* upstream/main:
  Update PULL_REQUEST_TEMPLATE.md (#15788)
  New Crowdin updates (#15787)
  Update heylogs to 0.18.0 and use github-actions format (#15786)
  Grand refactoring of the AI features (#15688)
  Chore(deps): Bump com.fasterxml:aalto-xml in /versions (#15782)
  Chore(deps): Bump org.junit:junit-bom from 6.0.3 to 6.1.0 in /versions (#15783)
  Fix default value for unwanted characters (#15743)
  Fix runner tag
  Fix runner for JBang (PR)
  Fix duplicate finder progress counter incrementing on empty queue polls (#15781)
  Refine JabKit CLI: positional input argument and check command group (#15759)
  Ignore exception in unregisterListener to prevent exception (#15761)
  Fix wrong usage of "key" (#15779)
  Fix Hayagriva export to nest identifiers under serial-number (#15750)

* Refactor chat history

* refactor chat history + start working on UI

* initroduce 1 class. idk why

* Move out database listeners

* middle work

* middle work

* middle work

* middle work

* Start working on tasks

* middle work

* Somewhat AiSummary

* middle work

* summary almost done

* middle work

* Finish AI summary

* Refactor locations

* Start working on AI chat UI

* Use single chat history repository

* Remove old classes

* Remove old classes x2

* Middle work

* middle work

* finish chat?

* Something works

* bug fixing

* fix Ai chat

* Implement status window

* Fixings

* rename identifiers

* Make group window

* Fix window close

* Some arch changes

* Add ADRs

* Some changes

* mid work

* Almost finished

* Add draft reqs

* fix ai chat

* add ai OFT

* add adr for messages

* Empty messages package

* add ADR on messages

* add new types

* Migrate to new type (todo: migrate messages v1->v2, show debug, export, oft)

* feat: mid work on removing

* refactor(ai): refactor AiDatabaseListener and AiFeature

* refactor(ai): remove AiTemplateKind

* refactor(ai): fix AiDatabaseListeners

* refactor(ai): start AiSummarizationLogic

* refactor(ai): petit change

* refactor(ai): remove some user message templates

* refactor(ai): add ai library id

* refactor(ai): refactor ChatIdentifier with ai library id

* refactor(ai): refactor AiSummaryIdentifier

* refactor(ai): rename AI identifiers

* refactor(ai): rename AI identifiers

* refactor(ai): remove AiTemplateKind.java

* refactor(ai): change tokenizators parameters

* refactor(ai): remove customimplementations package

* refactor(ai): make TokenizationAiFeature

* refactor(ai): change order of the responsibility chain

* refactor(ai): use string templates

* chore(gui): remove unused file

* fix(ai): add system message to the chat

* refactor(ai): use hash for ingested documents tracking

* refactor(ai): add GenerateEmbeddingsAiDatabaseListener and move files

* refactor(ai): add transfer of summaries

* refactor(ai): petit change

* refactor(ai): simplify templates

* refactor(ai): fix name

* refactor(ai): add factories

* refactor(ai): refactor summarization + add RAM layer

* refactor(ai): make migrations

* refactor(ai): remove features

* refactor(ai): remove AiChatLogic and introduce GenerateRagResponseTask

* refactor(ai): clean-up code

* Initial plan

* Initial plan

* feat: implement AI export feature for chat and summary

Agent-Logs-Url: https://github.com/InAnYan/jabref/sessions/97e3a57c-35e3-47c4-872e-370399072f7f

Co-authored-by: InAnYan <13097618+InAnYan@users.noreply.github.com>

* fix: use correct BibEntryWriter API for BibTeX serialization

Agent-Logs-Url: https://github.com/InAnYan/jabref/sessions/97e3a57c-35e3-47c4-872e-370399072f7f

Co-authored-by: InAnYan <13097618+InAnYan@users.noreply.github.com>

* refactor: extract helper methods to reduce duplication in AiChatView export

Agent-Logs-Url: https://github.com/InAnYan/jabref/sessions/97e3a57c-35e3-47c4-872e-370399072f7f

Co-authored-by: InAnYan <13097618+InAnYan@users.noreply.github.com>

* Implement follow-up questions feature for AI chat

Agent-Logs-Url: https://github.com/InAnYan/jabref/sessions/a7d0619b-da31-4013-b6bc-c945b94570c6

Co-authored-by: InAnYan <13097618+InAnYan@users.noreply.github.com>

* Fix magic number and duplicate regex pattern issues from code review

Agent-Logs-Url: https://github.com/InAnYan/jabref/sessions/a7d0619b-da31-4013-b6bc-c945b94570c6

