---
title: "Getting started with krt"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Getting started with krt}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r, include = FALSE}
knitr::opts_chunk$set(collapse = TRUE, comment = "#>")
```

```{r setup}
library(krt)
```

A **Key Resources Table** (KRT) lists the resources a study used and generated,
each paired with a persistent identifier. `krt` models resources around a
neutral, typed core schema and lets you validate, enrich, render, and export
them.

## Build a table

```{r}
k <- new_krt("Dopaminergic neuron study", study_type = "wet-lab")

k <- add_resource(k, "Antibody", "Rabbit Anti-TH",
                  vendor = "Millipore", catalog_number = "AB152",
                  rrid = "RRID:AB_390204", new_or_reuse = "reuse",
                  notes = "Dilution 1:500")

k <- add_resource(k, "Software/code", "Fiji", version = "2.14.0",
                  rrid = "RRID:SCR_002285", new_or_reuse = "reuse")

k <- add_resource(k, "Dataset", "Processed counts",
                  doi = "10.5281/zenodo.11111111", new_or_reuse = "new")
k
```

Identifiers are stored in their own typed fields (`catalog_number`, `rrid`,
`doi`, ...); they are only combined into a compound string at export time. The
author-facing table is a *view* of the underlying records:

```{r}
as.data.frame(k)[, c("resource_type", "display_name", "rrid", "doi")]
```

## Validate

Validation runs structural and semantic rules, with conditional packs that fire
only for the relevant resource types. Severity depends on the profile.

```{r}
validate_krt(k, profile = "generic")
```

Under the stricter ASAP profile, a missing identifier becomes an error:

```{r}
summary(validate_krt(k, profile = "asap"))
```

## Normalize and export

```{r}
k <- normalize_ids(k)

# Lossless canonical formats
cat(substr(write_krt_json(k), 1, 120))
```

Tabular and profile exports are lossy views and warn about it:

```{r}
cat(suppressWarnings(export_krt(k, format = "asap")))
```

## Render for a manuscript

```{r}
cat(render_krt(k, format = "md", profile = "star-methods"))
```

## Provenance

Every step is recorded:

```{r}
as.data.frame(krt_provenance(k))[, c("activity", "software")]
```