DisconnectionDPS

disconnectionDPS is a hacked version of disconnection.f90 that can read in pathsample generated output files min.data, and ts.data. It also has the ability of producing postscripts with color. The disconnection programs are currently only compiled with the Portland Fortran compilers. The input parameters are read from a file called dinfo.

A Minimal example of dinfo

     delta   15
     first   -490.0
     levels  10
     minima  min.data
     ts      ts.data
     connectmin <integer>
     identify
     

delta - energy difference between levels

first - qualifies highest energy state (the energy at the top of the graph, which should be chosen in combination with delta and levels so that you include everything you need to in the graph)

levels - specifies number energy levels

minima - file name of pathsample-style database file for minima. The first column must contain the energies; other columns are ignored.

ts - file name of pathsample-style database file for transition states. 1st column: ts energy, 2nd and 3rd columns ignored, 4th and 5th must contain the indices of the connected minima. Other columns ignored.

(Note that the format of the minima and ts data files is different from that expected by the original disconnection program...)

connectmin - in case there are disjoint sets of stationary points in the network (2 or more "islands" of stationary points connected only among themselves), we need to specify a minimum in the one set we are using to plot the tree (other islands will be ignored). It's usual to choose the index of one of the end point minima here.

identify - branch endpoints are labeled with the index of the minima. May be used in conjunction with pathsample's EXTRACTMIN keyword.

There are many other flags for centering the trunk, output files, output appearance.

e.g. TSTHRESH and NCONNMIN, which have the same meaning as in PATHSAMPLE.

All the keywords are documented at the top of the disconnectionDPS.f90 file itself.

The minima can be reorganized with manipulate.f90. Nodes are only moved horizontally.

The minima can be coloured with the keyword TRMIN.

     TRMIN <n> <max> <file> <file>

From Vanessa, ' Label n different sections of the graph in colour as specified by the minima in each file, one file for each section. Each file is a list of numbers of minima, one per line as for PICK. max is the total number of minima, currently used for array allocation. Colours are chosen automatically to spread over a rainbow spectrum from red to purple) in the order the files are specified but colours can be specified individually at both COLOURMARKER in this file.

From Livia:

IDENTIFY_NODE <max_min> - If present, the nodes are labelled with the format N1_N2, where N1 is the number of the level, N2 is the number of the node at that level. The label is only printed if the number of minima corresponding to that node is smaller than <max_min>. With this information you can pick the numbers of minima correspondig to that node from the node_numbers file, if you also use the keyword DUMPNUMBERS (this allows you to print lonely any branch of the graph)

IDENTIFY_NODE_SIZE <max_min2> - If present, the nodes are labelled with number of minima corresponding to them. The label is only printed if the number of minima below that node is smaller than <max_min2>

LABELSIZE <n> - Set the size of the fonts in case of the labels (for IDENTIFY, IDENTIFY_NODE ...) Default is 10 pt

--User:mp466 15:15, 1 July 2008 (BST)