---
title: "Reproducible Cached Retrieval and Batching"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Reproducible Cached Retrieval and Batching}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

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

This vignette shows reproducible high-throughput retrieval using:

- `pc_config()` for centralized policy.
- `pc_request(..., cache = TRUE)` deterministic caching.
- `pc_batch()` chunked execution.

```{r setup}
library(PubChemR)
library(tibble)
library(dplyr)
```

## Configure defaults

```{r}
cfg <- pc_config(
  rate_limit = 5,
  retries = 3,
  timeout = 60
)
cfg
```

## Clear cache for a clean run

```{r}
pc_cache_clear()
```

## Cached property retrieval

```{r eval=FALSE}
first <- pc_property(
  identifier = c(2244, 3672, 2519),
  properties = c("MolecularWeight", "XLogP", "TPSA"),
  namespace = "cid",
  cache = TRUE
)

second <- pc_property(
  identifier = c(2244, 3672, 2519),
  properties = c("MolecularWeight", "XLogP", "TPSA"),
  namespace = "cid",
  cache = TRUE
)

# second$from_cache is expected to be TRUE
as_tibble(second)
```

## Chunked batch execution

```{r eval=FALSE}
ids <- c("aspirin", "ibuprofen", "caffeine", "acetaminophen", "naproxen")

batch <- pc_batch(
  ids = ids,
  fn = function(chunk_ids) {
    pc_identifier_map(
      identifier = chunk_ids,
      namespace = "name",
      to = "cids",
      domain = "compound",
      cache = TRUE
    )
  },
  chunk_size = 2
)

batch
as_tibble(batch)
```