Compound list not generating an accurate graph

[djwetstede]

Describe the bug
When using a compound list as an input instead of genes, the knowledge graph is not generated correctly. It wrongly labels the input compounds as genes. Furthermore, code that is fully functional with genes will not work with compounds. If it is possible for a compound list to be generated into a graph, is there an example provided somewhere?

To Reproduce
Steps to reproduce the behavior:

1. Provide the Gene List used in the workflow

• The lists are provided in the notebooks below.

2. Provide the script/notebook where the error occur

• This is the notebook for genes, where everything is working correctly:
  https://github.com/djwetstede/pyBiodatafuse/blob/intact_annotator_wip/examples/test_intact_gene.ipynb
  -This is the notebook for compounds. The output of the annotator has the same output format as the genes, and uses the same function in generator.py:
  https://github.com/djwetstede/pyBiodatafuse/blob/intact_annotator_wip/examples/test_intact_compound.ipynb

4. State at which point the code is failing or resulting unexpected behavior.

• During the graph generation step, as the edges are wrongly created. Most likely due to the nodes not being correctly identified as genes.

5. Add any relevant information to reproduce the problemExpected behavior
   For the generation to work, there should probably be an input option for the user to specify the type of input. At the moment, it looks like it's always expecting genes.

The current intact generator function is also not really built around compounds, but even if it was, there's no current function to define input nodes as compounds.

Screenshots
If applicable, add screenshots to help explain your problem.

With genes as input:

Same function with 2 compounds input. They are linking to themselves, even though they should be linking to genes they are interacting with: