seconds (5 minutes), configurable per cron via `activeDeadlineSeconds`
+- **Adhoc Environments**: Crons are disabled in adhoc environments
 
 ### Resource Configuration
 

src/graphorm/AGENTS.md

@@ -224,7 +224,9 @@ export const resolvers = {
 };
 ```
 
-### Query with Custom Filtering
+### Query with Custom Filtering and Read Replica
+
+The optional fourth parameter enables read replica routing for eventually consistent reads:
 
 ```typescript
 export const resolvers = {
@@ -239,7 +241,7 @@ export const resolvers = {
             .limit(1);
           return builder;
         },
-        true // Use read replica
+        true // for better performance for read queries are required, take into account potential replication lag if doing reads right after writes
       );
     }
   }
@@ -265,6 +267,42 @@ export const resolvers = {
 };
 ```
 
+### Query by Hierarchy
+
+Use `queryByHierarchy` when you need to query a nested field from the resolve tree:
+
+```typescript
+export const resolvers = {
+  Query: {
+    searchPosts: async (_, args, ctx: Context, info) => {
+      return graphorm.queryByHierarchy<GQLPost>(
+        ctx,
+        info,
+        ['posts', 'edges', 'node'], // Path to the nested field
+        (builder) => {
+          builder.queryBuilder.where(`${builder.alias}.visible = true`);
+          return builder;
+        },
+        true // readReplica
+      );
+    }
+  }
+};
+```
+
+### Query Paginated Integration
+
+Use `queryPaginatedIntegration` for non-database data (e.g., external APIs) while still returning Relay-style pagination:
+
+```typescript
+const results = await graphorm.queryPaginatedIntegration<ExternalItem>(
+  (nodeSize) => false, // hasPreviousPage
+  (nodeSize) => nodeSize >= limit, // hasNextPage
+  (node, index) => base64(`cursor:${index}`), // nodeToCursor
+  async () => fetchFromExternalAPI(args), // fetchData callback
+);
+```
+
 ## Configuration Patterns
 
 ### Required Columns
@@ -354,11 +392,12 @@ fields: {
 
 ## Best Practices
 
-### 1. Always Use `beforeQuery` for Filtering
+### 1. Always Use the Query Builder Callback for Filtering
 
-Don't fetch all data and filter in JavaScript. Filter at the database level:
+Don't fetch all data and filter in JavaScript. Filter at the database level using the builder callback parameter:
 
 ```typescript
+// The callback receives { queryBuilder, alias } and must return the modified builder
 (builder) => {
   builder.queryBuilder
     .andWhere(`${builder.alias}."userId" = :userId`, { userId: ctx.userId })
@@ -395,10 +434,13 @@ Always paginate lists to avoid fetching too much data:
 graphorm.queryPaginated(
   ctx,
   info,
-  hasPreviousPage,
-  hasNextPage,
-  nodeToCursor,
-  beforeQuery
+  (nodeSize) => hasPreviousPage(nodeSize),
+  (nodeSize) => hasNextPage(nodeSize),
+  (node, index) => nodeToCursor(node, index),
+  (builder) => {
+    builder.queryBuilder.limit(limit).offset(offset);
+    return builder;
+  }
 );
 ```
 
@@ -450,6 +492,17 @@ requiredColumns: [
 4. **Required Columns**: Only add truly required columns to avoid unnecessary data fetching
 5. **Transform Functions**: Keep transforms lightweight - avoid database calls in transforms
 
+## Available Methods
+
+| Method | Description | Returns |
+|--------|-------------|---------|
+| `query<T>()` | Query multiple results | `Promise<T[]>` |
+| `queryOne<T>()` | Query single result or null | `Promise<T \| null>` |
+| `queryOneOrFail<T>()` | Query single result or throw | `Promise<T>` |
+| `queryPaginated<T>()` | Query with Relay pagination | `Promise<Connection<T>>` |
+| `queryByHierarchy<T>()` | Query nested field from resolve tree | `Promise<T[]>` |
+| `queryPaginatedIntegration<T>()` | Paginate non-DB data | `Promise<Connection<T>>` |
+
 ## Related Files
 
 - **Core Implementation**: `src/graphorm/graphorm.ts`