opencodekit 0.15.4 → 0.15.6

package/dist/template/.opencode/skill/cloudflare/references/pipelines/README.md

@@ -0,0 +1,664 @@
+# Cloudflare Pipelines Skill
+
+Expert guidance for working with Cloudflare Pipelines - ETL streaming data platform for ingesting, transforming, and loading data into R2.
+
+## Overview
+
+Cloudflare Pipelines ingests events, transforms them with SQL, and delivers to R2 as Apache Iceberg tables or Parquet/JSON files. It provides:
+
+- **Streams**: Durable, buffered queues for event ingestion via HTTP or Workers
+- **Pipelines**: SQL-based transformations between streams and sinks
+- **Sinks**: Destinations for processed data (R2 Data Catalog or R2 storage)
+
+**Status**: Open beta (Workers Paid plan required)
+**Pricing**: Currently no charge beyond standard R2 storage/operations
+
+## Architecture
+
+```
+Data Sources → Streams → Pipelines (SQL) → Sinks → R2
+                 ↑          ↓                 ↓
+            HTTP/Workers   Transform      Iceberg/Parquet
+```
+
+### Core Components
+
+1. **Streams**: Buffer and store incoming events
+   - Structured (with schema validation) or unstructured (raw JSON)
+   - HTTP endpoints and Worker bindings
+   - Can be read by multiple pipelines
+
+2. **Pipelines**: Execute SQL transformations
+   - Filter, transform, enrich data
+   - Cannot be modified after creation (delete/recreate required)
+   - SQL reference: SELECT, INSERT, scalar functions
+
+3. **Sinks**: Write to destinations
+   - **R2 Data Catalog**: Apache Iceberg tables with ACID guarantees
+   - **R2 Storage**: Parquet or JSON files
+   - Exactly-once delivery guarantee
+
+## Common Use Cases
+
+- **Analytics pipelines**: Server logs, clickstream, telemetry
+- **Data warehousing**: ETL into queryable Iceberg tables
+- **Event processing**: Mobile/IoT events with enrichment
+- **Ecommerce analytics**: User events, purchases, product views
+- **Log aggregation**: Application/server logs with filtering
+
+## Setup & Configuration
+
+### Quick Start
+
+```bash
+# Interactive setup (recommended)
+npx wrangler pipelines setup
+
+# Manual setup
+npx wrangler r2 bucket create my-bucket
+npx wrangler r2 bucket catalog enable my-bucket
+npx wrangler pipelines streams create my-stream --schema-file schema.json
+npx wrangler pipelines sinks create my-sink --type r2-data-catalog \
+  --bucket my-bucket --namespace default --table my_table \
+  --catalog-token YOUR_TOKEN
+npx wrangler pipelines create my-pipeline \
+  --sql "INSERT INTO my_sink SELECT * FROM my_stream"
+```
+
+### Schema Definition
+
+**Structured streams** (recommended for validation):
+
+```json
+{
+  "fields": [
+    {
+      "name": "user_id",
+      "type": "string",
+      "required": true
+    },
+    {
+      "name": "event_type",
+      "type": "string",
+      "required": true
+    },
+    {
+      "name": "amount",
+      "type": "float64",
+      "required": false
+    },
+    {
+      "name": "tags",
+      "type": "list",
+      "required": false,
+      "items": {
+        "type": "string"
+      }
+    },
+    {
+      "name": "metadata",
+      "type": "struct",
+      "required": false,
+      "fields": [
+        {
+          "name": "source",
+          "type": "string",
+          "required": false
+        }
+      ]
+    }
+  ]
+}
+```
+
+**Supported types**: `string`, `int32`, `int64`, `float32`, `float64`, `bool`, `timestamp`, `json`, `binary`, `list`, `struct`
