KG Catalog Init: DBLP

[kurzum]

Best way in my opinion is to let ChatGPT generate a python script. I already did one in bash:
https://chatgpt.com/share/68cac118-9188-8003-8fae-4bb09864d8ba

• produce a script that takes titel description databus URI from the yaml, then runs and uploads info to databus. try to make an initial prototype. I gave it the html, but it might also need a daatabus template
• write in the yaml that it is auto-generated. ignore the other stuff like sparqlendpoint, etc for now.

Milan's issue spec:

DBLP release cycle definition:

• frequency: once a month
• previous releases: not removed
• location: dump published at https://drops.dagstuhl.de/entities/collection/10.4230/dblp.rdf.ntriples
• dump form: single file as .nt.gz, e.g. https://drops.dagstuhl.de/storage/artifacts/dblp/rdf/2025/dblp-2025-10-01.nt.gz

Integrate DBLP in the knowledge graph catalog as follows:

• check regularly for new release
• if new release, then add to the catalog

[m1ci]

Implemented first version of auto-update for DBLP.

How it works?

• a github action (daily_check.yml) daily checks for new version of each KG
  • the action triggers a python script (daily_check.py) which iterates each KGs and if there is a release update script referenced in their KG metadata YAML, then it executes it
  • for DBLP there is an update script (update_dblp_yaml.py) and it is referenced in the dblp yaml metadata using the check-new-release field. (check-new-release: update_dblp_yaml.py)
  • if a new dblp version is found, the KG metadata (YAML) is updated and the changes are pushed
• another github action picks up the new release (due to YAML changes) and publishes it on the databus

In addition:

• a note is added to the version description that the this version release has been auto-generated. See https://databus.dbpedia.org/knowledge-graph-catalog/dblp/monthly-snapshot/2025-10-01

@kurzum please check

[kurzum]

Questions:

• did you generate update_dblp_yaml.py with AI?

Comments:

• shasum is missing -> 4.4GB is very small
• metadata.yml does not mention autogeneration. It would be partially locked for editing or overwritten, right? #######
• same as with yago, textual docu should be 50% better, i.e. not be minimal.
• update_dblp_yaml.py is technical not correct, it's if_checkNewVersion=true_then_update
  petra suggests these names:

Verb-based	check_and_update_dblp_yaml.py	Most natural, describes both actions clearly.
Conditional style	update_dblp_yaml_if_new.py	Explicit about the condition (nice middle ground).
Functional style	dblp_yaml_auto_update.py	Suggests automatic update behavior.
Event-based	dblp_yaml_on_new_version.py	Implies trigger or watcher style.

Otherwise seems ok.

[m1ci]

Thanks for the feedback.

Questions:

• did you generate update_dblp_yaml.py with AI?

Partially.

Comments:

• shasum is missing -> 4.4GB is very small

Fixed. provided correct shasum - manually at the moment.
As for the size. Fixed the size. Should be correct now. (But still 4-5 GB - thats the real size).
See https://databus.dbpedia.org/knowledge-graph-catalog/dblp

• metadata.yml does not mention autogeneration. It would be partially locked for editing or overwritten, right? #######

Added info in the description that this version has been autogenerated. See https://databus.dbpedia.org/knowledge-graph-catalog/dblp/monthly-snapshot/2025-10-01

I would still enable that person after auto-generation of the version the version metadata can still be edited manually.

• same as with yago, textual docu should be 50% better, i.e. not be minimal.

Improved textual description of the group (KG), artifact, and version. However an issue is that markdown seems not to work for group and artifacts descriptions. But works for versions.
Created an databus issue for this: dbpedia/databus#212

• update_dblp_yaml.py is technical not correct, it's if_checkNewVersion=true_then_update
  petra suggests these names:

Verb-based	check_and_update_dblp_yaml.py	Most natural, describes both actions clearly.
Conditional style	update_dblp_yaml_if_new.py	Explicit about the condition (nice middle ground).
Functional style	dblp_yaml_auto_update.py	Suggests automatic update behavior.
Event-based	dblp_yaml_on_new_version.py	Implies trigger or watcher style.

Renamed to dblp_release_auto_update.py

[m1ci]

@kurzum please if you have a look at DBLP and provide another round of feedback. Thanks