# Attribution

Every component is under a permissive licence. Preserve each project's
copyright and licence notice if you redistribute.

| Component | Licence | Pinned commit | Source |
|---|---|---|---|
| duckdb/duckdb | MIT | `fc9d6e93409a` | https://github.com/duckdb/duckdb |
| pola-rs/polars | MIT | `9ee0dc5b818a` | https://github.com/pola-rs/polars |
| apache/airflow | Apache-2.0 | `484a87b54bce` | https://github.com/apache/airflow |
| ClickHouse/ClickHouse | Apache-2.0 | `c83ecfd4417a` | https://github.com/ClickHouse/ClickHouse |
| trinodb/trino | Apache-2.0 | `2cf6ed7e29ee` | https://github.com/trinodb/trino |
| apache/superset | Apache-2.0 | `549508b340b2` | https://github.com/apache/superset |
| streamlit/streamlit | Apache-2.0 | `7716a0c0bbbb` | https://github.com/streamlit/streamlit |
| apache/kafka | Apache-2.0 | `432141b93a47` | https://github.com/apache/kafka |
| debezium/debezium | Apache-2.0 | `15bcfe1892b4` | https://github.com/debezium/debezium |
| apache/iceberg | Apache-2.0 | `411c38b58c52` | https://github.com/apache/iceberg |
| dagster-io/dagster | Apache-2.0 | `54577dad9a95` | https://github.com/dagster-io/dagster |
| prefecthq/prefect | Apache-2.0 | `2338265f6588` | https://github.com/prefecthq/prefect |

## What each does

### duckdb/duckdb
- Licence: MIT
- Role: In-process analytical database
- Full SQL analytics on files without a server. Point it at Parquet or CSV and query immediately - the fastest path from raw data to an answer.

### pola-rs/polars
- Licence: MIT
- Role: Fast dataframe library
- Multi-threaded dataframes with a lazy API. Handles datasets that make pandas give up, and uses far less memory.

### apache/airflow
- Licence: Apache-2.0
- Role: Workflow orchestration
- The default scheduler for data pipelines. DAGs as Python, retries, backfills and a UI that shows what actually ran.

### ClickHouse/ClickHouse
- Licence: Apache-2.0
- Role: Columnar OLAP database
- Aggregates billions of rows in milliseconds. The engine you reach for when queries on a row store stop finishing.

### trinodb/trino
- Licence: Apache-2.0
- Role: Distributed SQL query engine
- Queries data where it lives - Postgres, S3, Kafka - without moving it first. One SQL dialect across many sources.

### apache/superset
- Licence: Apache-2.0
- Role: Business intelligence and dashboards
- Connect a database, build charts, share dashboards. Replaces a BI subscription with something you host.

### streamlit/streamlit
- Licence: Apache-2.0
- Role: Data apps in pure Python
- Turn a script into an interactive app without touching frontend code. The quickest way to let someone else explore your data.

### apache/kafka
- Licence: Apache-2.0
- Role: Event streaming platform
- The backbone for real-time data. Durable, ordered logs that multiple consumers can read independently.

### debezium/debezium
- Licence: Apache-2.0
- Role: Change data capture
- Streams row-level changes out of your database as they happen. Keeps a warehouse in sync without batch jobs.

### apache/iceberg
- Licence: Apache-2.0
- Role: Open table format
- Schema evolution, time travel and snapshots on plain object storage. Your data stays readable by anything, not locked in a warehouse.

### dagster-io/dagster
- Licence: Apache-2.0
- Role: Data orchestration with typed assets
- Models pipelines as assets rather than tasks, so you can see what exists, what is stale, and what depends on what.

### prefecthq/prefect
- Licence: Apache-2.0
- Role: Workflow orchestration
- Python-native flows with minimal boilerplate. Easier to adopt than Airflow when your pipelines are still simple.

## Note

Source is fetched at the pinned commits above rather than vendored.
No upstream project is modified.