Database processors

List of database processors:

• Database relations
  • ZooRelation
  • RelationsPopulator
• High-level FLM content processing
  • ZooFLMProcessor
• Low-level FLM content object field compilation
  • FLMSimpleContentCompiler
• Git “last modified” information
  • GetGitLastModifiedDbProcessor
• Index searchable content
  • SearchableTextFieldset
  • SearchableTextProcessor
  • FLMSearchableDocTextValuesAssembler

Base class

import { ZooDbProcessorBase } from '@phfaist/zoodb/dbprocessor/base';

class ZooDbProcessorBase()

Base class for a ZooDb database processor.

Methods in this class are called at various points when loading (or reloading) the zoo data; subclasses should reimplement them to do something useful.

ZooDbProcessorBase.ZooDbProcessorBase

ZooDbProcessorBase.initialize_zoo()

This method is called after data is loaded into zoo, before process_zoo().

In case the database processor does not handle reloads separately from initial loads (e.g., if process_zoo_update_objects() is not reimplemented), then both of those events call process_zoo(). However, the initialize_zoo() is only called once upon the initial zoo load event.

This method is await’ed when called by the main zoo instance, so it can be declared async or return a Promise.

ZooDbProcessorBase.install_zoo(zoodb)

Will be called very early by the ZooDb object to make itself aware to the database processor. You don’t need to reimplement this function. It simply stores the zoodb instance in this.zoodb.

ZooDbProcessorBase.prepare_zoo_update_objects(db_objects)

Called before certain objects are reloaded and updated to new values. The argument db_objects has the structure db_objects[object_type][object_id] = new_object_data and contains the new, updated data that will be loaded into the zoo. The current, old object data that will be replaced is available in this.zoodb.objects.

This method is await’ed when called by the main zoo instance, so it can be declared async or return a Promise.

ZooDbProcessorBase.process_data_dump(data, options)

Process the given data, which is a ZooDb raw export data, so as to remove any annoying object instances in view of serialization.

For instance, an FLM content processor will likely want to replace FLMFragment objects by e.g. the corresponding FLM code representation.

Return the data as properly modified. You need to return the data even if you modify it in-place (but read the warning below!).

Warning

EXTRA CARE IS NECESSARY WHEN ANY MODIFICATIONS OF THE DATA ARE NECESSARY. TL,DR: You can modify objects as long as you assign the object’s fields (as declared in the schema) directly. If the object field value is an object instance (through some db processor), do NOT directly modify that object, IT IS THE DBPROCESSOR’S RESPONSIBILITY TO MAKE A DEEP COPY OF THE FIELD’S VALUE.

The zoodb’s data dump procedure cannot begin by making a deep copy of the data, because the field values might have been processed to custom object instances e.g. with circular references, and we might not know how to copy such objects. It is therefore the database processor’s responsibility to make sure that suitable copies are made whenever appropriate.

The data itself and data.db are already new object instances. The data.db.schemas object is already a full deep copy of the original schemas. The data.db.objects is copied recursively only up to the object properties that are fields as declared in the schema. Fields that are objects themselves and that have subfields declared in the schema are also recursed into. The _zoodb properties of each object is also already a deep copy of the original data. However the field values cannot be copied at the beginning of the export by the ZooDb base dump method, because we might not know how to copy the object. In other words, while the general object structure is copied, the object’s field values are not.

In all cases, you need to return the resulting data.

This method is await’ed when called by the main zoo instance, so it can be declared async or return a Promise.

ZooDbProcessorBase.process_zoo()

Perform any processing on the zoo data that is relevant to this database processor. The zoodb instance is available as this.zoodb.

Unless you reimplement process_zoo_update_objects(), this method is also called when certain objects are reloaded and updated. In this case, make sure that this method can be called multiple times; don’t assume that the data in this.zoodb is fresh and hasn’t been processed yet.

This method is await’ed when called by the main zoo instance, so it can be declared async or return a Promise.

ZooDbProcessorBase.process_zoo_update_objects(db_objects)

Perform any processing that is relevant to this database processor, on the specified objects that have just been reloaded and updated in the zoo.

The default implementation calls process_zoo() to re-process the entire zoo.

The argument db_objects has the structure db_objects[object_type][object_id] = new_object_data and contains the new, updated data that will be loaded into the zoo.

This method is await’ed when called by the main zoo instance, so it can be declared async or return a Promise.