Relative path for dpf import and batch files

[atillack]

This PR adds treatment of relative paths to both dpf import and batch files.

With this PR, when a ligand, grid map, etc. are specified without absolute paths in either a dpf or a batch file when called in from a different location (i.e. autodock_gpu_128wi -B screenA/file.lst) the relative path of the dpf or batch file is prepended to each entry that does not have an absolute path.

While the changes itself should be relatively minor, extensive testing in multiple situations will be required.

[diogomart]

Seems like receptor and flexres filenames get cropped at an underscore which is part of the filename.

$ cat ligs-inert/batch2-d33e098.filelist
../rec/receptor_rigid.maps.fld
../rec/receptor_flex.pdbqt-
../ligs-reactive/sufex1.pdbqt
../out-d33e098/b2-r-1
../ligs-reactive/sufex2.pdbqt
../out-d33e098/b2-r-2
sufex1.pdbqt
../out-d33e098/b2-i-1
sufex2.pdbqt
../out-d33e098/b2-i-2

$ ./adgpu-d33e098-PR225 -B ligs-inert/batch2-d33e098.filelist -T F2,F3,F1=F/P1,P3,P2=P/H6,H5,1H5,H4=HD/C6,C5,C4=Cl/1O4,O4,O6,O5=OA/I2,I1,I3=I/N6,N5,N4=NA/S8,S9,S7=Si/B2,B3,B1=B/S2,S1,S3=S/C2,1C3,C1,1C2,C3=C/S6,S4,S5=SA/B7,B9,B8=BR/B6,B4,B5=Br/A1,A2,A3=A/C9,C8,C7=CL/N1,N3,N2=N -I modpairs.dpf
AutoDock-GPU version: v1.5.3-50-gd33e098821de785deb2547bd8c7fdbf37c4eecb7

Using 4 OpenMP threads

Running 4 docking calculations

Kernel source used for development:      ./device/calcenergy.cl                  
Kernel string used for building:         ./host/inc/stringify.h                  
Kernel compilation flags:                 -I ./device -I ./common -DN128WI   -cl-mad-enable
OpenCL device:                           Quadro RTX 6000
(Thread 2 is setting up Job #2)
(Thread 1 is setting up Job #1)
(Thread 4 is setting up Job #4)
(Thread 5 is setting up Job #3)
Error: Can't open flexibe residue data file ligs-inert/../rec/receptor.
Error: Can't open flexibe residue data file ligs-inert/../rec/receptor.

Error in init_liganddata, stopped job.

Error in setup of Job #1:
(   Grid map file: ligs-inert/../rec/receptor_rigid.maps.fld )
(   Ligand file: ligs-inert/../ligs-reactive/sufex1.pdbqt )
(   Flexible residue: ligs-inert/../rec/receptor )
Error: Can't open flexibe residue data file ligs-inert/../rec/receptor.

[atillack]

@diogomart Thank you - good catch!

[diogomart]

New problem: relative path not prefixed to output (resnam)

$ cat ligs-inert/batch2-dec719d.filelist
../rec/receptor_rigid.maps.fld
../rec/receptor_flex.pdbqt-
../ligs-reactive/sufex1.pdbqt
../out-dec719d/b2-r-1
../ligs-reactive/sufex2.pdbqt
../out-dec719d/b2-r-2
sufex1.pdbqt
../out-dec719d/b2-i-1
sufex2.pdbqt
../out-dec719d/b2-i-2

$ ./adgpu-dec719d-PR225 -B ligs-inert/batch2-dec719d.filelist  -I modpairs.dpf -T F2,F3,F1=F/P1,P3,P2=P/H6,H5,1H5,H4=HD/C6,C5,C4=Cl/1O4,O4,O6,O5=OA/I2,I1,I3=I/N6,N5,N4=NA/S8,S9,S7=Si/B2,B3,B1=B/S2,S1,S3=S/C2,1C3,C1,1C2,C3=C/S6,S4,S5=SA/B7,B9,B8=BR/B6,B4,B5=Br/A1,A2,A3=A/C9,C8,C7=CL/N1,N3,N2=N
AutoDock-GPU version: v1.5.3-51-gdec719d057145d4069ef0d86a8136a0e9fbb8b61

Using 4 OpenMP threads

Running 4 docking calculations

Kernel source used for development:      ./device/calcenergy.cl                  
Kernel string used for building:         ./host/inc/stringify.h                  
Kernel compilation flags:                 -I ./device -I ./common -DN128WI   -cl-mad-enable
OpenCL device:                           Quadro RTX 6000
(Thread 4 is setting up Job #1)
(Thread 2 is setting up Job #2)
(Thread 3 is setting up Job #4)
(Thread 1 is setting up Job #3)

Error: Specified directory "../out-dec719d" for output files (e.g. with `--resnam`) does not exist.

Error: Specified directory "../out-dec719d" for output files (e.g. with `--resnam`) does not exist.

Error: Specified directory "../out-dec719d" for output files (e.g. with `--resnam`) does not exist.

Error: Specified directory "../out-dec719d" for output files (e.g. with `--resnam`) does not exist.
Segmentation fault

[atillack]

@diogomart Thank you! Fixed.

[diogomart]

Filelists work as expected now. 👍
DPF it works with existing inputs (didn't try relative paths).