[Beta Test] Corese-GUI v5.0.0 Snapshot – Share your feedback!

[remiceres]

We are opening this discussion to gather user feedback on the latest snapshot of Corese-GUI v5.0.0. Your testing and bug reports are essential to stabilize the upcoming official release.

Thank you for your contribution to the Corese project.

Latest release: https://github.com/corese-stack/corese-gui/releases/tag/dev-prerelease

[FabienGandon]

First things first... this is great !

[MA-inria]

Merci pour tous qui a contribué à ce travail !!
Congrats!

[YoucTagh]

When launching the application, there is no feedback while it loads. One suggestion would be to have a splash screen with a loader while this is happening.

[remiceres]

Thank you very much for your feedback!

You are right, I have now added a splash screen. It is available in the latest pre-release, so feel free to test it and let me know what you think.

Please make sure to uninstall the previous version before installing this one. Since the version number has not changed, it may cause installation conflicts on some systems.

[MA-inria]

https://github.com/corese-stack/corese-gui/releases/tag/dev-prerelease

[YoucTagh]

When I loaded a graph with prefixes

@prefix owl: <http://www.w3.org/2002/07/owl#> .
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
@prefix ex: <http://www.example.org#> .

ex:alice a ex:Person;
  ex:firstName "Alice"@en.

ex:bob a ex:Person;
  ex:firstName "Bob"@en.

The generated graph didn't preserve the existing label ex, instead renaming it 'ns1'.

[remiceres]

Thank you for your feedback.

This is a known issue. The prefix renaming (e.g., ex becoming ns1) does not come from Corese-GUI itself but from Corese-Core’s current prefix handling during graph processing and serialization.

This behavior is already identified and should be addressed as part of the ongoing Corese-Core refactoring. The fix is therefore expected in an upcoming Corese-Core release.

For the moment, I prefer not to implement a temporary workaround in Corese-GUI, as this could introduce inconsistencies later. It is better to handle this properly at the Core level. I will discuss it further with the team to confirm the best approach.

[MA-inria]

OS: Windows
I have the old version installed on windows. when the installation is finished, the installer called "corese-gui" , but actually it launches the old "corese-gui", we have to make sure that the installer is calling corese-gui from the installation folder, not from the PATH variable

[remiceres]

Thank you for your feedback.

Unfortunately, I do not have full control over this behavior. The installation system has completely changed since the 4.x.x versions, and conflicts may occur if an older version is still present on the system (especially if it was installed using a different method).

For this reason, I strongly recommend fully uninstalling any previous 4.x.x version before installing Corese-GUI 5.x.x.

Depending on how Corese-GUI was installed, here are the recommended commands:

Windows

iwr -useb https://raw.githubusercontent.com/corese-stack/corese-gui-swing/develop/packaging/scripts/install-windows.ps1 | iex

macOS

curl -fsSL https://raw.githubusercontent.com/corese-stack/corese-gui-swing/develop/packaging/scripts/install-macos.sh -o /tmp/corese.sh && bash /tmp/corese.sh

Linux

curl -fsSL https://raw.githubusercontent.com/corese-stack/corese-gui-swing/develop/packaging/scripts/install-linux.sh -o /tmp/corese.sh && bash /tmp/corese.sh

Flatpak (if installed via Flatpak)

flatpak remove --delete-data --user fr.inria.corese.CoreseGui

[MA-inria]

Corese-GUI is running multiple instances, it is better to limit the number of instances to one, so that even if you run twice corese-gui, only a "Single Process" is allowed. I never tried it with jar, but with Windows applications it is fully supported.

[remiceres]

Thank you for the suggestion.

In the current design, multiple instances are intentionally allowed.
Corese-GUI does not rely on a global workspace or shared locked resources, and running multiple instances can actually be useful (e.g., loading different datasets side by side for comparison).

For now, I prefer to keep this behavior, as it provides more flexibility for advanced users.

[MA-inria]

Tag: enhancement
If we click the open button, and then select "Cancel" (we did not select any file)
The warning message appears, about the incorrect format.
Optimisation no warning should be shown

[remiceres]

Thank you for your feedback.

