Streaming data from Neo4j to Gephi - "Invalid UTF-8 start byte 0xfc"

Viewed 405

I have recently started working with Neo4j and I am interested in visualizing my graph in Gephi. To do that, I am trying to use the apoc procedure

CALL apoc.gephi.add(null,'workspace1', paths) yield nodes, relationships, time
RETURN nodes, relationships, time

Some of my nodes have accented characters such as ö or å and that seems to be giving me trouble because I obtain the following error:

Neo.ClientError.Procedure.ProcedureCallFailed: Failed to invoke procedure 'apoc.gephi.add': Caused by: org.codehaus.jackson.JsonParseException: Invalid UTF-8 start byte 0xfc at [Source: apoc.export.util.CountingInputStream@599da1f9; line: 19, column: 125]  

On Gephi I obtain some of the nodes but not all of them or the relationships. This doesn't happen when I work with a database without special characters (like the movie database).

I am using

  • Windows 7
  • Neo4j Desktop 1.1.9
  • Neo4j 3.4.1
  • APOC 3.4.0.1
  • Gephi 0.9.2
  • Spanish locale

This is a similar problem to this one but it remains unsolved. Does anyone have any ideas?

2 Answers

It looks like this is an unresolved bug in the windows version of Gephi. So you will need to do one of the following

  • Run Gephi on linux inside a VM or Docker container
  • Remove the special characters from the data (I believe this is triggered by non ASCII characters like 'ä', 'Ä', 'ü', 'Ü', 'ö', 'Ö', 'ß')
  • Update Gephi to a newer version (if one is out)

I found a possible workaround while the issue is solved. If the graph from Neo4j is exported as a graphml and then imported on Gephi, there are no problem with the non ASCII characters.

Related