Jenner compatibility tests for prepare-chat-for-nsrr.sas

[lwsinclair]

Jenneranalytics.com provides an API that runs SAS code, with support for more than 200 SAS procedures. You can also use it with Anthropic Claude Code AI in a collaborative workspace. It's available for Mac on the Apple App Store, and by license for Windows and Linux.

The data-prep logic in scripts/prepare-chat-for-nsrr.sas runs on it directly. This PR adds a small jenner-check/ directory with four self-contained bundles, each lifted from a distinct part of that script and paired with a tiny synthetic input so it runs in isolation:

jenner-check/
├── t003_ahi_computation/      # AHI + sleep-maintenance-efficiency formulas
├── t004_harmonized_checks/    # PROC MEANS / PROC FREQ of the harmonized vars
├── t005_psg_decimal_hours/    # PSG lights/onset restore + decimal-hour conversion
├── t006_missing_value_recode/ # sentinel-code-to-missing recodes
└── run_jenner.sh              # ./run_jenner.sh --all

Each bundle keeps the original statements verbatim; only the external libname/Excel sources are swapped for small inline datalines. Run them with ./run_jenner.sh --all, or open any script.sas in the hosted workspace.

What stood out reading the script: the provenance is unusually disciplined — every dropped variable carries a one-line reason, and the harmonization block names the exact CHAT source variable behind each nsrr_* output.

Merge, close, or ignore as you see fit — no response expected. To opt out of future PRs like this, comment no-more-prs or open an issue titled jenner-check: opt out.