You are absolutely right, this warning message should not appear when simply clicking “Cancel” without selecting a file. I have corrected this behavior.

You can test the fix in the latest pre-release.

Please make sure to uninstall the previous version before installing the latest pre-release. Since the version number has not changed, it may cause installation conflicts on some systems.

[MA-inria]

Env: Windows, using installer
Problems loading a graph

• I am trying to open a .ttl file, it shows the number of nodes, but the graph area is empty, did I miss a step? I tried the "CONSTRUCT" query as well.
• When opening the same file twice, the number of nodes increments by 1 each time 5347,5348,5349, 5350 and so on
• Is there a way to see the list of loaded RDF files? I can see in the status bar the number of sources only, but cannot see exactly what they are.

After doing more tests, the problem seems happening with this file, If I close the program, and reopen it, then try with another file, it works.
While If I try to open a working file, after trying with the problematic one, nothing is show, the problematic file make the software enters in an unstable state
I cannot attach the .ttl here, here is the link :
https://minio-storage.apps.confianceai-public.irtsysx.fr/etaia-storage/documents_delivered/ontology_confiance.ttl

[MA-inria]

By checking the logs , I can see :

2026-02-25 14:14:45.654 [Corese-Worker-2] INFO  f.i.c.g.core.service.RdfDataService - Loading RDF file: C:\Users\userame\Downloads\ontology_confiance.ttl
2026-02-25 14:14:46.046 [Corese-Worker-2] INFO  f.i.c.g.core.service.RdfDataService - Successfully loaded 5347 triples from file.
2026-02-25 14:14:46.524 [Corese-Worker-2] WARN  f.i.c.g.c.s.GraphProjectionService - Sanitized 28933 malformed JSON-LD quoted value(s) in graph snapshot (double-quoted: 1, literal-inner-quotes: 28932).
2026-02-25 14:14:46.632 [JavaFX Application Thread] ERROR f.i.c.g.c.graph.GraphDisplayWidget - [JS] Graph drawing error: | JSON Parse error: Invalid escape character :
parse@[native code]

N.B. The GUI shows the success dialog when opening the file

[remiceres]

Hello,

Thank you for the detailed report (and for sharing the file/link), this was very helpful.

1) Graph area empty while nodes are counted

You did not miss any step. This was a UI rendering bug in the graph visualization layer (more precisely in the transformation between the loaded RDF data and the D3-based visual graph).

The data was correctly loaded and could be queried, but the visual graph failed to render in some cases, which made the graph area appear empty.

I have fixed this issue. The graph should now display correctly in the latest pre-release.

2) Count increases when opening the same file multiple times

What you see is expected in this case.

Corese stores RDF as a set, so exact duplicate triples are ignored when reloading the same file.

However, your file contains a blank node (anonymous node) in the ontology declaration.
Blank nodes get a new internal ID at each import, so triples involving that blank node are considered new.

That is why your counter increases by only +1 each time (5347 → 5348 → 5349), not by the full file size.

3) Viewing the list of loaded RDF sources

Currently, the status bar only displays the number of sources.

However, you can see the detailed list of loaded files/URLs in two ways:

• Go to the Logs page (4th button with the “i” icon in the sidebar).
• In the Data → Graph view, click the Reload button (arrow icon), which lists all currently loaded sources (files and URIs).

Thank you for the suggestion. I’ve noted it for possible future improvements.

---

Thank you again for your thorough testing and detailed feedback.

Please make sure to uninstall the previous version before installing the latest pre-release. Since the version number has not changed, it may cause installation conflicts on some systems.

[MA-inria]

tag: enhancement

• If you enter the Settings menu, you cannot exit from it by clicking outside, you should click other menus. suggestion: adding a close/collapse button, or closing the settings panel when clicking outside of it ,
• The cursor shape is the same (hand), whether you are selecting a node or "panning" the graph. suggestion using different cursors for paning or selecting, to give the user a visual feedback, whether he is on a blank area, or on a node because at some zoom level, you think you are clicking the node (to drag it), but actually your clicking outside, and vice versa.
• Addin the possibility to focus on the "node card", in order to copy some of the text

**Questions: **

• In the windows installer version, where is the default log files location ?
• Where can I see the list of loaded RDF files ?

