Add RabbitMQ connector and remove MSSQL connector

Replace MSSQL connector with RabbitMQ stream connector supporting
both input and output operations. Includes Rust connector implementation,
Python bindings, integration tests, and an ETL example project.

Author: danghoangnhan

Cargo.lock

@@ -83,6 +83,7 @@ pyo3 = { version = "0.25.0", features = ["abi3-py310", "multiple-pymethods"] }
 pyo3-async-runtimes = "0.25.0"
 pyo3-log = "0.12.4"
 qdrant-client = "1.15.0"
+rabbitmq-stream-client = "0.11.0"
 questdb-rs = "4.0.5"
 rand = "0.9.1"
 rayon = "1.10.0"
@@ -104,7 +105,6 @@ sysinfo = "0.35.1"
 tantivy = "0.22.1"  # Note: don't bump this dependency before the RAG integration test failure is investigated
 tempfile = "3.20.0"
 thiserror = "1.0.63"
-tiberius = { version = "0.12", default-features = false, features = ["tds73", "winauth", "rustls"] }
 timely = { path = "./external/timely-dataflow/timely", features = ["bincode"] }
 tokio = { version = "1.45.1", features = ["rt-multi-thread"] }
 tokio-util = { version = "0.7", features = ["compat"] }

Cargo.toml

@@ -0,0 +1,25 @@
+build:
+	docker compose up -d
+
+stop:
+	docker compose down -v --remove-orphans
+	docker rmi etl-rabbitmq-pathway:latest
+	docker rmi etl-rabbitmq-stream-producer:latest
+
+connect:
+	docker compose exec -it pathway bash
+connect-prod:
+	docker compose exec -it stream-producer bash
+connect-rabbitmq:
+	docker compose exec -it rabbitmq bash
+psql:
+	docker compose exec -it postgres psql -U pathway -d etl_db
+
+logs:
+	docker compose logs pathway
+logs-prod:
+	docker compose logs stream-producer
+logs-rabbitmq:
+	docker compose logs rabbitmq
+logs-postgres:
+	docker compose logs postgres

examples/projects/rabbitmq-ETL/Makefile

@@ -0,0 +1,93 @@
+# ETL with RabbitMQ Streams in / PostgreSQL out
+
+A streaming ETL pipeline built with Pathway that reads from RabbitMQ Streams,
+validates, transforms, aggregates, and joins data, then writes enriched results
+to PostgreSQL in real time.
+
+## What Pathway replaces
+
+This example is inspired by three ETL architectures that traditionally require
+many moving parts:
+
+| Traditional approach | What Pathway replaces |
+|---|---|
+| Python/Java workers consuming from queues | `pw.io.rabbitmq.read` |
+| Staging databases for validation | UDF-based filtering |
+| Airflow DAGs for orchestration | Reactive streaming engine |
+| Celery task queues for processing | Single streaming process |
+| Separate aggregation jobs | `groupby().reduce()` |
+| Multi-step load pipelines | `pw.io.postgres.write` |
+
+**Result:** One Pathway process replaces 5+ services.
+
+## Architecture
+
+```
+Producer (rstream)           Pathway (single process)              PostgreSQL
++----------------+    +------------------------------+    +------------------+
+| employees_raw  |--->| pw.io.rabbitmq.read          |    |                  |
+| orders_raw     |--->|   -> validate (UDFs)         |--->| enriched_employees|
+| listings_raw   |--->|   -> transform               |    | enriched_listings |
++----------------+    |   -> aggregate (groupby)     |    | (snapshot mode)  |
+                      |   -> join (left joins)        |    +------------------+
+                      | pw.io.postgres.write           |
+                      +------------------------------+
+```
+
+## Data sources
+
+- **Employees** (10 records): HR data with intentional quality issues (empty names,
+  invalid emails, negative salaries) to demonstrate validation
+- **Orders** (15+ records): Purchase orders linked to employees by `employee_id`
+- **Listings** (10+ records): Rental property listings linked to employees via
+  `agent_employee_id` (inspired by rental platform crawlers)
+
+## ETL pipeline
+
+1. **Read** three RabbitMQ streams via `pw.io.rabbitmq.read`
+2. **Validate** using UDFs: email format, phone format, positive salary/price, non-empty fields
+3. **Transform**: normalize emails/departments to lowercase, parse string amounts to float,
+   concatenate names, derive boolean flags
+4. **Aggregate**: order stats per employee (`total_orders`, `total_spent`),
+   listing stats per agent (`total_listings`, `avg_listing_price`)
+5. **Join**: left-join employees with order stats and listing stats;
+   left-join listings with agent names
+6. **Write** to PostgreSQL snapshot tables (auto-updated as new data arrives)
+
+## Launching
+
+```bash
+docker compose up -d
+# or
+make
+```
+
+## Checking results
+
+Connect to PostgreSQL:
+
+```bash
+make psql
+```
+
+Then query the enriched tables:
+
+```sql
+SELECT * FROM enriched_employees ORDER BY employee_id;
+SELECT * FROM enriched_listings ORDER BY listing_id;
+```
+
+As the producer continues sending new orders and listings, re-running these queries
+will show updated aggregates in real time.
+
+## Monitoring
+
+- **Pathway logs**: `make logs`
+- **Producer logs**: `make logs-prod`
+- **RabbitMQ Management UI**: http://localhost:15672 (guest/guest)
+
+## Stopping
+
+```bash
+make stop
+```

examples/projects/rabbitmq-ETL/README.md

@@ -0,0 +1,59 @@
+version: "3.7"
+name: etl-rabbitmq
+networks:
+  etl-rabbitmq-network:
+    driver: bridge
+services:
+  rabbitmq:
+    image: rabbitmq:3.13-management
+    environment:
+      RABBITMQ_SERVER_ADDITIONAL_ERL_ARGS: "-rabbitmq_stream advertised_host rabbitmq -rabbitmq_stream advertised_port 5552"
+    ports:
+      - 5552:5552
+      - 5672:5672
+      - 15672:15672
+    command: bash -c "rabbitmq-plugins enable --offline rabbitmq_stream && rabbitmq-server"
+    healthcheck:
+      test: rabbitmq-diagnostics -q ping
+      interval: 10s
+      timeout: 5s
+      retries: 5
+    networks:
+      - etl-rabbitmq-network
+  postgres:
+    image: postgres:16
+    environment:
+      POSTGRES_USER: pathway
+      POSTGRES_PASSWORD: pathway
+      POSTGRES_DB: etl_db
+    ports:
+      - 5432:5432
+    volumes:
+      - ./sql/init-db.sql:/docker-entrypoint-initdb.d/init-db.sql
+    healthcheck:
+      test: ["CMD-SHELL", "pg_isready -U pathway"]
+      interval: 5s
+      timeout: 5s
+      retries: 5
+    networks:
+      - etl-rabbitmq-network
+  stream-producer:
+    build:
+      context: .
+      dockerfile: ./producer-src/Dockerfile
+    depends_on:
+      rabbitmq:
+        condition: service_healthy
+    networks:
+      - etl-rabbitmq-network
+  pathway:
+    build:
+      context: .
+      dockerfile: ./pathway-src/Dockerfile
+    depends_on:
+      rabbitmq:
+        condition: service_healthy
+      postgres:
+        condition: service_healthy
+    networks:
+      - etl-rabbitmq-network

examples/projects/rabbitmq-ETL/docker-compose.yml

@@ -0,0 +1,6 @@
+FROM python:3.10
+
+RUN pip install -U pathway
+COPY ./pathway-src/etl.py etl.py
+
+CMD ["python", "etl.py"]