Metadata-Version: 2.4
Name: apache-airflow-providers-ailake
Version: 0.0.10
Summary: Apache Airflow provider for AI-Lake Format — hook, operators, and snapshot sensor
Project-URL: Repository, https://github.com/ThiagoLange/iceberg-ai-deltalakehouse
Project-URL: Bug Tracker, https://github.com/ThiagoLange/iceberg-ai-deltalakehouse/issues
Author-email: Thiago Egon Lange <thiago.egon@gmail.com>
License: MIT OR Apache-2.0
Keywords: ailake,airflow,iceberg,lakehouse,provider,vector
Classifier: Framework :: Apache Airflow
Classifier: Framework :: Apache Airflow :: Provider
Classifier: License :: OSI Approved :: Apache Software License
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Classifier: Topic :: Database
Requires-Python: >=3.9
Requires-Dist: apache-airflow>=2.6
Description-Content-Type: text/markdown

# apache-airflow-providers-ailake

Apache Airflow provider for [AI-Lake Format](https://github.com/ThiagoLange/iceberg-ai-deltalakehouse) — an Iceberg-compatible file format that unifies tabular data, embeddings, and HNSW vector indexes in a single Parquet file.

## Installation

```bash
pip install apache-airflow-providers-ailake
```

## Components

### Hook

`AilakeHook` — connects to an AI-Lake table on object storage (S3/GCS/Azure).

```python
from airflow_providers_ailake.hooks.ailake import AilakeHook

hook = AilakeHook(conn_id="ailake_default")
```

### Operators

`AilakeWriteOperator` — writes a batch of rows (with embeddings) to an AI-Lake table.

`AilakeSearchOperator` — runs a vector similarity search and pushes results to XCom.

### Sensor

`AilakeSnapshotSensor` — waits until a new Iceberg snapshot appears on the table (useful for triggering downstream DAGs after a write).

## Requirements

- Apache Airflow >= 2.6
- Python >= 3.9
- AI-Lake SDK (`ailake` Python package) installed in the Airflow worker environment

## Links

- [Source](https://github.com/ThiagoLange/iceberg-ai-deltalakehouse)
- [Issue tracker](https://github.com/ThiagoLange/iceberg-ai-deltalakehouse/issues)
