Semgraph Malformed JSON Output

🐛 Bug Reports

🌍 Environment

Your operating system and version: Ubuntu 18.04
Your FreeLing version: FreeLing 4.1

:memo: Bug Description

Given a FreeLing instance initialized with Mode=Server, Output Level=Semgraph and Output=JSON, at the time of generating the JSON Semgraph, a comma (',') is added for each paragraph of the document. In case the paragraph does not contain any sentence, no paragraph JSON object is added to the semgraph report, which implies that the aformentioned added comma malforms the JSON semgraph report.

Example:

Given a document with 3 paragraphs where the 2nd one has no sentences, this would be the resulting output:

{ "paragraphs" : [
   { "sentences" : [
      { "id":"1"}
     ]
    }
,
,
   { "sentences" : [
      { "id":"2"}
     ]
   }
]}

The comma (',') situated between the two sentences malforms the JSON.

💥 Reproducing

FreeLing Server initialization command: freeling/bin/analyze -f es.cfg --outlv semgraph --output json --nortk --flush --server --sense ukb --nec --numb --date --quant

TALP-UPC / FreeLing