Patrice Lopez

Hi @rodyoukai ! Thanks for the issue, I don't see anything wrong in the training examples for URL and ISSN. ```xml Azaola, Elena (2009). El comercio con el dolor y...

Hello ! The encoding of the results follows the TEI, so URL are encoded like this by definition: ``` ``` `` has no type, and target URL is defined by...

> In your example (above) type parameter exists in idno tag... What is your input reference? With ISSN keyword (e.g. `ISSN: 1390-3691.`) it works normally, but without (`1390-3691.`), it's just...

With this input reference, the type ISSN appears with the current system (https://grobid.science-miner.com). In your training data, did you add systematically the ISSN prefix in the `` field, for example:...

In this example the `ISSN:` is outside the `` mark-up? Should be: ```xml Vargas Reyes, Bryan, Ariza Santamaría, Rosembert (2020). Liberación de la madre tierra: entre la legitimidad y los...

Hi @elonzh ! Thanks for the case! Is it open access and CC-BY? I think the reference segmenter model would need first a decent amount of training data with more...

Dear @chrismattmann and @sujen1412 Sorry to bother, I am trying to get things alright to publish the java artefacts in maven central via Gradle and so on. Could you add...

Thank you @sujen1412 !!

... set-up everything in branch https://github.com/kermitt2/grobid/tree/maven-central This is incredibly time consuming... There are also different URLs to upload stuff either as release or snapshot, not sure how to manage that...

Hi @de-code ! Are you using CrossRef or biblio-glutton as consolidation? You can have a look at https://github.com/kermitt2/biblio-glutton#matching-accuracy for a study of consolidation accuracy (evaluation on 17.015 bibliographical references/DOI pairs)....