Re: [Rdkit-discuss] Having trouble getting RDKIT to recognize LiAsF6

2017-11-25 Thread Chris Earnshaw 
Hi Yoolhee

If this is something you need fixed *now*, you have access to the RDKit
source code and you're confident to build a new version of RDKit, then try
the following.

1. from the root folder of the RDkit distribution, find the folder
Code/GraphMol/
2. in that folder, copy atomic_data.cpp to atomic_data.cpp.bak (so you can
restore things if necessary)
3. edit atomic_data.cpp as follows
  - find the line for phosphorus that starts '15   P' (probably line 42
in the file)
  - note that the line ends '3   5   7 \n \'
  - find the line for arsenic, starting '33  As' (probably line 60)
  - this line ends '3  5 \n \'
  - change the line by inserting 7 immediately after the 5 (that's a
tab character)
  - the end of the arsenic line should now look like the end of the
phosphorus line
  - save the file
4. rebuild RDKit

I wouldn't recommend trying this if you haven't previously built RDKit from
source.


@Greg - is it worth adding higher valence states for the halogens as well
while fixing As and Sb? At the moment we have
Cl  1
Br  1
I   1  2  5
Shouldn't these all be '1  3  5  7' ?

Best regards,
Chris

On 24 November 2017 at 21:31, Yoolhee Kim  wrote:

> Chris,
>
> Thank you very much for your reply! I'm not very familiar with RDKIT, and
> I was wondering if you could elaborate how to fix the problem of adding '7'
> in the As entry so that the valence state is recognized.
>
> Best regards,
> Yoolhee Kim
>
> On Tue, Nov 21, 2017 at 8:22 AM, Chris Earnshaw 
> wrote:
>
>> Hi
>>
>> The entries for P and As in RDKit's atomic_data.cpp are -
>> 15  P   0.750.892.0830.974  5   31
>> 30.97376163 3   5   7
>> 33  As  1.211.2 1.8574.922  5   75
>> 74.9215965   3   5
>>
>> So the required 'connectivities' for PF6- are present in the file
>> (final 3 values, '7' being the relevant one here), but not for As. I'm
>> not quite sure why this should be so as the phosphorus is P(V) in PF6,
>> but this appears to be the cause. You'll have to add the '7' in the As
>> entry and rebuild RDKit to allow AsF6- to be processed. A similar fix
>> would be needed for antimony too. Someone please let me know if
>> there's a more convenient solution!
>>
>> There are similar issues for the halogens as well. Only iodine has
>> values > 1, so by default it's not possible to construct e.g.
>> chlorates, or bromates, and no perhalates are allowed.
>>
>> Regards,
>> Chris Earnshaw
>>
>> On 20 November 2017 at 23:03, Yoolhee Kim 
>> wrote:
>> > Hello,
>> >
>> > I'm trying to get RDKIT to recognize LiAsF6 from the SMILES Formula
>> > "[Li+].F[As-](F)(F)(F)(F)F" but it seems that RDKIT will not read it
>> > correctly and is unable to create a molecule object.
>> >
>> > RDKIT is able to understand LiPF6 from the SMILES formula
>> > "[Li+].F[P-](F)(F)(F)(F)F" so I tried using that and replacing P with As
>> > using both replaceatom and replacesubstruct functions, but that had some
>> > issues as well.
>> >
>> > Has anyone faced a similar issue or have any suggestions?
>> >
>> > Thanks!
>> >
>> > -Yoolhee
>> >
>> > 
>> --
>> > 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
>> >
>>
>
>
>
> --
> Yoolhee Kim
> M.S. in Energy Science Technology and Policy
> Carnegie Mellon University
> (714) 326-5820 | yoolh...@andrew.cmu.edu
>
--
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

Re: [Rdkit-discuss] Having trouble getting RDKIT to recognize LiAsF6

2017-11-24 Thread Greg Landrum 
On Fri, Nov 24, 2017 at 10:32 PM, Yoolhee Kim 
wrote:

>
> Thank you for your reply!
>
> I'm not super familiar with RDKIT... could you please elaborate on how to
> fix this problem?
>
>
Ah, sorry, it's something that needs to be fixed in the C++ code. The bug
report is filed:
https://github.com/rdkit/rdkit/issues/1668
and a fix is on the way.

-greg
--
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

Re: [Rdkit-discuss] Having trouble getting RDKIT to recognize LiAsF6

2017-11-22 Thread Greg Landrum 
This is happening because the code doesn't recognize 7 as a valid valence
state of As.
This is, as you note, inconsistent with how P is handled, which doesn't
make a lot of sense.

Fortunately, it's a fixable problem. :-)

On Tue, Nov 21, 2017 at 12:03 AM, Yoolhee Kim 
wrote:

> Hello,
>
> I'm trying to get RDKIT to recognize LiAsF6 from the SMILES Formula
> "[Li+].F[As-](F)(F)(F)(F)F" but it seems that RDKIT will not read it
> correctly and is unable to create a molecule object.
>
> RDKIT is able to understand LiPF6 from the SMILES formula
> "[Li+].F[P-](F)(F)(F)(F)F" so I tried using that and replacing P with As
> using both replaceatom and replacesubstruct functions, but that had some
> issues as well.
>
> Has anyone faced a similar issue or have any suggestions?
>
> Thanks!
>
> -Yoolhee
>
> 
> --
> 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
>
>
--
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

Re: [Rdkit-discuss] Having trouble getting RDKIT to recognize LiAsF6

2017-11-21 Thread Chris Earnshaw 
Hi

A quick follow-up to my last message, w.r.t. halogens. The issue is
not that higher oxidation state halogen ions can't be constructed at
all, but they end up with bizarre charges on the halogen which are
frequently not recognised by other software.

Chris

On 21 November 2017 at 08:22, Chris Earnshaw  wrote:
> Hi
>
> The entries for P and As in RDKit's atomic_data.cpp are -
> 15  P   0.750.892.0830.974  5   31
> 30.97376163 3   5   7
> 33  As  1.211.2 1.8574.922  5   75
> 74.9215965   3   5
>
> So the required 'connectivities' for PF6- are present in the file
> (final 3 values, '7' being the relevant one here), but not for As. I'm
> not quite sure why this should be so as the phosphorus is P(V) in PF6,
> but this appears to be the cause. You'll have to add the '7' in the As
> entry and rebuild RDKit to allow AsF6- to be processed. A similar fix
> would be needed for antimony too. Someone please let me know if
> there's a more convenient solution!
>
> There are similar issues for the halogens as well. Only iodine has
> values > 1, so by default it's not possible to construct e.g.
> chlorates, or bromates, and no perhalates are allowed.
>
> Regards,
> Chris Earnshaw
>
> On 20 November 2017 at 23:03, Yoolhee Kim  wrote:
>> Hello,
>>
>> I'm trying to get RDKIT to recognize LiAsF6 from the SMILES Formula
>> "[Li+].F[As-](F)(F)(F)(F)F" but it seems that RDKIT will not read it
>> correctly and is unable to create a molecule object.
>>
>> RDKIT is able to understand LiPF6 from the SMILES formula
>> "[Li+].F[P-](F)(F)(F)(F)F" so I tried using that and replacing P with As
>> using both replaceatom and replacesubstruct functions, but that had some
>> issues as well.
>>
>> Has anyone faced a similar issue or have any suggestions?
>>
>> Thanks!
>>
>> -Yoolhee
>>
>> --
>> 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
>>

--
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