[remiceres]

Hello,

Thank you for your feedback.

Settings panel

I’m not sure I fully understood the point, so please feel free to reply to this comment if I misunderstood your suggestion.

In the current design, Settings is implemented as a full page, at the same hierarchical level as the other pages (Data, Query, Validation, etc.). It is not displayed as an overlay on top of the other pages.

For that reason, we intentionally did not add a close button or the ability to dismiss it by clicking outside, as that would not really fit with the page-based navigation model of the application.

Cursor feedback when panning / selecting nodes

Good suggestion, I have implemented this change.

Now:

• When the mouse is over the background, the cursor appears as a directional cross, indicating that the graph can be panned.
• When the mouse is over a node, the cursor changes to a grab hand, indicating that the node can be dragged.

This should make the interaction clearer when navigating the graph.

Selecting text from the node card

Good idea as well, this has now been implemented.

You can test it in the latest pre-release.
Simply click on a node, and the node card will remain visible so that its content can be selected and copied.

Questions

Where are the log files located (Windows installer)?

In the current installer version, logs were written to:

C:\Users\user\logs

This location was not very explicit, so I have changed it.
Starting from the latest pre-release, runtime files are stored in standard OS locations.

Runtime Files

Corese GUI stores logs and preferences in user-writable OS locations.

Windows

• Logs: %LOCALAPPDATA%\Corese GUI\logs\
• Preferences: %APPDATA%\Corese GUI\preferences\preferences.properties

macOS

• Logs: ~/Library/Logs/Corese GUI/
• Preferences: ~/Library/Application Support/Corese GUI/preferences/preferences.properties

Linux

• Logs: $XDG_STATE_HOME/corese-gui/logs/
  (fallback: ~/.local/state/corese-gui/logs/)
• Preferences: $XDG_CONFIG_HOME/corese-gui/preferences/preferences.properties
  (fallback: ~/.config/corese-gui/preferences/preferences.properties)

Where can I see the list of loaded RDF files?

You can access the list of loaded sources in two ways:

1. Go to the Logs page (4th button with the “i” icon in the sidebar).
2. In Data → Graph view, click the Reload button (arrow icon), which lists all currently loaded sources (files and URIs).

Please make sure to uninstall the previous version before installing the latest pre-release. Since the version number has not changed, it may cause installation conflicts on some systems.

[FabienGandon]

Some feedback here:
Very nice new version

• I was surprised by the fact that “Data>Reload” deactivates “RDFS” if it has been activated.
• Since the Validation has RDF as output, we could also have a “Graph” tab in the validation to show that graph
• It would be nice to have a way to duplicate a query tab to start a new query from another one
• It would be nice to be able to change the name / to be able save to another file
• In the table for literals should we display its content or the literal and type/language ?
  Bla bla bla
  or
  “Bla bla bla”, “Bla bla bla”@fr , “bla bla”^^xsdy:int …
• When I open many query tabs there is no arrows to scroll back to the first tabs

[MA-inria]

For the query tabs, the good thing is that currently we can move to the first tab by "click and drag".
Even a horizontal "mouse scroll" is working. until implementing the arrows.

[MA-inria]

[remiceres]

Hello Fabien,

Thank you for your feedback.

I was surprised by the fact that “Data > Reload” deactivates “RDFS” if it has been activated.

I agree with you on this point. I have modified the behavior so that “Data > Reload” no longer automatically deactivates the currently enabled inference rules.

Since the Validation has RDF as output, we could also have a “Graph” tab in the validation to show that graph.

You are right. Since SHACL validation produces RDF as output, I have now enabled the Graph tab for validation results as well.

It would be nice to have a way to duplicate a query tab to start a new query from another one.

