Two equivalent SMILES encodings produce different similarity results

`C[n]1c(=O)ccc2cc(ccc12)Nc1[n]c([n]cc1Cl)N1C[C@H](C)C[C@H](C)C1`
`CN1C(=O)C=CC2=CC(=CC=C12)NC1=NC(=NC=C1Cl)N1C[C@H](C)C[C@H](C)C1`
inchi key `MYXSMUMMYLUNSW-OKILXGFUSA-N`

Depending on which SMILES string is used to initialize the molecule object, different similarity results are produced. This should be reproducible by producing a minimal bingo nosql db which includes the molecules listed below.

SMILES1 similarity results:
```
UVLOSTAZXZSNCP-UHFFFAOYSA-N (0.65)
ASOBSKNWRJNQBK-UHFFFAOYSA-N (0.65)
PETCVZZPKYJZAU-UHFFFAOYSA-N (0.63)
```

SMILES2 similarity results:
```
MUSYPBPUEKXCSS-UHFFFAOYSA-N (0.73)
JHZOJILVCGEIFB-UHFFFAOYSA-N (0.74)
ADSRAQFBQNFKGE-UHFFFAOYSA-N (0.73)
```

Indigo is configured with the following, using bingo nosql.
```
"ignore-stereochemistry-errors": True,
"standardize-charges": True,
"standardize-keep-largest": True,
"ignore-closing-bond-direction-mismatch": True,
"ignore-bad-valence": True,
"standardize-stereo": True,
"standardize-neutralize-zwitterions": True,
"standardize-clear-unusual-valences": True,
"similarity-type", "sim"
```