+
+**Unstructured streams** (no validation, single `value` column):
+```bash
+npx wrangler pipelines streams create my-stream
+```
+
+## Writing Data to Streams
+
+### Via Workers (Recommended)
+
+**Configuration** (`wrangler.toml`):
+```toml
+[[pipelines]]
+pipeline = "<STREAM_ID>"
+binding = "STREAM"
+```
+
+**Or JSON** (`wrangler.jsonc`):
+```jsonc
+{
+  "$schema": "./node_modules/wrangler/config-schema.json",
+  "pipelines": [
+    {
+      "pipeline": "<STREAM_ID>",
+      "binding": "STREAM"
+    }
+  ]
+}
+```
+
+**Worker code**:
+```typescript
+export default {
+  async fetch(request, env, ctx): Promise<Response> {
+    const event = {
+      user_id: "12345",
+      event_type: "purchase",
+      product_id: "widget-001",
+      amount: 29.99
+    };
+    
+    // Send single or multiple events
+    await env.STREAM.send([event]);
+    
+    return new Response('Event sent');
+  },
+} satisfies ExportedHandler<Env>;
+```
+
+**Batch sending**:
+```typescript
+const events = [
+  { user_id: "user1", event_type: "view" },
+  { user_id: "user2", event_type: "purchase", amount: 50 }
+];
+await env.STREAM.send(events);
+```
+
+### Via HTTP
+
+**Endpoint format**: `https://{stream-id}.ingest.cloudflare.com`
+
+**Without auth** (for testing):
+```bash
+curl -X POST https://{stream-id}.ingest.cloudflare.com \
+  -H "Content-Type: application/json" \
+  -d '[
+    {
+      "user_id": "user_12345",
+      "event_type": "purchase",
+      "product_id": "widget-001",
+      "amount": 29.99
+    }
+  ]'
+```
+
+**With authentication**:
+```bash
+curl -X POST https://{stream-id}.ingest.cloudflare.com \
+  -H "Content-Type: application/json" \
+  -H "Authorization: Bearer YOUR_API_TOKEN" \
+  -d '[{"event": "data"}]'
+```
+
+**Required permission**: Workers Pipeline Send
+
+## SQL Transformations
+
+### Basic Patterns
+
+**Pass-through**:
+```sql
+INSERT INTO my_sink SELECT * FROM my_stream
+```
+
+**Filtering**:
+```sql
+INSERT INTO my_sink
+SELECT * FROM my_stream
+WHERE event_type = 'purchase' AND amount > 100
+```
+
+**Field selection**:
+```sql
+INSERT INTO my_sink
+SELECT user_id, event_type, timestamp, amount
+FROM my_stream
+```
+
+**Field transformation**:
+```sql
+INSERT INTO my_sink
+SELECT
+  user_id,
+  UPPER(event_type) as event_type,
+  timestamp,
+  amount * 1.1 as amount_with_tax,
+  CONCAT(user_id, '_', product_id) as unique_key
+FROM my_stream
+```
+
+**Conditional logic**:
+```sql
+INSERT INTO my_sink
+SELECT
+  user_id,
+  event_type,
+  CASE
+    WHEN amount > 1000 THEN 'high_value'
+    WHEN amount > 100 THEN 'medium_value'
+    ELSE 'low_value'
+  END as customer_tier
+FROM my_stream
+WHERE event_type IN ('purchase', 'refund')
+```
+
+## Sink Configuration
+
+### R2 Data Catalog (Iceberg Tables)
+
+**Create sink**:
+```bash
+npx wrangler pipelines sinks create my-sink \
+  --type r2-data-catalog \
+  --bucket my-bucket \
+  --namespace my_namespace \
+  --table my_table \
+  --catalog-token YOUR_CATALOG_TOKEN \
+  --compression zstd \
+  --roll-interval 60 \
+  --roll-size 100
+```
+
+**Options**:
+- `--compression`: `zstd` (default), `snappy`, `gzip`, `lz4`, `uncompressed`
+- `--roll-interval`: Seconds between writes (default: 300)
+- `--roll-size`: Max file size in MB before rolling
+- `--target-row-group-size`: Parquet row group size in MB (default: 256)
+
+**Querying with R2 SQL**:
+```bash
