Please create a tool wiht:
tool inlinks.nt > backlinks.nt

Company A submits links to the repo and can then see which external sources are backlinking to it.

• remove from master
• in dev try to move as many as possible into the links folder

<#1> a void:Linkset ; dc:author <amit> ; dbp:ntriplefilelocation <http://downloads.dbpedia.org/2016-04/links/nytimes_links.ttl.bz2> ; void:objectsTarget <http://data.nytimes.com/> ; dct:description "nytimes information copied statically" .

1. extend description
2. test whether full external URI download works

shacl

• sh:xone only one way
• if two ways exist, take the best one, in general not the static

error reports

• should go to @holycrab13 via info.json
• via travis CI to people who send pull requests

the easiest is if validate repo produces the input data for Jan in json https://www.mkyong.com/java/json-simple-example-read-and-write-json/

I transfered this issue from the extraction framework github.
Originaly posted by @VladimirAlexiev

http://dbpedia.org/page/3M says that nyt:3M is
http://data.nytimes.com/N38115567920937854392.
However, that is long obsolete and doesn't resolve.

If you go to https://tools.wmflabs.org/sqid/#/browse?type=properties
and do "filter labels" by "new york" or "nyt", you'll see 5 new URL patterns that work.

• Eg for "New York Times Semantic Concept: Organization":
  http://api.nytimes.com/svc/semantic/v2/concept/name/nytd_org/3M%20Company.json?fields=all&api-key=
• The odd one amongst these 5 is "NYT topic ID":
  https://www.wikidata.org/wiki/Property:P3221

(Note: I wrote this in some notes Dec 2016, so don't remember what is "odd" about P3221, and have no clue how to go from N38115567920937854392 to nytd_org/3M Company)

https://github.com/dbpedia/links/blob/master/links/dbpedia.org/viaf.org/scripts/makeLinks.sh

• protect with .htaccess and .htpwd
• sync on download server
• Table should have at least two columns with linkset name and download link for backlinks
• optionally the table could also link to the linkset in the repo and links download server

generate a simple automatic overview of all the links that are available and how they are created

https://github.com/silk-framework/silk

Problem:
http://downloads.dbpedia.org/links/ has too many folders:
2016-10/ 10-Oct-2016 08:23 -
links/ 17-Oct-2016 17:32 -
tools/ 17-Oct-2016 17:32 -
types/ 17-Oct-2016 17:31 -
README.md

should only contain:
2016-10/ 10-Oct-2016 08:23 -

proposed solution:
clone repo in a non-public place, then link only the folder with the links vie ln -s

https://github.com/dbpedia/extraction-framework/blob/master/scripts/src/main/bash/process-geonames.txt

• find the error that produces 0 links (if metadata is wrong links can still be generated, this is a message for jan)
• remove static duplicates from the metadata.ttl ntriplefilelocation
• change order of execution in java-maven generatelinks

below is a list of improvements for linkviz.

• could you move the tool to links/tools/linkviz ?
• linkviz as github.io page

Above the table add (as a summary X links to Y datasets):

• number of total links in the latest release
• number of datasets

Add the following fields in the overview table

• Max links for every dataset

data for the tasks below

• @akirsche will provide the links to github and the metadata.ttl in a new data file

Improve the leftmost column "Linkset"

• Informative linkset names in the first, ideally rdfs:label from metadata.ttl, if label is missing use the current name and display a warning, that the label is missing
• link to github repo folder, displayed as edit pen
• Check for valid metadata.ttl
• for us it was unclear that the title can be clicked, so that the extra info opens

For each linkset chart

• Use a more compact bar style (see e.g. http://lodlaundromat.org/visualizations/ Datasets)

for each linkset info (on the right side)

• Add the content of metadata.ttl to the right pane,

• if any data is missing in metadata.ttl this can be displayed as a warning

• Fix error with firefox browser and d3js

Hi Robert,
can you run the cronjob weekly for a while?
so we can check whether it works properly.

http://downloads.dbpedia.org/links/2016-11-18/xxx.dbpedia.org/

this is a post processing steps that resolves redirects to the final target

MapObjectsUris / MapSubjectsUris (depending on where the DBpedia resource is)
https://github.com/dbpedia/extraction-framework/tree/master/scripts/src/main/scala/org/dbpedia/extraction/scripts

Pls tell

required @holycrab13

add warnings in GUI

• dct:description "logainm.ie is like a geonames for places within Ireland." .
• dct:license http://creativecommons.org/publicdomain/zero/1.0/ ;
• rdfs:label
• void:objectsTarget http://data.logainm.ie/place/ ;
• dc:author https://github.com/jimregan ;

also @akirsche

• check the technical things, exactly one method for contribution

We had a few syntax errors in some files that took time to identify & fix.
The idea is to create a TravisCI script that takes all rdf files in the repo and tries to validate them with rapper.
Travis has a new container based infrastructure were we can use normal bash scripts to do the validation
http://docs.travis-ci.com/user/migrating-from-legacy/

This approach will also validate all incoming pull requests

API which takes any URL and returns a canoical DBpedia / Wikidata URI

are some of the most important links

1. Go to http://dbpedia.org/page/Professor
2. Click on the Wikidata link
3. Error page appears

Feel free to copy this report to any more appropriate issue tracker.
Thanks!

https://github.com/dbpedia/extraction-framework/blob/master/scripts/src/main/scala/org/dbpedia/extraction/scripts/RecodeUris.scala

this needs to run as a post processing step

https://github.com/dbpedia/extraction-framework/blob/master/scripts/src/main/bash/process-yago.txt

• choose good examples for each contribution way and put in readme
• static ways should be marked as discouraged
• merge readme from tools/java-maven

• URIs for Linksets should start with a letter, there is an old Jena incompatibility and if the linksets are called <#1> the code will not work. They should rather be called
• scripts should always have a dedicated outputfile as parameter
• each linkset can have a combination of several ntriplefiles, scripts, sparqlqueries and linkConfs, however only one endpoint
• check whether sparql endpoint is active

https://github.com/dbpedia/links/wiki/How-To-Contribute-Links-to-DBpedia#guide