Issues

• PageQuery title patterns and getDiscussionArchives fail on PostgreSQL columns with a nondeterministic collation

  🐛Bug📖DocumentationModule: API
  dkpro/dkpro-jwpl#751 · Filter by author rzo1 opened on Oct 1, 2026 · 2.4.0
  

• Unescaped '_' in LIKE patterns of getDiscussionArchives and the template fragment lookups matches any character

  🐛BugModule: APIModule: Util
  dkpro/dkpro-jwpl#750 · Filter by author rzo1 opened on Oct 1, 2026 · 2.4.0
  

• PageQuery title pattern matches case-insensitively on case-insensitive collations

  🐛BugModule: API
  dkpro/dkpro-jwpl#748 · Filter by author rzo1 opened on Oct 1, 2026 · 2.4.0
  

• Wikipedia.getPageByExactTitle(String) binds the title with blanks and never finds a multi-word title

  🐛BugModule: API
  dkpro/dkpro-jwpl#747 · Filter by author rzo1 opened on Oct 1, 2026 · 2.4.0
  

• Wikipedia.getCategories(String) does not normalize the title and returns categories of case variants on case-insensitive collations

  🐛BugModule: API
  dkpro/dkpro-jwpl#746 · Filter by author rzo1 opened on Oct 1, 2026 · 2.4.0
  

• Wikipedia.getPageIds(String) does not normalize the title and returns all case variants on case-insensitive collations

  🐛BugModule: API
  dkpro/dkpro-jwpl#745 · Filter by author rzo1 opened on Oct 1, 2026 · 2.4.0
  

• Wikipedia.getPage(String) follows the column collation and returns case, accent or ß/ss variants of the title

  🐛BugModule: API
  dkpro/dkpro-jwpl#744 · Filter by author rzo1 opened on Oct 1, 2026 · 2.4.0
  

• Backslashes in XML dump titles and texts are escaped twice for the database import

  🐛BugModule: DataMachineModule: TimeMachineModule: WikiMachine
  dkpro/dkpro-jwpl#683 · Filter by author rzo1 opened on Sep 23, 2026 · 2.4.0

• Support PostgreSQL as a database backend

  🆕EnhancementModule: APIModule: DataMachineModule: RevisionMachine
  ⭐️ Enhancement

  dkpro/dkpro-jwpl#674 · Filter by author rzo1 opened on Sep 23, 2026 · 2.3.1
  

• Performance audit: overview of automated findings (Hibernate/JDBC queries, DataMachine throughput)

  🆕Enhancementmajor
  dkpro/dkpro-jwpl#619 · Filter by author rzo1 opened on Sep 22, 2026 · 2.4.0

• TimeMachine: every parsed row is handed to all N snapshot versions, one after another, on the parser thread

  🆕EnhancementModule: TimeMachineModule: WikiMachine
  dkpro/dkpro-jwpl#613 · Filter by author rzo1 opened on Sep 22, 2026 · 2.4.0

• RevisionMachine: lazy Revision.getRevisionText() repeats the index lookup and re-fetches the diff chain; lazy RevisionIterator selects an unused Revision blob

  🆕EnhancementModule: RevisionMachine
  ⭐️ Enhancement

  dkpro/dkpro-jwpl#610 · Filter by author rzo1 opened on Sep 22, 2026 · 2.3.1
  

• WikipediaInfo: constructor loads every page to count them, fan-out loads every link target, category overlap is O(C²), and the lazy-init guards never fire

  🐛BugjavaModule: API
  🐛 Bug

  dkpro/dkpro-jwpl#602 · Filter by author rzo1 opened on Sep 22, 2026 · 2.3.1
  1

• CategoryGraph cache returns the wrong graph for subsets, and graph/map files use slow Java serialization

  🐛BugModule: API
  dkpro/dkpro-jwpl#598 · Filter by author rzo1 opened on Sep 22, 2026 · 2.4.0
  1

  

• RevisionMachine: add an opt-in binary (MEDIUMBLOB) revision storage mode to SQLDatabaseWriter; the existing binaryTask path is dead and broken

  🆕EnhancementModule: RevisionMachine
  dkpro/dkpro-jwpl#596 · Filter by author rzo1 opened on Sep 22, 2026 · 2.4.0

• Template info generator: use_revision_iterator=false rebuilds every revision from its full revision via getRevision(pageId, timestamp)

  🆕EnhancementModule: Util
  dkpro/dkpro-jwpl#591 · Filter by author rzo1 opened on Sep 22, 2026 · 2.4.0

• ModularParser: every SpanManager edit walks all managed spans, and each content element rescans the page-wide link/template/tag/nowiki/math lists

  🆕EnhancementModule: Parser
  dkpro/dkpro-jwpl#586 · Filter by author rzo1 opened on Sep 22, 2026 · 2.4.0

• PageQueryIterable loads and parses every candidate page one at a time, and getNumberOfPages(PageQuery) builds the full result just to count it

  🆕EnhancementModule: API
  dkpro/dkpro-jwpl#574 · Filter by author rzo1 opened on Sep 22, 2026 · 2.3.1
  

• WikipediaTemplateInfo.getArticleRevisionPairs: O(n²) LinkedList.contains, three revision-block reads and one full-text Page load per candidate revision

  🆕EnhancementModule: Util
  dkpro/dkpro-jwpl#562 · Filter by author rzo1 opened on Sep 22, 2026 · 2.3.1
  

• WikipediaTemplateInfo.getRevisionPairs*: three queries plus a diff-BLOB range scan per revision just to get its id, then one template query per revision

  🆕EnhancementModule: Util
  dkpro/dkpro-jwpl#561 · Filter by author rzo1 opened on Sep 22, 2026 · 2.4.0

• Template info dump is built in memory as one giant String, and the configured maxAllowedPackets is ignored

  🐛Bug🆕EnhancementModule: Util
  dkpro/dkpro-jwpl#560 · Filter by author rzo1 opened on Sep 22, 2026 · 2.4.0

• DataMachine: run independent pipeline stages (and multi-part XML parts) in parallel, deterministically

  🆕EnhancementjavamajorModule: DataMachineModule: WikiMachine⚙️Refactoring
  ⭐️ Enhancement

  dkpro/dkpro-jwpl#559 · Filter by author rzo1 opened on Sep 22, 2026 · 2.4.0
  1

  

• DataMachine: drop the revision pass and the boxed textId-to-pageId map when rejoining text.bin, and stop writing unused category texts

  🆕EnhancementModule: DataMachine
  dkpro/dkpro-jwpl#555 · Filter by author rzo1 opened on Sep 22, 2026 · 2.4.0

• Linktarget resolver keeps one String title per target (including unused talk pages): resolve lt_id to page/category ids once instead

  🆕EnhancementModule: WikiMachine
  dkpro/dkpro-jwpl#553 · Filter by author rzo1 opened on Sep 22, 2026 · 2.4.0

• Replace char-by-char StreamTokenizer SQL dump parsing with a byte-level tokenizer that only materialises the columns it needs

  🆕EnhancementModule: WikiMachine
  dkpro/dkpro-jwpl#552 · Filter by author rzo1 opened on Sep 22, 2026 · 2.3.1