+export WRANGLER_R2_SQL_AUTH_TOKEN=YOUR_API_TOKEN
+
+npx wrangler r2 sql query "warehouse_name" "
+SELECT user_id, event_type, COUNT(*) as event_count
+FROM default.my_table
+WHERE event_type = 'purchase'
+GROUP BY user_id, event_type
+LIMIT 100"
+```
+
+### R2 Storage (Raw Files)
+
+**JSON format**:
+```bash
+npx wrangler pipelines sinks create my-sink \
+  --type r2 \
+  --bucket my-bucket \
+  --format json \
+  --path analytics/events \
+  --partitioning "year=%Y/month=%m/day=%d" \
+  --roll-interval 60 \
+  --roll-size 100 \
+  --access-key-id YOUR_KEY \
+  --secret-access-key YOUR_SECRET
+```
+
+**Parquet format** (better compression/performance):
+```bash
+npx wrangler pipelines sinks create my-sink \
+  --type r2 \
+  --bucket my-bucket \
+  --format parquet \
+  --compression zstd \
+  --path analytics/events \
+  --partitioning "year=%Y/month=%m/day=%d/hour=%H" \
+  --target-row-group-size 256 \
+  --roll-interval 300 \
+  --roll-size 100 \
+  --access-key-id YOUR_KEY \
+  --secret-access-key YOUR_SECRET
+```
+
+**File organization**:
+```
+bucket/analytics/events/
+  year=2025/
+    month=01/
+      day=11/
+        uuid-1.parquet
+        uuid-2.parquet
+```
+
+## Wrangler Commands Reference
+
+### Setup & Management
+
+```bash
+# Interactive setup (creates stream, sink, pipeline)
+npx wrangler pipelines setup
+
+# List all pipelines
+npx wrangler pipelines list
+
+# Get pipeline details
+npx wrangler pipelines get <PIPELINE_ID>
+
+# Delete pipeline
+npx wrangler pipelines delete <PIPELINE_ID>
+```
+
+### Streams
+
+```bash
+# Create stream with schema
+npx wrangler pipelines streams create my-stream --schema-file schema.json
+
+# Create unstructured stream
+npx wrangler pipelines streams create my-stream
+
+# List streams
+npx wrangler pipelines streams list
+
+# Get stream details
+npx wrangler pipelines streams get <STREAM_ID>
+
+# Delete stream (deletes dependent pipelines and buffered events!)
+npx wrangler pipelines streams delete <STREAM_ID>
+```
+
+### Sinks
+
+```bash
+# Create R2 Data Catalog sink
+npx wrangler pipelines sinks create my-sink \
+  --type r2-data-catalog \
+  --bucket my-bucket \
+  --namespace default \
+  --table my_table \
+  --catalog-token TOKEN
+
+# Create R2 storage sink
+npx wrangler pipelines sinks create my-sink \
+  --type r2 \
+  --bucket my-bucket \
+  --format parquet \
+  --compression zstd
+
+# List sinks
+npx wrangler pipelines sinks list
+
+# Get sink details
+npx wrangler pipelines sinks get <SINK_ID>
+
+# Delete sink
+npx wrangler pipelines sinks delete <SINK_ID>
+```
+
+### Pipelines
+
+```bash
+# Create with inline SQL
+npx wrangler pipelines create my-pipeline \
+  --sql "INSERT INTO my_sink SELECT * FROM my_stream"
+
+# Create with SQL file
+npx wrangler pipelines create my-pipeline \
+  --sql-file transform.sql
+
+# View pipeline
+npx wrangler pipelines get <PIPELINE_ID>
+
+# List all pipelines
+npx wrangler pipelines list
+
+# Delete pipeline
+npx wrangler pipelines delete <PIPELINE_ID>
+```
+
+## Authentication & Permissions
+
+### R2 Data Catalog Token
+
+Required permissions: **R2 Admin Read & Write**
+
+Create in dashboard:
+1. Go to R2 > Manage API tokens
+2. Create Account API Token
+3. Select "Admin Read & Write" permission
+4. Save token value
+
+### R2 Storage Credentials
+
+Required permissions: **Object Read & Write**
+
+Create via Wrangler or dashboard for access key ID and secret access key.
