DEXAnalyzer

Nexthink NQL API toolkit for querying, caching, and analyzing Digital Employee Experience data.

Features

• NQL API client — OAuth 2.0 auth, v2/execute (fast, 1k limit) and v1/export (unlimited, async) endpoints
• Persistent query cache — Parquet-based file cache with configurable TTL to avoid repeated API calls
• Analysis tools — One file per tool, operating on cached DataFrames
• AI summaries — Optional Claude-powered natural language insights (requires Anthropic API key)

Setup

pip install -e .

Copy .env.example to .env and fill in your credentials:

NEXTHINK_INSTANCE=yourcompany
NEXTHINK_REGION=eu
NEXTHINK_CLIENT_ID=...
NEXTHINK_CLIENT_SECRET=...

Optional settings:

ANTHROPIC_API_KEY=sk-ant-...    # Enables AI summaries
NQL_LOG_LEVEL=INFO              # DEBUG, INFO, WARNING (default)

Analysis Tools

Call CPU Analysis

Compares CPU efficiency of Teams vs Zoom during calls by matching 15-minute execution samples to call time windows on the same device.

python -m nql_analyzer.tools.call_cpu_analysis

How it works:

1. Downloads call data (#call_analysis_calls) and execution samples (#call_analysis_executions) via the export API
2. Maps call apps to binaries (Teams → msteams/ms-teams.exe, Zoom → zoom.us/zoom.exe)
3. For each call, finds execution samples fully contained within the call window on the same device
4. Computes average CPU % per second from matched samples, extrapolates to full call duration
5. Reports summary by app+platform and detail by binary+platform with percentiles

Project Structure

src/nql_analyzer/
├── config.py       # Settings from .env (instance, region, credentials)
├── auth.py         # OAuth 2.0 token management with auto-refresh
├── client.py       # NQL API client (execute + export)
├── cache.py        # Parquet file cache with TTL
└── tools/
    └── call_cpu_analysis.py   # Teams vs Zoom CPU efficiency

NQL Queries

The analysis tools rely on pre-saved NQL queries that you must create in your Nexthink instance before running any tool:

1. Log in to your Nexthink web console
2. Navigate to Administration > NQL API queries
3. Create each query listed in nql_queries.md using the exact queryId and NQL syntax provided
4. Ensure the API credentials in your .env have permission to execute these queries

License

MIT — see LICENSE.