Thank you for the suggestion. I noted the idea, but I prefer not to add it in this version for now. We can consider it for a future release.

It would be nice to be able to change the name / to be able save to another file.

This was already discussed with Aline. We can discuss it together again to decide the best approach. I have noted the suggestion and we can revisit it for a future version.

In the table for literals should we display its content or the literal and type/language?

I agree that it is better to display the full information. I have therefore modified it so that literals are now displayed with their datatype or language tag.

When I open many query tabs there is no arrows to scroll back to the first tabs.

Currently the interface provides horizontal scrolling with a shadow effect to indicate that additional tabs are available. For now I think this is sufficient, but we could consider adding arrows in the future if more users request it.

[gmontenegrou]

Environment: Windows (installer version)

First of all — it’s amazing! Great job on this new version 👏

Here is some feedback after testing:

1. After installing the new version, both the old and the new versions appeared on my desktop. The only visible difference between them is the name:
   Old version: Corese-GUI
   New version: corese-gui

2. The adaptive view worked well with a graph containing 452 triples.

However, when loading a knowledge graph with 4160 triples, the adaptive view displayed no labels, and the entire Visualization tab became unstable / buggy (which is comprehensible)

This suggests possible scalability or rendering issues when handling larger graphs.

4. I really liked the new export feature in the query tab. In particular the default file names that are automatically generated .

5. I tested SHACL validation with 16 defined shapes in my larger KG. After more than 30 minutes, the validation was still running. I could not find any way to abort or stop the validation process. No related logs were displayed in the System Logs tab. Because of this, I could not determine whether: The shapes were invalid, or the process was simply taking a very long time. I had to close and reopen the application to continue testing. It would be helpful to: Add a “Stop” or “Cancel” option for long-running validations.

Thanks !!

[remiceres]

Hello Génesis,

Thank you for your feedback and for the kind comment, I’m really glad you like the new version.

Installation / application name

Unfortunately, the installation system changed completely between versions prior to 5.0.0 and the new 5.0.0 release. Because of this, the new installer cannot automatically detect or uninstall older versions.

As a result, the previous version must be uninstalled manually.

I also noticed the naming inconsistency you mentioned. I have corrected this: the application is now named Corese-GUI again (instead of corese-gui) in the latest build.

You can test this in the latest pre-release:
https://github.com/corese-stack/corese-gui/releases/tag/dev-prerelease

If you want to test the latest pre-release, I recommend uninstalling the current 5.0.0 snapshot first. Since the version number has not changed, keeping both installed may cause conflicts.

Adaptive graph view

For the graph visualization, I implemented a progressive degradation strategy based on the number of triples. The goal is to keep the interface responsive while still allowing users to visualize the graph.

When the graph is small, all visualization features are enabled:

• zoom
• panning
• tooltips
• node labels
• edge labels
• ...

As the graph grows larger, some features are progressively disabled in order to keep the visualization smooth and avoid slowing down the application.

For a graph of around 4160 triples, some of these optimizations are therefore activated (for example labels may be hidden).

Beyond a certain threshold, the graph visualization is completely frozen. This is intentional: it allows the graph to still be displayed without degrading the overall application performance. In practice, graphs of that size are usually difficult to explore visually anyway.

For very large graphs, the visualization may even be fully paused, ensuring that the rest of the application remains responsive.

However, you pointed out an important issue: the feedback provided to the user about these optimizations was not very clear. I have therefore improved this by:

• Fix the graph status indicator
• Improving tooltips explaining which optimizations are currently active

This improvement is included in the latest pre-release.

SHACL validation

The SHACL performance issue is related to Corese-Core, and unfortunately it is not something I can easily fix at the GUI level. Performance should improve significantly in the future with the ongoing refactoring of Corese-Core.

However, you are absolutely right about the missing ability to interrupt a validation.

This functionality already existed for SPARQL queries but it was missing for SHACL validation I have now added a *Stop/Cancel mechanism for SHACL validations as well.

This is also available in the latest pre-release.

Thanks again for your detailed feedback.