+
+### HTTP Ingest Token
+
+Required permissions: **Workers Pipeline Send**
+
+For authenticated HTTP ingestion endpoints.
+
+## Best Practices
+
+### Schema Design
+- ✅ Use structured streams for validation
+- ✅ Mark critical fields as `required: true`
+- ✅ Use appropriate types (`int64` for timestamps, `float64` for decimals)
+- ❌ Avoid overly nested structs (query performance)
+- ❌ Don't change schemas after creation (recreate stream)
+
+### Performance
+- **Low latency**: Set `--roll-interval 10` (smaller files, more frequent)
+- **Query performance**: Set `--roll-interval 300` and `--roll-size 100` (larger files, less frequent)
+- Use `zstd` compression for best ratio, `snappy` for speed
+- Increase `--target-row-group-size` for analytical workloads
+
+### SQL Transformations
+- ✅ Filter early (`WHERE` clauses reduce data volume)
+- ✅ Select only needed fields (reduces storage costs)
+- ✅ Use functions for enrichment (CONCAT, UPPER, CASE)
+- ❌ Cannot modify pipelines after creation (plan carefully)
+- ❌ No JOINs across streams (single stream per pipeline)
+
+### Workers Integration
+- ✅ Use Worker bindings (no token management)
+- ✅ Batch events when possible (`send([event1, event2, ...])`)
+- ✅ Handle send errors gracefully
+- ❌ Don't await send in critical path if latency matters (use `ctx.waitUntil()`)
+
+```typescript
+// Fire-and-forget pattern
+export default {
+  async fetch(request, env, ctx) {
+    const event = { /* ... */ };
+    
+    // Don't block response on send
+    ctx.waitUntil(env.STREAM.send([event]));
+    
+    return new Response('OK');
+  }
+};
+```
+
+### HTTP Ingestion
+- ✅ Enable auth for production endpoints
+- ✅ Configure CORS if sending from browsers
+- ✅ Send arrays (not single objects) for batch efficiency
+- ✅ Handle 4xx/5xx responses with retries
+
+### Monitoring
+- Check stream buffer status (dashboard or API)
+- Monitor pipeline processing rate
+- Review R2 storage growth
+- Query data regularly to verify pipeline health
+
+## Limits (Open Beta)
+
+| Resource | Limit |
+|----------|-------|
+| Streams per account | 20 |
+| Sinks per account | 20 |
+| Pipelines per account | 20 |
+| Payload size per request | 1 MB |
+| Ingest rate per stream | 5 MB/s |
+
+Request increases: [Limit Increase Form](https://forms.gle/ukpeZVLWLnKeixDu7)
+
+## Troubleshooting
+
+### Events not appearing in R2
+- Wait 10-300 seconds (depends on `--roll-interval`)
+- Check pipeline status: `npx wrangler pipelines get <ID>`
+- Verify stream has data (check dashboard metrics)
+- Confirm sink credentials are valid
+
+### Schema validation failures
+- Events accepted but dropped if invalid
+- Check event structure matches schema exactly
+- Verify required fields are present
+- Check data types (e.g., strings not numbers)
+
+### Worker binding not found
+- Verify `wrangler.toml`/`wrangler.jsonc` has correct `pipeline` ID
+- Redeploy Worker after adding binding
+- Check binding name matches code (`env.STREAM`)
+
+### SQL errors
+- SQL cannot be modified after creation
+- Recreate pipeline with corrected SQL
+- Verify stream and sink names in SQL match actual resources
+- Check SQL syntax against reference docs
+
+## Complete Example: Ecommerce Analytics
+
+**1. Create schema** (`ecommerce-schema.json`):
+```json
+{
+  "fields": [
