Thanks for a detailed analysis.
If possible it would be very useful to have a CDK News article describing the design, implementation and limitations of the parser - sort of sumarizing the various threads between you and Dazhi
I am very sorry for not replying to your previous message.
Here is the explanation to the failed tests.
Done. A bit more than 20% of the parsing still fails.OK i have checked the report and these are all expected to me.link: http://cheminfo.informatics.indiana.edu/~rguha/code/java/nightly/test/result-smarts.html
Here is how and why, From Bottom
1) Testcase: testSingleBond status failed,
I have a problem in recognising the single bond now because the single bond and the negative sign is in
conflict and there is no way to understand when single bond will arrive and when negative sign, see
O-C (Oxygen negative charge and has any bond with carbon) or (Oxygen single bonded with carbon)
which one should we pick up. So at present single bond is not perceived i.e. ("-" means negative sign .)
( What will be the solution?)
This is working,I have checked it many times.
the SMARTS contains blank space " " and smart parser does not consume it.
the original expression is this
parse("[c;$([*Cl]),$([*H1])]1ccc(O)c(C)c1 or Cc1:c(O):c:c:[$(cCl),$([cH])]:c1");
the correct one should be
we do not support R at this stage but it is simple to incorporate by changing 'r' to 'r'|'R'
In the atom expression function the valencyatom() is missing that's why this error occured.
But last time when i committed it i have added , i think, well i will update the file
6) Up to to 199 same reasons as 5
7)198 to 195 the expression #16 is causing an error why this occurring i don't know.
But i will remove this error.
8)183,182 same reason as 4
9)178 giving the error because of digit 12 well this is my mistake i have added only DIGIT not DIGITS
before all tokens but i will remove this error now.
8)177 the bond / and \ are not yet supported steriochemistry related stuff are left.
9)176 the token @ is not yet supported as above
10)175 this is certainly a failure from my point of view because the operator are not yet supported for bonds
up to 173 same reason.
11) 172 ,163 ,162 ,161 is same as 4
12) 145 is same as 9
13) 139 to 116 same as 7
12) 101 same as 5
13) 98 ,33 the dot operator is not supported.
14) The biggest error " java.lang.ArrayIndexOutOfBoundsException: -1" has to be handled
OK i will handle this exception this is certainly my mistake.
Note that these tests only test if parsing fails or not. It does not yet testThe resultant QueryAtomContainer is correct in all the cases except those expression having a dot operator.
if the resulting QueryAtomContainer is correct, nor if the right (number of)
substructures are found
You can see the graphical tester link http://sushilronghe.tripod.com/cdk/Unused Code and code style problem
Yes this is clean up required from my side i will do it as soon as possibleJavaDoc coverage
Yes this work is certainly left from my sideThis month is last month of my M.S. Degree i have to search for PH.D.and many formalities are left in my M.S. project molpharm.
also i have to prepare my project thesis etc.
So this month i will be really short of time.But i will try to complete the work that i have suppose to complete.-
Center of Pharmacoinformatics
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
Cdk-devel mailing list