Installation
============

Requirements: Python 3.10+

There are three ways to install metapod depending on your environment.

From PyPI (released version)
----------------------------

If you have internet access and want the latest stable release:

.. code-block:: bash

   pip install metapod-probes

This gives you file scanners (CSV, JSON, Parquet, Avro), Excel, dbt, SQLite,
and Python transform probes - no database drivers needed.

From source (local development)
-------------------------------

If you cloned the repository or want to modify the code locally:

.. code-block:: bash

   # Clone the repository
   git clone https://github.com/mario-ambrosino/metapod.git
   cd metapod

   # Editable install - code changes take effect immediately
   pip install -e "."

   # With development tools (pytest, ruff)
   pip install -e ".[dev]"

   # With specific extras (e.g. Azure + Soda)
   pip install -e ".[azure,soda]"

   # Verify
   metapod list-probes

The ``-e`` (editable) flag means Python uses your local source directory
directly. You don't need to reinstall after modifying code.

**Running without installing** - if you just want to try it without pip:

.. code-block:: bash

   git clone https://github.com/mario-ambrosino/metapod.git
   cd metapod
   python cli.py list-probes
   python cli.py run "sqlite.*" --config config.yaml

This works because ``cli.py`` adds the project root to ``sys.path``.
However, the ``metapod`` command won't be available in your PATH.

Air-gapped install (no internet)
--------------------------------

For environments where outbound traffic is blocked (corporate networks,
secure zones, Azure private endpoints):

**Step 1** - On a machine WITH internet, build the offline bundle:

.. code-block:: bash

   git clone https://github.com/mario-ambrosino/metapod.git
   cd metapod
   bash scripts/build_airgapped.sh                  # core only
   bash scripts/build_airgapped.sh --extras azure   # with Azure SDKs

This creates ``dist/metapod-probes-airgapped-X.Y.Z.tar.gz`` containing
the wheel + all dependencies + install scripts.

**Step 2** - Transfer the tarball (USB, internal share, whatever works).

**Step 3** - On the air-gapped machine:

Linux / Mac:

.. code-block:: bash

   tar xzf metapod-probes-airgapped-*.tar.gz
   cd metapod-probes-airgapped-*/
   bash install.sh

Windows PowerShell:

.. code-block:: powershell

   Expand-Archive metapod-probes-airgapped-*.tar.gz
   cd metapod-probes-airgapped-*/
   . .\install.ps1

For Presidio PII detection in air-gapped mode, also download the spaCy
model on the connected machine:

.. code-block:: bash

   pip download en_core_web_lg -d ./models/
   # Transfer models/ to air-gapped machine, then:
   pip install --no-index --find-links ./models/ en_core_web_lg

Database extras
---------------

.. code-block:: bash

   pip install 'metapod-probes[oracle]'           # Oracle (oracledb)
   pip install 'metapod-probes[postgres]'         # PostgreSQL (psycopg2)
   pip install 'metapod-probes[snowflake]'        # Snowflake
   pip install 'metapod-probes[bigquery]'         # Google BigQuery
   pip install 'metapod-probes[mysql]'            # MySQL / MariaDB
   pip install 'metapod-probes[db2]'              # IBM DB2
   pip install 'metapod-probes[duckdb]'           # DuckDB
   pip install 'metapod-probes[mongodb]'          # MongoDB
   pip install 'metapod-probes[elasticsearch]'    # Elasticsearch

Cloud extras
------------

.. code-block:: bash

   pip install 'metapod-probes[azure]'            # Azure (ADLS, Blob, ADF, File Share)
   pip install 'metapod-probes[aws]'              # AWS (S3)

Quality, profiling, and compliance extras
-----------------------------------------

.. code-block:: bash

   pip install 'metapod-probes[soda]'             # Soda Core (data quality checks)
   pip install 'metapod-probes[ydata]'            # YData Profiling (statistics, anomalies)
   pip install 'metapod-probes[presidio]'         # Presidio PII (~75 entity types)

Everything
----------

.. code-block:: bash

   pip install 'metapod-probes[all]'

Verify
------

.. code-block:: bash

   metapod list-probes
   # Should show 197 probes across 49 sources