# What's New

Highlights from recent releases and commits.

## Schema & Placement

- **SchemaPlacer** — NXDL-guided auto-placement for SPEC and other inputs.
- **Application definition bias** — prefer fields from a chosen application definition (for example
  `NXxas`).
- **Bundled NXDL** — if `external_references/nexus/nexus_definitions` is present, it is used as the
  default `--nxdl-root`.
- **Units-aware scoring** — units such as `eV`, `K`, `deg` steer placement to appropriate fields.
- **CSV metadata** — enrich variables with `description`/`units` values that propagate into
  generated definitions.

## HDF5 & Templates

- **Links vs extract modes** — control how definitions are generated from existing HDF5/NeXus files.
- **Per-scan outputs** — SPEC workflow (`-f`) emits one file per scan plus a master with external
  links.
- **Single-scan templates** — `-t` creates reusable templates expanded across scans during
  conversion.
- **YAML output** — optional YAML definitions when generating templates.
- **Name sanitisation** — invalid dataset/group names are normalised, preserving the original label
  in `@long_name`.

## Inputs & Plugins

- **SPEC** — robust parser with normalised keys and common aliasing.
- **DTA/DAT (batteries)** — RAW/non-RAW/temperature readers plus combined folder generation.
- **PEAXIS** — folder generator and SIF/XAS/DAT parsers.
- **MPES** — generator + parser for multi-photon emission spectroscopy HDF5 data; activated by
  `-b mpes`.
- **YAML** — generator + parser for YAML-format data files with optional SchemaPlacer placement.
- **JSON-LD** — generator + parser for fixed-width/delimited text described via CDI/Schema.org
  JSON-LD (`--jsonld-structure FILE`).
- **Diamond B18 XAFS** — unified ASCII/NeXus parser + NXxas generator; every parsed variable is now linked to the produced definition.
- **HDF5/NeXus** — generator + parser, building definitions from existing files.
- **TIFF** — image-centric definitions with detector/data placeholders and NXdata linkage.
- **Plugin system** — plug-and-play modules; auto-discovered without explicit registration.

## CLI & Batch

- **Batch conversion** — recursion, glob filters, dry-run summaries.
- **Folder generation modes** — `--single-file` vs `--multi-file`.
- **Beamline context** — IKFT, batteries, PEAXIS, MPES modes.
- **Batteries analysis** — `--batteries-analysis` derives EIS/SoC/IV datasets during conversion.
- **DTA/RAW user name** — `--user-name NAME` embeds a user field in generated pair templates;
  otherwise prompted interactively and cached for the run.
- **Prompt literals** — `.nxd` values of the form `?"Prompt text"` ask for user input once per
  run and reuse the answer for all files in a batch.
- **ICAT ingestion** — `-I NUM` appends a `proposal_<NUM>/` subfolder to outputs.
- **Utilities** — `--license`, `--notice`, `--list-beamlines`, CSV export for variables.
- **Auto-generate NXD** — during conversion, create definitions per file when scanning directories.
- **Validation** — `--validate` runs `punx` on generated `.nxs` outputs.
- **Value export** — `--export-values-csv` writes HDF5 dataset values to CSV with optional prefix
  filtering.