Co-authored-by: InAnYan <13097618+InAnYan@users.noreply.github.com>

* Fix constant placement and remove duplicate constant for code review feedback

Agent-Logs-Url: https://github.com/InAnYan/jabref/sessions/a7d0619b-da31-4013-b6bc-c945b94570c6

Co-authored-by: InAnYan <13097618+InAnYan@users.noreply.github.com>

* Changes before error encountered

Agent-Logs-Url: https://github.com/InAnYan/jabref/sessions/d86d0212-7cca-4a28-b941-2d31a9b6bfbc

Co-authored-by: InAnYan <13097618+InAnYan@users.noreply.github.com>

* Address review comments: BackgroundTask for follow-up questions, ViewModel actions, cleanup

Agent-Logs-Url: https://github.com/InAnYan/jabref/sessions/96988cba-e6b3-47de-b08a-5099969cbca9

Co-authored-by: InAnYan <13097618+InAnYan@users.noreply.github.com>

* Update AiTab.java

* Update AiTab.java

* refactor(ai): update embeddings to use file hash

* Update GenerateFollowUpQuestions.java

* refactor: move export logic to view models, add interfaces, add markdown export template

- Introduce AiChatExporter and AiSummaryExporter interfaces
- Add configurable markdown export Velocity template to AiPreferences
  (AiDefaultTemplates, AiPreferences, JabRefCliPreferences, AiTemplateRenderer)
- Add ExportMessage helper class to AiTemplateRenderer for Velocity template context
- Rewrite AiChatMarkdownExporter to use Velocity template rendering
- Make AiChatJsonExporter and AiChatMarkdownExporter implement AiChatExporter
- Make AiSummaryJsonExporter and AiSummaryMarkdownExporter implement AiSummaryExporter
- Move exportMarkdown/exportJson from AiChatView to AiChatViewModel
- Move exportMarkdown/exportJson from AiSummaryView to AiSummaryViewModel
- Add markdown export template tab to AI preferences UI (AiTab.fxml, AiTab.java, AiTabViewModel)
- Fix wrong Javadoc in AiChatJsonExporter (said Markdown, now says JSON)

Agent-Logs-Url: https://github.com/InAnYan/jabref/sessions/0643754f-446a-4022-89a2-03f5a4e38465

Co-authored-by: InAnYan <13097618+InAnYan@users.noreply.github.com>

* refactor: improve method names per code review feedback

Agent-Logs-Url: https://github.com/InAnYan/jabref/sessions/0643754f-446a-4022-89a2-03f5a4e38465

Co-authored-by: InAnYan <13097618+InAnYan@users.noreply.github.com>

* refactor: add AiMetadata record, remove ExportMessage, use ChatMessage in templates

- Create AiMetadata record in org.jabref.model.ai with aiProvider, model, timestamp
- Update AiChatExporter and AiSummaryExporter interfaces to accept AiMetadata
- Remove ExportMessage helper class from AiTemplateRenderer; pass ChatMessage directly
- Update renderMarkdownChatExport to accept AiMetadata and List<ChatMessage>
- Update default template to use $message.role().getDisplayName() and $message.content()
- Update AiChatJsonExporter to use AiMetadata (single export method)
- Update AiChatMarkdownExporter to pass metadata and filtered ChatMessages
- Update AiSummaryMarkdownExporter and AiSummaryJsonExporter to accept AiMetadata
- Update AiChatViewModel.buildMetadata() from ChatModel
- Update AiSummaryViewModel.buildMetadata() from AiSummary

Agent-Logs-Url: https://github.com/InAnYan/jabref/sessions/ef6acf07-7477-4765-b518-b687ea537b01

Co-authored-by: InAnYan <13097618+InAnYan@users.noreply.github.com>

* style: convert /// comments to standard Javadoc in AiMetadata and AiTemplateRenderer

Agent-Logs-Url: https://github.com/InAnYan/jabref/sessions/ef6acf07-7477-4765-b518-b687ea537b01

Co-authored-by: InAnYan <13097618+InAnYan@users.noreply.github.com>

* refactor: embed AiMetadata in AiSummary; remove AiMetadata.empty()

- Remove AiMetadata.empty() factory method
- Add Jackson annotations to AiMetadata for JSON serialization
- AiSummary now stores AiMetadata (replacing timestamp/aiProvider/model fields)
- Update BibEntrySummarizator to build AiMetadata when creating AiSummary
- Update SummariesMigration to build AiMetadata when converting old summaries
- Update AiSummaryShowingView to use summary.metadata().xxx()
- Update AiSummaryViewModel.buildMetadata() to return summary.metadata() directly
- Update AiChatViewModel.buildMetadata() to inline empty-model fallback

Agent-Logs-Url: https://github.com/InAnYan/jabref/sessions/2ad7966f-6d03-41db-8c8e-17e6f58a421d

