Cache Dataset

CacheDataset<D> wraps any dataset and caches the loaded/saved value in memory. Subsequent loads return the cached value without hitting the underlying dataset.

Requires the std feature.

Definition

pub struct CacheDataset<D: Dataset> {
    pub dataset: D,
    #[serde(default)]
    pub take: bool,
    cache: Arc<Mutex<Option<D::LoadItem>>>,
}

Usage

Wrap any dataset to add caching:

#[derive(Serialize, Deserialize)]
struct Catalog {
    readings: CacheDataset<PolarsCsvDataset>,
}

readings:
  dataset:
    path: data/readings.csv
    separator: ","

Behavior

First load() — delegates to the inner dataset, caches the result, returns it
Subsequent load() calls — returns the cached value without re-reading the file
save() — writes to the inner dataset and updates the cache
html() — delegates to the inner dataset
remover() — clears the cache and removes the inner dataset’s value (used by RetentionHook)

Taking the value

With take: true (or CacheDataset::new(ds).with_take(true)), the first load() after a save() moves the value out of memory instead of cloning it, and later loads read the inner dataset again:

checkpoint:
  dataset:
    path: ckpt/epoch_3.mpk
  take: true

This suits a value with a single consumer, such as one checkpoint in an epoch chain: the next epoch reads the weights straight from memory, skipping deserialization, and memory never holds more than the checkpoints not yet consumed. Without take, every epoch’s CacheDataset would keep its weights for the rest of the run. A second reader still works — it loads from disk.

When to use

Use CacheDataset when a dataset is read by multiple nodes and the underlying I/O is expensive:

(
    Node { name: "analyze", input: (&cat.readings,), .. },
    Node { name: "validate", input: (&cat.readings,), .. },
    Node { name: "summarize", input: (&cat.readings,), .. },
)

Without caching, readings would be loaded from disk three times. With CacheDataset<PolarsCsvDataset>, it’s loaded once and served from memory for the remaining reads.

Constraints

The inner dataset must satisfy:

D::LoadItem: Clone + Send — so the cached value can be cloned on each load
D::SaveItem: Clone + Into<D::LoadItem> — so saves can update the cache
PondError: From<D::Error> — for error conversion

Keyboard shortcuts