[MaillPierre]

Enhancement:

• Syntax check in the SPARQL query editor.
  I try to send a wrong CONSTRUCT query:

CONSTRUCT {
?s ?p ?o
} LIMIT 10

but the correct version should have been:

CONSTRUCT WHERE {
?s ?p ?o
} LIMIT 10

A warning before submission would be nice if query parsing fails, before clicking "send", as Yasgui does:

[remiceres]

Hello Pierre,

Thank you for your feedback.

I completely agree that having syntax checking directly in the SPARQL editor would be very helpful.

The idea is noted, and this is something we plan to integrate in a future version.

[MaillPierre]

Bug, error while loading file via URL.
Tried to load https://github.com/Wimmics/IndeGx/raw/refs/heads/endpoint_status/catalogs/catalog.latest-status.ttl .
Got the following message:

fr.inria.corese.gui.core.service.RdfDataService$RdfLoadException: Failed to load RDF URI 'https://github.com/Wimmics/IndeGx/raw/refs/heads/endpoint_status/catalogs/catalog.latest-status.ttl': (internal_error) Unexpected error: java.security.InvalidAlgorithmParameterException: the trustAnchors parameter must be non-empty
	at fr.inria.corese.gui.core.service.RdfDataService.loadUri(RdfDataService.java:177)
	at fr.inria.corese.gui.core.service.data.DefaultDataWorkspaceService.loadUri(DefaultDataWorkspaceService.java:71)
	at fr.inria.corese.gui.feature.data.support.DataLoadingSupport.loadUris(DataLoadingSupport.java:59)
	at fr.inria.corese.gui.feature.data.DataViewController.lambda$executeLoadUris$0(DataViewController.java:468)
	at fr.inria.corese.gui.feature.data.DataViewController.lambda$runAsyncOperationWithLoading$0(DataViewController.java:711)
	at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1090)
	at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:614)
	at java.base/java.lang.Thread.run(Thread.java:1474)
Caused by: javax.net.ssl.SSLException: (internal_error) Unexpected error: java.security.InvalidAlgorithmParameterException: the trustAnchors parameter must be non-empty
	at java.base/sun.security.ssl.Alert.createSSLException(Alert.java:132)
	at java.base/sun.security.ssl.TransportContext.fatal(TransportContext.java:376)
	at java.base/sun.security.ssl.TransportContext.fatal(TransportContext.java:319)
	at java.base/sun.security.ssl.TransportContext.fatal(TransportContext.java:314)
	at java.base/sun.security.ssl.SSLSocketImpl.handleException(SSLSocketImpl.java:1710)
	at java.base/sun.security.ssl.SSLSocketImpl.startHandshake(SSLSocketImpl.java:470)
	at java.base/sun.security.ssl.SSLSocketImpl.startHandshake(SSLSocketImpl.java:426)
	at java.base/sun.net.www.protocol.https.HttpsClient.afterConnect(HttpsClient.java:483)
	at java.base/sun.net.www.protocol.https.AbstractDelegateHttpsURLConnection.connect(AbstractDelegateHttpsURLConnection.java:187)
	at java.base/sun.net.www.protocol.http.HttpURLConnection.getInputStream0(HttpURLConnection.java:1377)