Co-authored-by: InAnYan <13097618+InAnYan@users.noreply.github.com>

* refactor: move export responsibility to AiSummaryShowingViewModel

- AiSummaryShowingViewModel now takes AiPreferences, FieldPreferences,
  BibEntryTypesManager, DialogService and owns the export logic directly
- AiSummaryShowingViewModel gains entryProperty (FullBibEntry) for export
- AiSummaryShowingView gains @Inject for GuiPreferences, DialogService,
  BibEntryTypesManager; passes them to AiSummaryShowingViewModel
- AiSummaryShowingView exposes entryProperty() so AiSummaryView can bind it
- AiSummaryView.setupBindings() now also binds summaryShowing.entryProperty()
- AiSummaryView no longer has exportMarkdown/exportJson FXML handlers
- AiSummaryViewModel export methods and related imports removed
- AiSummary.fxml: removed onExportMarkdown/onExportJson handler wiring
  (export is now fully self-contained in AiSummaryShowingView)

Agent-Logs-Url: https://github.com/InAnYan/jabref/sessions/a09f6212-dde3-4c26-9ee5-a22d1fe3bfa3

Co-authored-by: InAnYan <13097618+InAnYan@users.noreply.github.com>

* style: remove trailing blank line in AiSummaryShowingView

Agent-Logs-Url: https://github.com/InAnYan/jabref/sessions/a09f6212-dde3-4c26-9ee5-a22d1fe3bfa3

Co-authored-by: InAnYan <13097618+InAnYan@users.noreply.github.com>

* refactor(ai): add chat in memory cache

* refactor(ai): make compile

* refactor(ai): clean ups

* refactor(ai): fix migrations

* refactor(ai): quick fix for migrations

* refactor(ai): fix

* refactor(ai): fix migration

* refactor(ai): fix a bit the follow up questions

* refactor(ai): fix chat history scroll

* feat: redesign AI chat - move model to status window, add export there, move clear button right, add confirm dialog

Agent-Logs-Url: https://github.com/InAnYan/jabref/sessions/37c13a31-503d-4677-996a-8987a7aa7a22

Co-authored-by: InAnYan <13097618+InAnYan@users.noreply.github.com>

* refactor: extract formatChatModelLabel into named method in AiChatStatusView

Agent-Logs-Url: https://github.com/InAnYan/jabref/sessions/37c13a31-503d-4677-996a-8987a7aa7a22

Co-authored-by: InAnYan <13097618+InAnYan@users.noreply.github.com>

* refactor(ai): clean quickly

* refactor: move chat model building and export to AiChatStatusViewModel; bind chatHistory from AiChatViewModel

Agent-Logs-Url: https://github.com/InAnYan/jabref/sessions/07c35f08-0d20-4817-a568-88778f4e9c81

Co-authored-by: InAnYan <13097618+InAnYan@users.noreply.github.com>

* refactor(ai): add embedding model download cache

* refactor(ai): clear ingested documents on embedding model change

* refactor(ai): ui change

* refactor(ai): ui change for Ai Tab

* refactor(ai): restore regenerate message func

* refactor(ai): fix adr order

* refactor(ai): fix adr order

* refactor(ai): fix chatting requirements

* refactor(ai): change future feature

* refactor(ai): clean and fix file hasher tests

* refactor(ai): remove persited file ingestor test

* refactor(ai): refactor plain citation parsing with llm

* refactor(ai): less metadata changes

* refactor(ai): improve comment

* refactor(ai): refactor EmbeddingSimilarityMetric

* refactor(ai): quick modules change

* refactor(ai): remove ResolvedGroup

* refactor(ai): update AI docs

* refactor(ai): Rename PredefinedEmbeddingModel

* refactor(ai): change ChatMessage.Role

* refactor(ai): change ChatHistoryRecord

* refactor(ai): cleanups

* refactor(ai): remove ListenersHelper.java

* refactor(ai): remove CitationKeyCheck.java

* refactor(ai): fix AiPreferences

* refactor(ai): revert MVStoreBase

* refactor(ai): remove BibEntryListComparatorById

* refactor(ai): do not save if deleted

* refactor(ai): simplify chunked summarization logic

* refactor(ai): clean

* refactor(ai): clean

* refactor(ai): refactor

* refactor(ai): move exporters

* refactor(ai): static class refactors

* refactor(ai): remove comment

* refactor(ai): refactor answer engines

* refactor(ai): remove comment

* refactor(ai): cleanup + ingestion

* refactor(ai): cleanup

* refactor(ai): refactor ingestion + clean ups + follow-up questions

* refactor(ai): refactor bindings

* refactor(ai): simplify chat message

* refactor(ai): change UI ai chat

* refactor(ai): remove PropertiesHelper.java

* refactor(ai): block chat history during loading

* refactor(ai): refactor to use one status pane

* refactor(ai): move files + add tooltip

* refactor(ai): refactor bindings + cleanup

* refactor(ai): refactor ai default preferences

* refactor(ai): add API key changes in Preferences

* refactor(ai): convert javadoc to markdown

* Revert "refactor(ai): convert javadoc to markdown"

This reverts commit 5ed90c4.

