Better error handling in similarity search

[mnowotka]

This URLs:

https://www.ebi.ac.uk/chembl/api/data/similarity/TTZHDVOVKQGIBA-IQWMDFIBSA-N/100 https://www.ebi.ac.uk/chembl/api/data/similarity/CC(C)OC(=O)[C@H](C)N[P@](=O)(OC[C@H]1O[C@@H](N2C=CC(=O)NC2=O)[C@](C)(F)[C@@H]1O)Oc3ccccc3/100

generates 500 (details below), instead of some kind of 4XX, this should be fixed.

ORA-29902: error in executing ODCIIndexStart() routine 
ORA-20100: MDL-0279: Ctlib warning: [RDMFIL]RDMFIL:Invalid bond stereo, bond=10, stereo=6
MDL-0423: Unable to read molfile, error=3004 MDL-1620: Unable to read converted SMILES molfile
MDL-1051: Molstructure search query is not a valid molecule 
MDL-0976: Molecule index search initialization failed ORA-06512: 
"C$MDLICHEM62.MDL_MXIXMDL", line 329 ORA-06512: at
"C$MDLICHEM62.MDL_MXIXMDL", line 296

Details: http://sark.ebi.ac.uk:9000/chembl/chembl_new_webservices/group/91/

[mnowotka]

Fixed.