	at java.base/sun.net.www.protocol.http.HttpURLConnection.getInputStream(HttpURLConnection.java:1302)
	at java.base/sun.net.www.protocol.https.HttpsURLConnectionImpl.getInputStream(HttpsURLConnectionImpl.java:223)
	at fr.inria.corese.gui.core.service.RdfDataService.openUriStreamInternal(RdfDataService.java:209)
	at fr.inria.corese.gui.core.service.RdfDataService.openUriStream(RdfDataService.java:192)
	at fr.inria.corese.gui.core.service.RdfDataService.lambda$loadUri$0(RdfDataService.java:167)
	at fr.inria.corese.gui.core.service.RdfDataService.parseWithRetry(RdfDataService.java:228)
	at fr.inria.corese.gui.core.service.RdfDataService.loadUri(RdfDataService.java:167)
	... 7 more
Caused by: java.lang.RuntimeException: Unexpected error: java.security.InvalidAlgorithmParameterException: the trustAnchors parameter must be non-empty
	at java.base/sun.security.validator.PKIXValidator.<init>(PKIXValidator.java:96)
	at java.base/sun.security.validator.Validator.getInstance(Validator.java:173)
	at java.base/sun.security.ssl.X509TrustManagerImpl.getValidator(X509TrustManagerImpl.java:308)
	at java.base/sun.security.ssl.X509TrustManagerImpl.checkTrustedInit(X509TrustManagerImpl.java:183)
	at java.base/sun.security.ssl.X509TrustManagerImpl.checkTrusted(X509TrustManagerImpl.java:198)
	at java.base/sun.security.ssl.X509TrustManagerImpl.checkServerTrusted(X509TrustManagerImpl.java:132)
	at java.base/sun.security.ssl.CertificateMessage$T13CertificateConsumer.checkServerCerts(CertificateMessage.java:1279)
	at java.base/sun.security.ssl.CertificateMessage$T13CertificateConsumer.onConsumeCertificate(CertificateMessage.java:1172)
	at java.base/sun.security.ssl.CertificateMessage$T13CertificateConsumer.consume(CertificateMessage.java:1115)
	at java.base/sun.security.ssl.SSLHandshake.consume(SSLHandshake.java:421)
	at java.base/sun.security.ssl.HandshakeContext.dispatch(HandshakeContext.java:477)
	at java.base/sun.security.ssl.HandshakeContext.dispatch(HandshakeContext.java:448)
	at java.base/sun.security.ssl.TransportContext.dispatch(TransportContext.java:199)
	at java.base/sun.security.ssl.SSLTransport.decode(SSLTransport.java:172)
	at java.base/sun.security.ssl.SSLSocketImpl.decode(SSLSocketImpl.java:1506)
	at java.base/sun.security.ssl.SSLSocketImpl.readHandshakeRecord(SSLSocketImpl.java:1421)
	at java.base/sun.security.ssl.SSLSocketImpl.startHandshake(SSLSocketImpl.java:455)
	... 18 more
Caused by: java.security.InvalidAlgorithmParameterException: the trustAnchors parameter must be non-empty
	at java.base/java.security.cert.PKIXParameters.setTrustAnchors(PKIXParameters.java:202)
	at java.base/java.security.cert.PKIXParameters.<init>(PKIXParameters.java:121)
	at java.base/java.security.cert.PKIXBuilderParameters.<init>(PKIXBuilderParameters.java:105)
	at java.base/sun.security.validator.PKIXValidator.<init>(PKIXValidator.java:93)
	... 34 more

