search

dancsi / OrthoCCTools

Computational coiled-coil peptide design toolkit, written in modern C++
https://www.dancsi.net/icipa
MIT License
3 stars 2 forks source link

read_scores_binary does not load the IDs from the fasta file #23

Closed ajasja closed 4 years ago

ajasja commented 8 years ago

read_scores_binary does not load the IDs from the fasta file

The IDs are from P1 ... PN+1 But if the IDs in the fasta file are different, then the solver will return the wrong set.

TODO: give the fasta file also when loading the binary score file and load the correct IDs.

dancsi commented 8 years ago

Done, fixed in 7728a86cf3bd0ddef48fbe5ed1f4fc5ab8a37135. Now, the solver tries automatically to read the peptide names from [input file base name].fasta. Alternatively, you can specify the fasta name through the fasta-name command line argument

ajasja commented 8 years ago

Thanks! That was lightning quick:)

ajasja commented 8 years ago

Just a minor issue: If no fasta is given and no fasta with the same basename is present the solver freezes. I would prefer if it throws an exception or something.

ajasja commented 8 years ago

Hi!

Did you change anything else? I'm getting less sets with the newest version. I used to get two homodimeric sets now I get just one.

Attaching the log and output files

new.zip old.zip

dancsi commented 8 years ago

Something very strange is happening, I am hitting an assertion when running the solver under a debugger. This was not introduced by the latest commit, but was appearently silently ignored before. As soon as I debug that, I will get back to fixing this version.

ajasja commented 8 years ago

Great, thanks!

Can you also print the "version" of the solver in the log file (when you print the parameter values).

On 19 October 2016 at 17:08, Daniel Siladji notifications@github.com wrote:

Something very strange is happening, I am hitting an assertion when running the solver under a debugger. This was not introduced by the latest commit, but was appearently silently ignored before. As soon as I debug that, I will get back to fixing this version.

— You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub https://github.com/dancsi/CoiledCoilOrtho/issues/20#issuecomment-254841951, or mute the thread https://github.com/notifications/unsubscribe-auth/AABN1pE-Jt9uyRgw4uux0xmSzEi3Qh2Rks5q1jJbgaJpZM4Ka4Dp .

dancsi commented 8 years ago

I have just fixed something that looks like a bug in the mcqd code, and it is pushed to master. Hopefully it will solve your problem, as it was in the part of the clique solver that estimates the size of a candidate clique. I will fix the rest of the stuff tomorrow...

ajasja commented 8 years ago

Trying out now

On 19 October 2016 at 18:06, Daniel Siladji notifications@github.com wrote:

I have just fixed something that looks like a bug in the mcqd code, and it is pushed to master. Hopefully it will solve your problem, as it was in the part of the clique solver that estimates the size of a candidate clique. I will fix the rest of the stuff tomorrow...

— You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub https://github.com/dancsi/CoiledCoilOrtho/issues/20#issuecomment-254860181, or mute the thread https://github.com/notifications/unsubscribe-auth/AABN1vzLxkrwSOtWU_iULV2T9bJlDei6ks5q1kAKgaJpZM4Ka4Dp .

ajasja commented 8 years ago

Seems to be working fine:)