Commit Graph

218 Commits

Author SHA1 Message Date
Giambattista Bloisi b19643f6eb Dedup aliases, created when a dedup in a previous build has been merged in a new dedup, need to be marked as "deletedbyinference", since they are "merged" in the new dedup 2024-02-08 15:34:59 +01:00
Claudio Atzori 6fd25cf549 code formatting 2024-01-23 08:47:12 +01:00
Giambattista Bloisi 21a14fcd80 Reusable RunSQLSparkJob for executing SQL in Spark through Oozie Spark Actions
Implements pivots table update oozie workflow
2024-01-15 10:18:14 +01:00
Giambattista Bloisi 3c66e3bd7b Create dedup record for "merged" pivots
Do not create dedup records for group that have more than 20 different acceptance date
2024-01-10 22:59:52 +01:00
Giambattista Bloisi 10e135db1e Use dedup_wf_002 in place of dedup_wf_001 to make explicit a different algorithm has been used to generate those kind of ids 2024-01-10 22:59:52 +01:00
Giambattista Bloisi 831cc1fdde Generate "merged" dedup id relations also for records that are filtered out by the cut parameters 2024-01-10 22:59:52 +01:00
Giambattista Bloisi 1287315ffb Do no longer use dedupId information from pivotHistory Database 2024-01-10 22:59:52 +01:00
Giambattista Bloisi 02636e802c SparkCreateSimRels:
- Create dedup blocks from the complete queue of records matching cluster key instead of truncating the results
- Clean titles once before clustering and similarity comparisons
- Added support for filtered fields in model
- Added support for sorting List fields in model
- Added new JSONListClustering and numAuthorsTitleSuffixPrefixChain clustering functions
- Added new maxLengthMatch comparator function
- Use reduced complexity Levenshtein with threshold in levensteinTitle
- Use reduced complexity AuthorsMatch with threshold early-quit
- Use incremental Connected Component to decrease comparisons in similarity match in BlockProcessor
- Use new clusterings configuration in Dedup tests

SparkWhitelistSimRels: use left semi join for clarity and performance

SparkCreateMergeRels:
- Use new connected component algorithm that converge faster than Spark GraphX provided algorithm
- Refactored to use Windowing sorting rather than groupBy to reduce memory pressure
- Use historical pivot table to generate singleton rels, merged rels and keep continuity with dedupIds used in the past
- Comparator for pivot record selection now uses "tomorrow" as filler for missing or incorrect date instead of "2000-01-01"
- Changed generation of ids of type dedup_wf_001 to avoid collisions

