<!--
SPDX-FileCopyrightText: 2026 Arcangelo Massari <arcangelo.massari@unibo.it>

SPDX-License-Identifier: ISC
-->

# Triples

Counts RDF triples or quads in files using parallel processing. Supports ZIP, GZIP, and uncompressed files. The output dynamically shows "triples" or "quads" based on the RDF format (quads for `nquads`, triples for `nt` and `json-ld`).

## Usage

```bash
uv run python -m oc_meta.run.count.triples <DIRECTORY> [OPTIONS]
```

## Options

| Option | Default | Description |
|--------|---------|-------------|
| `--pattern` | `*.nq.gz` | Glob pattern for locating files |
| `--format` | `nquads` | RDF format: `nquads`, `nt`, `json-ld` |
| `--recursive` | false | Search subdirectories recursively |
| `--prov-only` | false | Count only files in `prov` subdirectories |
| `--data-only` | false | Count only files not in `prov` subdirectories |
| `--workers` | CPU count | Number of parallel workers |
| `--show-per-file` | false | Print count for each file |
| `--keep-going` | false | Continue processing even if errors occur |

## Examples

Count quads in gzip-compressed N-Quads files:

```bash
uv run python -m oc_meta.run.count.triples /data/rdf --recursive
```

Count triples in ZIP files containing JSON-LD:

```bash
uv run python -m oc_meta.run.count.triples /data/rdf --pattern "*.zip" --format json-ld --recursive
```

Count only data (exclude provenance):

```bash
uv run python -m oc_meta.run.count.triples /data/rdf --recursive --data-only
```

Count only provenance:

```bash
uv run python -m oc_meta.run.count.triples /data/rdf --recursive --prov-only
```

Show per-file counts with 8 workers:

```bash
uv run python -m oc_meta.run.count.triples /data/rdf --recursive --workers 8 --show-per-file
```