Class: BaseAdapter

undercomplicate.BaseAdapter()

new BaseAdapter()

Parameters:
Name Type Attributes Default Description
config.cache_enabled boolean <optional>
 true

Whether to enable the LRU cache, and store a copy of the normalized and parsed response data. Turned on by default for most remote requests; turn off if you are using another datastore (like Vuex) or if the response body uses too much memory.
config.cache_size number <optional>
 3

How many requests to cache. Track-dependent annotations like LD might benefit from caching more items, while very large payloads (like, um, TOPMED LD) might benefit from a smaller cache size. For most LocusZoom usages, the cache is "region aware": zooming in will use cached data, not a separate request
Source:

Methods

_annotateRecords(records, options) → {*}

Perform custom client-side operations on the retrieved data. For example, add calculated fields or perform rapid client-side filtering on cached data. Annotations are applied after cache, which means that the same network request can be dynamically annotated/filtered in different ways in response to user interactions.

This result is currently not cached, but it may become so in the future as responsibility for dynamic UI behavior moves to other layers of the application.

Parameters:
Name Type Description
records Array.<Object>
options Object
Source:
Returns:
Type
*

_buildRequestOptions(options, dependent_data) → {*}

Build an object with options that control the request. This can take into account both explicit options, and prior data.

Parameters:
Name Type Description
options Object

Any global options passed in via getData. Eg, in locuszoom, every request is passed a copy of plot.state as the options object, in which case every adapter would expect certain basic information like chr, start, end to be available.
dependent_data Array.<Object>

If the source is called with dependencies, this function will receive one argument with the fully parsed response data from each other source it depends on. Eg, ld(assoc) means that the LD adapter would be called with the data from an association request as a function argument. Each dependency is its own argument: there can be 0, 1, 2, ...N arguments.
Source:
Returns:

An options object containing initial options, plus any calculated values relevant to the request.

Type
*

_getCacheKey(options) → {*}

Determine how this request is uniquely identified in cache. Usually this is an exact match for the same key, but it doesn't have to be. The LRU cache implements a find method, which means that a cache item can optionally be identified by its node metadata (instead of exact key match). This is useful for situations where the user zooms in to a smaller region and wants the original request to count as a cache hit. See subclasses for example.

Parameters:
Name Type Description
options object

Request options from _buildRequestOptions
Source:
Returns:

This is often a string concatenating unique values for a compound cache key, like chr_start_end. If null, it is treated as a cache miss.

Type
*

_normalizeResponse(response_text, options) → {*}

Convert the response format into a list of objects, one per datapoint. Eg split lines of a text file, or parse a blob of json.

Parameters:
Name Type Description
response_text *

The raw response from performRequest, be it text, binary, etc. For most web based APIs, it is assumed to be text, and often JSON.
options Object

Request options. These are not typically used when normalizing a response, but the object is available.
Source:
Returns:

A list of objects, each object representing one row of data {column_name: value_for_row}

Type
*

_performRequest(options) → {Promise}

Perform the act of data retrieval (eg from a URL, blob, or JSON entity)

Parameters:
Name Type Description
options object

Request options from _buildRequestOptions
Source:
Returns:
Type
Promise

_postProcessResponse(records, options)

A hook to transform the response after all operations are done. For example, this can be used to prefix fields with a namespace unique to the request, like log_pvalue -> assoc:log_pvalue. (by applying namespace prefixes to field names last, annotations and validation can happen on the actual API payload, without having to guess what the fields were renamed to).

Parameters:
Name Type Description
records
options
Source:

getData(options, …dependent_data) → {Promise.<*>}

All adapters must implement this method to asynchronously return data. All other methods are simply internal hooks to customize the actual request for data.

Parameters:
Name Type Attributes Description
options object

Shared options for this request. In LocusZoom, this is typically a copy of plot.state.
dependent_data Array.<Array> <repeatable>

Zero or more recordsets corresponding to each individual adapter that this one depends on. Can be used to build a request that takes into account prior data.
Source:
Returns:
Type
Promise.<*>