+    {
+      "name": "user_id",
+      "type": "string",
+      "required": true
+    },
+    {
+      "name": "event_type",
+      "type": "string",
+      "required": true
+    },
+    {
+      "name": "product_id",
+      "type": "string",
+      "required": false
+    },
+    {
+      "name": "amount",
+      "type": "float64",
+      "required": false
+    },
+    {
+      "name": "timestamp",
+      "type": "timestamp",
+      "required": true
+    }
+  ]
+}
+```
+
+**2. Setup infrastructure**:
+```bash
+# Create bucket and enable catalog
+npx wrangler r2 bucket create ecommerce-data
+npx wrangler r2 bucket catalog enable ecommerce-data
+
+# Create stream
+npx wrangler pipelines streams create ecommerce-stream \
+  --schema-file ecommerce-schema.json
+
+# Create sink
+npx wrangler pipelines sinks create ecommerce-sink \
+  --type r2-data-catalog \
+  --bucket ecommerce-data \
+  --namespace default \
+  --table events \
+  --catalog-token $CATALOG_TOKEN \
+  --roll-interval 60
+
+# Create pipeline with transformation
+npx wrangler pipelines create ecommerce-pipeline \
+  --sql "INSERT INTO ecommerce_sink 
+         SELECT 
+           user_id,
+           UPPER(event_type) as event_type,
+           product_id,
+           amount,
+           timestamp,
+           CASE 
+             WHEN amount > 100 THEN 'high_value'
+             ELSE 'standard'
+           END as transaction_tier
+         FROM ecommerce_stream
+         WHERE event_type IN ('purchase', 'add_to_cart', 'view_product')"
+```
+
+**3. Configure Worker** (`wrangler.toml`):
+```toml
+name = "ecommerce-api"
+main = "src/index.ts"
+
+[[pipelines]]
+pipeline = "<STREAM_ID>"
+binding = "EVENTS"
+```
+
+**4. Send events** (`src/index.ts`):
+```typescript
+interface Env {
+  EVENTS: Pipeline;
+}
+
+export default {
+  async fetch(request: Request, env: Env): Promise<Response> {
+    if (request.method === 'POST') {
+      const data = await request.json();
+      
+      const event = {
+        user_id: data.userId,
+        event_type: data.eventType,
+        product_id: data.productId,
+        amount: data.amount,
+        timestamp: new Date().toISOString()
+      };
+      
+      try {
+        await env.EVENTS.send([event]);
+        return new Response('Event tracked', { status: 200 });
+      } catch (error) {
+        return new Response('Failed to track event', { status: 500 });
+      }
+    }
+    
+    return new Response('Method not allowed', { status: 405 });
+  }
+} satisfies ExportedHandler<Env>;
+```
+
+**5. Query results**:
+```bash
+export WRANGLER_R2_SQL_AUTH_TOKEN=$CATALOG_TOKEN
+
+npx wrangler r2 sql query "ecommerce-warehouse" "
+SELECT 
+  event_type,
+  transaction_tier,
+  COUNT(*) as event_count,
+  SUM(amount) as total_revenue
+FROM default.events
+WHERE event_type = 'PURCHASE'
+GROUP BY event_type, transaction_tier
+ORDER BY total_revenue DESC"
+```
+
+## Additional Resources
+
+- [Pipelines Documentation](https://developers.cloudflare.com/pipelines/)
+- [SQL Reference](https://developers.cloudflare.com/pipelines/sql-reference/)
+- [R2 Data Catalog](https://developers.cloudflare.com/r2/data-catalog/)
+- [Wrangler Commands](https://developers.cloudflare.com/workers/wrangler/commands/#pipelines)
+- [Apache Iceberg](https://iceberg.apache.org/)