similarity search problem & bitstring to string conversion

Hi,

a:
I used rdkit nodes for calculation of fingerprints which I fed to similarity search node.
It looks like the names of the compared columns in both tables has to be the same, strange but OK. But if they are the same  and I can finally run the node but it will give error:

Execute failed: Argument contains duplicates: [FP_RDKit_Morgan, FP_RDKit_Morgan]

What is the problem?

b: is it possible to convert bit string to string and save it into csv? or export it into R dataframe or pandas?

Please help. mx

 

If columns had different names you would need to map each column from the two tables. If you are doing Tanimoto you can use a fingerprint column so the mapping would be easy, but this would need a different implementation of the column handling for different similarity measures. 

I just ran the RDKit fingerprint node, then connected the output to both data table input ports on the SImilarity Search node and it works fine. Can you share your workflow? 

 

With regards to R it looks like the BitSet is passed as a series of integer values. You could save the String representation of the fingerprint to a CSV file...

 

Cheers

Sam

If columns had different names you would need to map each column from the two tables. If you are doing Tanimoto you can use a fingerprint column so the mapping would be easy, but this would need a different implementation of the column handling for different similarity measures. 

I just ran the RDKit fingerprint node, then connected the output to both data table input ports on the SImilarity Search node and it works fine. Can you share your workflow? 

 

With regards to R it looks like the BitSet is passed as a series of integer values. You could save the String representation of the fingerprint to a CSV file...

 

Cheers

Sam

Hi,

I have the same problem. "Execute failed: Argument contains duplicates: [Circular fingerprints, Circular fingerprints]"

I made sure to remove duplicates with groupby and I get this message anyway.

Is there a bug?

Thank you.

Julien