feat: scaffold ARM automation framework for Relativity workspace restoration

[davis-chad]

Summary

Scaffolds the Python-based ARM (Archive & Restore Manager) automation system for automating Relativity workspace restoration steps tracked in Smartsheet.

What's included

Core modules (arm_automation/)

Module	Purpose
config.py	Environment-variable-based configuration. Reads SMARTSHEET_API_KEY, REQUESTS_CA_BUNDLE (Zscaler support), and REQUEST_TIMEOUT. Produces an immutable Config dataclass with a .verify property ready for requests.
archive_parser.py	Parses Relativity ARM ArchiveInformation.xml files using xml.etree.ElementTree. Returns a flat dict[str, str] keyed by element path (e.g. ArchiveInformation/DatabaseInfo/ServerName). Also supports CSV export via to_csv(). Handles UNC paths via pathlib.Path.
smartsheet_client.py	Schema-agnostic Smartsheet REST API client built on requests. Supports get_sheet, list_columns, get_row, update_row, and add_rows. Does not hard-code column names/IDs — callers pass them explicitly. Respects the Config.verify property for TLS.
step_runner.py	Decorator-based step registry (@register_step) and execution framework. Steps receive a StepContext dataclass and return a StepResult. Exceptions are caught and surfaced as failed results with messages.

First step implementation (arm_automation/steps/)

Module	Step name	Description
parse_archive_info.py	parse_archive_info	Parses an ArchiveInformation.xml from a path in context.params["archive_xml_path"], optionally exports to CSV, and makes the parsed dict available in StepResult.outputs["archive_data"].

Tests (tests/)

28 tests covering:

• Config.from_env() — valid configs, missing/empty API key, frozen dataclass
• archive_parser.parse() — top-level fields, nested fields, UNC paths, XML attributes, error cases
• archive_parser.to_csv() — file creation, headers, content fidelity, parent directory creation
• Step runner — registration, duplicate detection, unknown steps, success/failure, context passing
• parse_archive_info step — end-to-end with real XML fixture

Other changes

• Updated requirements.txt with requests and pytest
• Updated README.md with full project documentation, usage examples, and design decisions

Design decisions

• REST API over Smartsheet SDK: Keeps the dependency footprint small and gives full control over HTTP behaviour for Zscaler environments.
• stdlib XML parsing: Uses xml.etree.ElementTree — no lxml needed.
• TLS verification never disabled: Config.verify always returns True or a CA bundle path.
• Schema-agnostic Smartsheet client: Column IDs and sheet schemas are never hard-coded.

How to test

pip install -r requirements.txt
pytest tests/ -v