Tried to use corese-command on the same URL and got no problem

[remiceres]

Hello Pierre,

Thank you for your feedback.

I was not able to reproduce the issue on my side. It may have already been fixed during one of the recent pre-releases.

Could you let me know whether you still encounter the problem with the latest pre-release?

[z-arnaud]

There is a button missing to view / remove existing .ttl files in the Data tab. We can only reload, which does not work for removing, particularly if there is only one .ttl loaded.

[remiceres]

Hello,

Thank you for your feedback.

I am not sure I fully understood what you were referring to. If my answer does not address your point, please feel free to reply in this thread.

The fact that it is not possible to manage TTL files individually (for example reloading or removing only one file) is actually a deliberate design choice.

The reason is related to how Corese loads data. When a file is loaded, the triples it contains are inserted into an in-memory graph representation. Once loaded, the triples are no longer linked to their original file. If multiple files are loaded, all triples are stored together in the same graph.

Because of this, it is not possible to know which triples came from which file after loading, and therefore it is not possible to manage files individually at that level.

The only way to remove a file is therefore to rebuild the graph. This is exactly what the current "Reload" action does: it clears the graph and reloads the selected files, while also reactivating the previously enabled rules or profiles.

For example, if you want to remove one file among several loaded files, you can simply uncheck that file and press Reload. The graph will then be rebuilt without that file.

If only one file is loaded and you want to remove it, the correct action is to use the "Clean" button rather than "Reload".

I hope this clarifies the current behavior.

[Sensaku]

Overall feeling: The GUI is a bit of fresh air and sounds promising.

• Export features work great (except de namespace issue related before that leads to 36K lines of header).
• I can't use the graph visualiser because it breaks at 50K nodes and is paused forever. I think this should be disabled because it consume a lot of ram even if the visualisation breaks.
• I found the local loading slower, maybe related to the visualisation.
• It would be nice to have a button to show all result without paging instead of putting a large number to cover all result.

I have small things and bug to report unfortunatly.
tag: bug

• Query:
  When loading multiple .rq file, the focus is always the first file loaded but the execute button is anchored to the last file loaded. That leads to wrong display, wrong result.
  Note: clicking back to any tab solve the execute button.
• Loading RDF graph locally:

• 164Mo of turtle (2M tiple) is enough to break the graph visualization and increase loading time exponentially (10 minutes to have not display at all).
• 4Go of turtle breaks the GUI.
• When the graph view is paused, we can't launch it again.

[remiceres]

Hello Arnaud,

Thank you for the detailed feedback and for the encouraging words.

Export features work great (except the namespace issue related before that leads to 36K lines of header).

Yes, this is related to the prefix handling issue previously mentioned. The current behavior comes from Corese-Core serialization and should be improved as part of the ongoing Corese-Core refactoring.

Previous response on this topic:
https://github.com/orgs/corese-stack/discussions/280#discussioncomment-15984662

I can't use the graph visualiser because it breaks at 50K nodes and is paused forever.

This behavior is intentional. The graph visualizer includes several protection mechanisms to prevent the UI from freezing when the graph becomes too large. When the graph exceeds certain thresholds, the visualization is progressively degraded and may eventually be paused entirely.

You can find more details in this previous reply:
https://github.com/orgs/corese-stack/discussions/280#discussioncomment-16070409

I found the local loading slower, maybe related to the visualisation.

I need to run more detailed tests to determine whether there is a measurable difference.

What takes time is mainly the graph rendering and the computation of graph statistics. However, this process runs in a separate background thread. As soon as the "data loaded" notification appears, Corese is fully functional even if the graph rendering is still ongoing. At that point the data can already be queried using SPARQL or validated with SHACL.

If the graph visualization is not needed, another option is to change the Graph Preview Limit in the settings. This allows you to disable the automatic graph rendering entirely.

It would be nice to have a button to show all results without paging.

Good suggestion. I have noted this idea for a possible improvement of the results table.

For now, the workaround is to set a large number of results per page. Note that this value is saved between tabs and across sessions, so you only need to set it once.

---

Bugs

When loading multiple .rq files, the focus is always the first file loaded but the execute button is anchored to the last file loaded.

Thank you for reporting this. This was a UI synchronization issue between the active editor and the execution context.

I have fixed the problem in the latest pre-release.

Please make sure to uninstall the previous snapshot before installing the new one, since the version number has not changed and keeping both versions may cause conflicts.

164 MB of Turtle (2M triples) breaks the graph visualization and increases loading time.

Yes, this is consistent with the current limitations of the visualization layer. The graph viewer is not designed to handle millions of triples directly. In such cases the visualization will either degrade heavily or pause entirely.

4 GB of Turtle breaks the GUI.

Corese-Core currently uses an in-memory storage model and is therefore not designed to handle datasets of that size.

When the graph view is paused, we can't launch it again.

When the graph exceeds a certain size, the visualization is intentionally disabled. At that scale the visualization would not be usable anyway and could significantly slow down or crash the application.

Thanks again for your thorough testing and detailed report.

[remiceres]

Hello everyone,

Thank you all for your feedback, testing, and detailed reports. Your contributions were extremely helpful in improving and stabilizing this new version.

Unless new issues appear, the official release should arrive in the next few days.

Thanks again for your help and support!