org.clojars.roklenarcic.paginator — org.clojars.roklenarcic/paginator 0.2.2

Paginator enables paginating multiple concurrent items with batching.

engine clj

(engine)

(engine async-fn)

Defines a minimal map describing the paging engine. Use with-* to add more options.

async-fn is a fn that takes a no-arg fn and runs it asynchronously, defaults to future-call

Default buffer size for result channel is 100. Default concurrency is 1, see with-concurrency docs for meaning of this setting.

entity-type clj

(entity-type paging-states)

It returns entity-type of the first paging-state in the collection of paging states

merge-expand-result clj

(merge-expand-result paging-state result xf)

Updates paging state with result map. Returns a collection with updated paging state + expansion.

Anything in the result map overwrites current values in paging-state. The exception are keys:

• :page (which increases by 1)
• :items are not added, but instead they are processed through transducer xf and added to the returned collection

merge-expand-results clj

(merge-expand-results paging-states results xf)

Updates paging states with the results + expansions, see merge-expand-result.

merge-result clj

(merge-result paging-state result)

Updates paging state with result map. Anything in the result map overwrites current values in paging-state. The exception are keys:

• :page (which increases by 1)
• :items values are combined

merge-results clj

(merge-results paging-states results)

Updates paging states with the results, see merge-result fn.

paginate! clj

(paginate! engine get-pages-fn entity-ids)

Loads pages with given entity-id pairs. Returns a vector of paging states. Any exceptions are rethrown. entity-ids should be a coll of pairs of entity-type and id.

paginate*! clj

(paginate*! engine get-pages-fn)

Returns a channel for inbound paging state maps and a channel where the paging state maps are returns with additional keys of :items or :exception.

They are returned in the map as :in and :out keys respectively.

get-pages-fn takes paging-states coll and returns new updated paging-states coll, use merge-result, merge-results functions to assist in updating paging states correctly with new items.

paginate-coll! clj

(paginate-coll! engine get-pages-fn entity-type ids)

Loads pages with given IDs. Returns a vector of vectors where each subvector represents items for the id at that index in input parameters. Any exceptions are rethrown.

paginate-one! clj

(paginate-one! engine get-page-fn)

Starts pagination on a single entity and returns items. It expects that there is only 1 result.

get-item-fn should take 1 parameter, the sole paging state, with entity-type ::singleton, id nil

paging-state clj

(paging-state entity-type entity-id)

Returns a base data-structure that describes initial state in paging of results based on some entity.

Initially it doesn't have a :page-cursor key. If the key is present and nil, the entity is considered to be finished caching.

throw-states-exceptions clj

(throw-states-exceptions states)

Throws any exceptions in the states given.

with-batcher clj

(with-batcher engine sorted?)

(with-batcher engine sorted? max-items)

(with-batcher engine sorted? max-items batch-fn)

Add batching configuration to engine map.

sorted? enables using sorted map instead of normal for keeping batches. This comes into play when you have multiple unfinished batches and one needs to be picked for processing, with sorted map the batches with lowest batch key go first.

max-items is the batch size. If pipeline stalls because no batches are available with max size, smaller batches are issued.

batch-fn is the function to generate batch key from paging state map. Defaults to :entity-type

with-concurrency clj

(with-concurrency engine concurrency)

with-result-buf clj

(with-result-buf engine buf-size)

The buffer size for the result channel.