DedupRecordFactory: use reduceGroups instead of mapGroups to decrease memory pressure
2024-01-10 22:59:52 +01:00
Claudio Atzori 431c6bb08a [dedup] added isLookupUrl to the graph consistency workflow definition, required now by the entity grouping phase 2023-12-06 11:06:46 +01:00
Claudio Atzori 03670bb9ce [dedup] use common saveParquet and save methods to ensure outputs are compressed 2023-10-16 10:55:47 +02:00
Claudio Atzori eed9fe0902 code formatting 2023-10-06 12:31:17 +02:00
Claudio Atzori 7b403a920f Merge branch 'beta' into consistency_keep_mergerels 2023-10-02 11:26:00 +02:00
Giambattista Bloisi 2caaaec42d Include SparkCleanRelation logic in SparkPropagateRelation
SparkPropagateRelation includes merge relations
Revised tests for SparkPropagateRelation
2023-09-04 11:33:20 +02:00
Giambattista Bloisi 6cc7d8ca7b GroupEntities and DispatchEntites are now merged in GroupEntitiesSparkJob 2023-08-30 10:43:31 +02:00
Giambattista Bloisi 6b1c05d118 Add sparkExecutorMemoryOverhead workflow config to set off-heap memory for Spark actions. If not explicitly set it is defaulted to 1Gb 2023-08-29 16:04:19 +02:00
Claudio Atzori bf35280ea6 code formatting 2023-08-29 11:11:00 +02:00
Claudio Atzori 58665a246c Merge branch 'beta' into propagate_relation_rewrite 2023-08-29 10:47:02 +02:00
Giambattista Bloisi d012aec0b3 Revert PropagateRelation's argument name from outputPath to graphOutputPath in consistency workflow (#8964) 2023-08-28 22:44:54 +02:00
Giambattista Bloisi a860e19423 Fix ensure all relations are written out, not only those managed by dedup 2023-08-28 15:36:02 +02:00
Giambattista Bloisi 0d7b2bf83d Rewrite SparkPropagateRelation exploiting Dataframe API 2023-08-28 10:34:54 +02:00
Giambattista Bloisi 95cd2b9b1e Make filterInvisible a mandatory parameter of DispathEntitiesSparkJob
Make filterInvisible a mandatory parameter of both dedup/consistency and graph/group oozie workflows
2023-08-10 11:53:48 +02:00
Giambattista Bloisi fab9920271 DispatchEntitiesSparkJob: manage all entity types together, support filtering by dataInfo.invisible flag 2023-08-09 15:41:43 +02:00
Giambattista Bloisi 97b6d1dc45 Filter ids by dataInfo.deletedbyinference and DataInfo.invisible flags
Filter relations also by dataInfo.invisible flag
2023-08-07 10:24:11 +02:00
Giambattista Bloisi af49424b59 Add a "CleanRelation" action after the PropagateRelation to filter out all relations that have been deleyted by inference or that are pointing to dangling entities 2023-08-04 14:27:39 +02:00
Giambattista Bloisi e64c2854a3 Refactor Dedup process to use Spark Dataframe API and intermediate representation with Row interface
JsonPath cache contention fixed by using a ConcurrentHashMap
Blacklist filtering performance improvement
Minor performance improvements when evaluating similarity
Sorting in clustered elements is deterministic (by ordering and identity field, instead of ordering field only)
2023-07-24 15:36:24 +02:00
Giambattista Bloisi 5e15f20e6e Fix entityMerger that was excluding the authors of the first entity in the list to merge 2023-07-21 00:46:54 +02:00
Giambattista Bloisi dba34505de Fix SparkStatsTest bug where parquet tables were incorrectly read as text files leading to unpredictable count() values 2023-07-19 14:24:52 +02:00
Sandro La Bruzzo 9963fd6d29 updated log to add subentity 2023-06-28 13:36:05 +02:00
Sandro La Bruzzo ed7e2ab6d1 reverted mistake on commit workflow.xml 2023-06-28 11:40:19 +02:00
Sandro La Bruzzo 9910ce06ae added to CreateSimRel the feature to write time log 2023-06-28 11:38:16 +02:00
Sandro La Bruzzo bd17c3edc8 added to CreateSimRel the feature to write time log 2023-06-28 11:20:58 +02:00
Claudio Atzori 909729a2fc [dedup] tweaking num partitions, minor changes 2023-05-17 10:16:22 +02:00
Claudio Atzori 062abfd669 fixed NPE, removed unused stuff 2022-12-06 12:04:00 +01:00
Claudio Atzori 0aa725083f extended dedup testing 2022-11-17 16:13:43 +01:00
Claudio Atzori 3dbc637d3e code formatting 2022-11-17 09:55:41 +01:00
Claudio Atzori ddff0e8999 merging duplicates using IdentifierComparator 2022-11-11 16:10:25 +01:00
Claudio Atzori 5af5a8ae42 added IdentifierComparator 2022-11-09 14:20:59 +01:00
Claudio Atzori 61319b2e83 updated dhp-schema version; set entity-level dataInfo before & after merging the fields from the group of duplicates 2022-03-25 16:38:33 +01:00
miconis c959639bd5 dependency updated to the new pace-core version 2022-03-15 16:33:03 +01:00
miconis 8991d097b4 bug fix in the DedupRecordFactory, DataInfo set before merge 2022-02-24 17:13:12 +01:00
Claudio Atzori 391aa1373b added unit test 2022-01-19 17:13:21 +01:00
Claudio Atzori 44a937f4ed factored out entity grouping implementation, extended to consider results from delegated authorities rather than identical records from other sources 2022-01-19 12:24:52 +01:00
Claudio Atzori f4538f3c4c cleanup 2021-11-19 11:33:10 +01:00
Claudio Atzori 2b46b87f56 fixed filtering criteria applied in SparkCopyRelationsNoOpenorgs to keep the parent/child relations from OpenOrgs 2021-11-19 11:30:29 +01:00
Claudio Atzori a24b9f8268 [dedup] trivial refactoring 2021-11-18 17:12:02 +01:00
Claudio Atzori c0750fb17c avoid non necessary count operations over large spark datasets 2021-11-18 17:11:31 +01:00
Claudio Atzori 0a727d325d [dedup] increased number of partitions in the consistency phase 2021-11-16 08:43:41 +01:00
miconis 611ca511db set configuration property in openorgs duplicates wf 2021-10-07 15:39:55 +02:00
miconis 9646b9fd98 implementation of the http call for the update of openorgs suggestions 2021-10-07 11:29:11 +02:00
miconis 853333bdde implementation of the whitelist for similarity relations 2021-09-20 16:21:47 +02:00