Sophia Leierseder authored 5 years ago and Johannes Goderbauer committed 5 years ago

(cherry picked from commit 52870197)

Sophia Leierseder authored 5 years ago and Johannes Goderbauer committed 5 years ago

(cherry picked from commit 775617d2)

If a merge has been successful, a new acitivity gets created on the target record of the merge
This has been implemented on Person and Organisation

It's now possible to have a scanner configuration without any prefilters. No threshold logic is applied, only the configured fields are used to search.
On the other hand is it now possible to have a scanner with only a prefilter.

Fixed error where if no duplicate scan result on creation of a new person could be found, the code tried to access .lenght...
Data_alias.aod: The auditSyncConfig was configured to audit itself which resulted in an error. Fixed.

RebuildAllDuplicateCaches_serverProcess.aod takes both person and organisation scanner (currently hard coded) and runs them. This is ought to be executed via the manager/other serverside ways
RebuildDuplicatesCache_serverProcess on the other hand should be used via code because this process expects the filterName and TargetEntity as parameters. The usages are commented out for the time beign becaus of incompatibilities in the process.execute* methods...
If a person or organisation is merged, the standardcommunication ways of the old record are reset to "no standard".

If duplicates have been found while saving a new contact, a notification will be shown, alerting the user

This has been integrated into the action to rebuild person and org duplicates

If a person is added, a scan for duplicates is triggered automatically. if a matching cluster exists the new record gets added, otherwise a new cluster gets created
The field, which is currently used by the demo duplicatescan (firstname) is now another type because the phonetic matcher creates seemingly wrong searchresults. If the scan for duplicates or the indexsearch results in seemingly unrelated results, set the indexFieldType to TEXT_NO_STOPWORDS.

refactorings

Feld DB_FIELD_NAME removed
Created fields for defaultvalues city,zip,... to use in duplicatesearchconfig
Custom fields added in subProcess to have them added to the index

S.Listl authored 5 years ago and Sebastian Listl committed 5 years ago

(cherry picked from commit f79a987f)

S.Listl authored 5 years ago and Sebastian Listl committed 5 years ago

(cherry picked from commit caac232e)

S.Listl authored 5 years ago and Sebastian Listl committed 5 years ago

(cherry picked from commit c708b72d)

The function ScanForDuplicates has been refactored and is now much more easy to use because it does much more stuff inside to have an easier api.
"RebuildDuplicatesCache" and "ScanForDuplicates" have been refactored.
The fields used in the prefilter don't have to be configured manually by the user. Instead, the fields defined by the prefilter now get added automatically to the list of fields to be used. In the case, that the field has been added via the users config anyways, it doesn't get added by code to ensure that each field is mentioned exactly once.
To reflect these changes, the javadoc including the example of ScanForDuplicates and RebuildDuplicates has been updated.
 Until it's possible to load values from a consumed entity dynamically via code, the possibility to select entity fields from consumed entities via ui has been removed. See ticket #1044577. Now only "native" fields directly from the base entity can be selected.
 Since the entity fields are now used in the configuration of the duplicate search, no db fields have to be configured any more. The field has been removed from liquibase

The Cache rebuilding has been changed. The values to check for duplicates are now loaded using the entities api instead of pure sql. If the cache for all duplicates of a scanner gets rebuilt, it's now done in blocks. This ensures that even while wokring on huge recordsets, the ressources are enough. It sure may take some time to check huge recordsets...
The flag "Use external Webservice" has been removed from the view because no native api to a duplicate scanning service exists, therefore this is unnecessary right now. The config of the resultfields has been removed aswell, if no external service is used, it's enough to work with the id which is always in the index's response.
The configuration of the indexfields has been changed, no manual connection between entity and db field is necessary because no sql is used any more. Instead, the fields can be selected via dropdown which offers the entity fields to select.
Some refactoring
Changes on the docks to reflect the changed datastructure of some arrays

The possibility to configure an external duplicate has been removed from the views because no default external service integration exists. Reactivate later maybe..

Removed unnecessary log outputs
Wrong behaviour removed. If a valid prefilter exists but the filterpatternfields contain no value, the indexsearch got triggered nonetheless. The results which have been found on the base of the prefilter had been defined as duplicates which was wrong.
Code examples have been added to the important functions like ScanForDuplicates or RebuildDuplicatesCache

Refactorings

Documentation
Example code to show the usage of the webservice callback

If a field which is configured to be used in the indexsearch contains no value, no pattern will be added @see _buildFilterPatternConfig

The flag if an external webservice should be used is now used in the code. If true, the parameterized callbackfunction gets called. See DuplicateScannerUtils._isUseExternalWebservice

The possibility to integrate a webservice call and the necessary formatting of the values has been developed. The function RebuildDuplicatesCache now awaits a function as parameter in which an external developer can implement the steps to format the values and call the desired webservice. There is no native webservice call implementation because the formfactor of the data send to a webservice is always different and how with the particular webservices has to be communicated cant be known by the standard implementation.
The customized function gets an array which holds values based on the configured resultFields. To access a value the key which is its index field ought to be used. The resultfields, which are known at this place, can be converted to their respective index field name pendants using DuplicateScannerUtils.TranslateEntityToIndexFields which returns some kind of map (access via key).

New function DuplicateScannerUtils.TranslateEntityToIndexFields

Old functions to rebuild the duplicate cache have been removed and the ones with the postfix "Index"(Which was also removed) replaced them
New function to load the DB_FIELDS from the IndexFieldConfig and concatenate them to use the resulting string as fields-to-select in a db select query

Refactoring to get a small glimpse of cleancodish structure
More checks and case handlings to increase stability

-> each consumer to ObjectTree can define several object types/ids which should be displayed

The Indexfieldconfig now contains the correct fields which are necessary to run the search.
The fields now can be configured via the client. Those infos are now used to dynamically load the configured fields an their values which are then internally used to run the prefilter as well as the real duplicate search.
NEXT UP => Configuring of the result fields. There's light at the end of the tunnel.