On Dec 18, 2016, at 6:32 PM, Brian Kelley wrote:
> >>> m.GetProp("_smilesAtomOutputOrder")
> '[3,2,1,0,]'
>
> Note that this returns the list as a string which is sub-optimal.
> GetPropsAsDict will convert these to proper python objects, however, this is
> considered a private member so you need to return these as well:
>
> >>> list(m.GetPropsAsDict(True,True)["_smilesAtomOutputOrder"])
> [3, 2, 1, 0]
For fun, here are a few timing numbers:
# Common setup
from rdkit import Chem
mol = Chem.MolFromSmiles("c1ccccc1Oc1ccccc1")
Chem.MolToSmiles(mol)'
import json
import ujson # third-party JSON decoder
import re
integer_pat = re.compile("[0-9]+")
# Get the string (give a lower bound)
mol.GetProp("_smilesAtomOutputOrder")'
10000 loops, best of 3: 31.3 usec per loop
Here are variations for how to get that information as a list of integers:
# Using Python's "eval()" to decode the list (this is generally UNSAFE!)
eval(mol.GetProp("_smilesAtomOutputOrder"))'
10000 loops, best of 3: 157 usec per loop
# Use the built-in json module (need to remove the terminal ",")
json.loads(mol.GetProp("_smilesAtomOutputOrder")[:-2]+"]")'
10000 loops, best of 3: 66.5 usec per loop
# Use the third-party "ujson" package, which is faster than json.
ujson.loads(mol.GetProp("_smilesAtomOutputOrder")[:-2]+"]")
10000 loops, best of 3: 41.2 usec per loop
("cjson" takes 49.7 usec per loop)
# Use the properties dictionary
mol.GetPropsAsDict(True,True)["_smilesAtomOutputOrder"]
1000 loops, best of 3: 462 usec per loop
# Parse it more directly
map(int, integer_pat.findall(mol.GetProp("_smilesAtomOutputOrder")))
10000 loops, best of 3: 89 usec per loop
Andrew
da...@dalkescientific.com
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
_______________________________________________
Rdkit-discuss mailing list
Rdkit-discuss@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/rdkit-discuss
