.. _auto_file_unzip:

*FILE UNZIP*
============

Extracts a .zip, .tar.gz, .tgz, or .tar archive attachment and saves every contained file as its own d3VIEW attachment. Returns the new attachment IDs so downstream file_read, file_parser, or file_search workers can operate on the extracted contents directly instead of the archive.

When to use
-----------

Tagged: ``agent``, ``archive``, ``attachments``, ``extract``, ``file``, ``tar``, ``tar.gz``, ``tgz``.

Inputs
------

.. list-table::
   :header-rows: 1
   :widths: 20 20 20 20 20 20

   * - Label
     - ID
     - Type
     - Default
     - Required
     - Description
   * - Archive Attachment
     - input_file
     - file
     - —
     - ✓
     - Attachment ID or name of the archive to extract; must be a .zip, .tar.gz, .tgz, or .tar file — required, no default.
   * - Flatten Nested Paths
     - flatten_paths
     - select
     - yes
     - 
     - Controls how nested archive paths are mapped to attachment names: 'yes' (default) joins path segments with '__' (e.g. bom_dir/part1.csv → bom_dir__part1.csv) to guarantee unique names; set 'no' to use only the raw filename, which may collide when the archive contains same-named files in different sub-directories.

Outputs
-------

.. list-table::
   :header-rows: 1
   :widths: 20 20 20 20

   * - Label
     - ID
     - Type
     - Description
   * - Agent Next Steps
     - agent_next_steps
     - textarea
     - Human-readable markdown block listing every extracted attachment's ID and name; LLM agents should read this first to obtain durable attachment IDs for subsequent tool calls rather than referencing the original archive.
   * - Extracted Attachments
     - attachments
     - dataset
     - Dataset with one row per extracted file containing columns id (integer attachment ID), name (flattened or raw filename string), and size (file size in bytes).
   * - File Count
     - count
     - number
     - Integer count of regular files extracted from the archive; zero if the archive contained no regular files.

Disciplines
-----------

- data.io.archive
- data.io.file

.. raw:: html

   <hr style="margin-top:2em">
   <p style="font-size:11px;color:#888">
   Auto-generated from <code>platform</code> schema. Worker id: <code>file_unzip</code>. Schema hash: <code>dbd9e4ac8e8d</code>. Hand-curated docs in <code>workerexamples/</code> override this page when present.
   </p>