Build a standard NextGenSwitch Python SDK

[masum0009]

Summary

Builds the first standard Python SDK for the official NextGenSwitch Programmable Voice API.

API clients

• Adds typed synchronous and asynchronous clients using HTTPX
• Implements create-call requests with inline XML or a hosted response URL
• Implements active-call modification through PUT /api/v1/call/{call_id}
• Sends the documented authentication headers
• Supports injected HTTPX clients/transports for custom networking and deterministic testing
• Adds context-manager lifecycle, configurable timeouts, safe call-ID encoding, response decoding, and transport errors

Voice XML

• Adds an ElementTree-backed fluent builder with automatic XML escaping
• Supports Say, Play, Gather, Dial, Record, Connect/Stream, Hangup, Pause, Redirect, Bridge, and Leave
• Supports nested Say and Play instructions and Stream parameters

Typed SDK behavior

• Adds immutable ApiResponse, GatherResult, and DialResult models
• Adds ApiError, ValidationError, TransportError, and base NextGenSwitchError
• Ships a py.typed marker for downstream type checking

Package quality

• Adds modern pyproject.toml packaging for Python 3.9+
• Adds Ruff, strict MyPy, pytest, build, and Twine validation
• Adds tests for sync/async requests, authentication, XML generation and escaping, call modification, errors, and callbacks
• Adds GitHub Actions across Python 3.9–3.13
• Adds MIT license, README, and practical create, modify, record, and AI-streaming examples

Documentation source

Implementation follows the official NextGenSwitch Programmable Voice API documentation.

Validation

• Local Python 3.12 checks passed: compile, Ruff, strict MyPy, pytest, wheel/sdist build, and Twine metadata validation
• GitHub Actions will repeat validation across Python 3.9–3.13
• This PR remains an unmerged draft