* refactor(ai): make it run

* refactor(ai): fix + fix checkstyle

* refactor(ai): add tests

* refactor(ai): convert comments

* refactor(ai): docs

* refactor(ai): update ADR 0057

* refactor(ai): revert checkstyle

* refactor(ai): add MVStore comment

* refactor(ai): update docs

* refactor(ai): add testing resources

* refactor(ai): update from my review

* refactor(ai): add context menu for chat messages

* refactor(ai): remove assertions

* refactor(ai): just fixes

* refactor(ai): fix tracing

* refactor(ai): add CHANGELOG entry

* refactor(ai): add answer engine combobox

* docs(adr): use CUID2 for aiLibraryId

Add ADR-0059 documenting the choice of CUID2 over v4 UUID for the
`aiLibraryId` written into shared `.bib` files.

Co-authored-by: Ruslan <ruslanpopov1512@gmail.com>
Co-authored-by: ThiloteE <73715071+ThiloteE@users.noreply.github.com>
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

* refactor(ai): rename vBox/buttonsVBox in AiChatMessageView

Rename FXML-injected fields to describe their UI role rather than their
JavaFX type: `vBox` -> `bubble` (carries the chat-bubble styleClass),
`buttonsVBox` -> `buttons`. Matching `fx:id`s in `AiChatMessage.fxml`
updated.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

* undo

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

* refactor(ai): tidy AiChatMessageView

Use StringUtil.makeSafe for the null-to-empty content fallback, rename
the lambda parameter `v` to `role` in the role-comparison maps, and add
a short comment explaining the orElse(false) on the pseudo-class
bindings.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

* refactor(ai): localize FileStatus and use Directories.getUserDirectory

In AiChatStatusViewModel, give FileStatus a private displayName field
populated through the constructor with Localization.lang(...) at each
constant, matching the ChatMessage.Role pattern.

Replace Path.of(System.getProperty("user.home")) with
Directories.getUserDirectory() in both AiChatStatusViewModel
(exportMarkdown / exportJson) and AiSummaryShowingViewModel.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

* refactor(ai): collapse 2-value State enums to BooleanProperty

AiEntryChatViewModel and AiGroupChatViewModel both had a State enum
with only AI_TURNED_OFF and CHATTING, bound through BindingsHelper.bindEnum
on top of aiPreferences.enableAiProperty().not(). That is a boolean
dressed as an enum. Replace it with a BooleanProperty `enabled` bound
directly to aiPreferences.enableAiProperty(), and update the matching
view bindings to enabledProperty()/.not(). The multi-valued State enums
in AiChatViewModel and AiSummaryViewModel are left alone.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

* Fix HTML block

* reactor(ai): fix markdown

* reactor(ai): fix new lines at the end

* reactor(ai): fix from review

* reactor(ai): fix do while loop

* reactor(ai): fix new lines

* Fix submodules

* reactor(ai): fix code style

* reactor(ai): fix annotations

* reactor(ai): fix links in markdown

* reactor(ai): fix CHANGELOG

* reactor(ai): update module-info.java

* reactor(ai): apply openrewrite

* reactor(ai): migrate to jackson 3

* reactor(ai): fix links

* reactor(ai): fix code style

* reactor(ai): fix arch tests

* reactor(ai): fix arch tests x2

* reactor(ai): add check if a file is already ingested

* reactor(ai): fix chat migration

* reactor(ai): fix code style

* fix(docs/requirements): add expert settings AI requirements

* reactor(ai): fix module info

* fix: fix module info

* fix: fix module info

* fix: build gradle

* reactor(ai): fix localization

* reactor(ai): fix tests

* Update CHANGELOG.md

* reactor(ai): fix order of ADR

* reactor(ai): add new good to the adr

* reactor(ai): refine requirements

* reactor(ai): refine null check

* reactor(ai): add qodo comments

* reactor(ai): add qodo comments

* reactor(ai): fix formatting

* reactor(ai): fix from qodo

* reactor(ai): move AI adrs

---------

Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Co-authored-by: InAnYan <13097618+InAnYan@users.noreply.github.com>
Co-authored-by: Oliver Kopp <kopp.dev@gmail.com>
Co-authored-by: ThiloteE <73715071+ThiloteE@users.noreply.github.com>
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>