Difference between revisions of "CompSciWeek5"

From Predictive Chemistry
Jump to: navigation, search
 
Line 3: Line 3:
 
* [http://www.aosabook.org/en/posa/working-with-big-data-in-bioinformatics.html Bioinformatics Code Design]
 
* [http://www.aosabook.org/en/posa/working-with-big-data-in-bioinformatics.html Bioinformatics Code Design]
   
= Class 1: Effective Design =
+
= Class 1 and 2: Code walkthrough =
* Code walk-through, parsing and using graphs
+
* Structuring a list min/max problem over multiple timescales
* Structured Code, Bioinformatics example from AOS Book
+
* Building and using a structured nearest neighbor graph
* Code Testing
 
* Source Code Versioning
 
** basic git
 
 
= Class 2: Using HPC Resources =
 
* Accessing binaries and libraries, using modules
 
* Using scratch space
 
* Submitting a job script
 
* Managing queued jobs
 
* Advanced scripting tips and tricks
 
** awk
 
   
 
= Example Codes =
 
= Example Codes =
  +
== List Min/Max ==
  +
See [[CompSciWeek1]].
   
 
== Using Graphs ==
 
== Using Graphs ==

Latest revision as of 11:04, 29 September 2014

Reading:

Class 1 and 2: Code walkthrough

  • Structuring a list min/max problem over multiple timescales
  • Building and using a structured nearest neighbor graph

Example Codes

List Min/Max

See CompSciWeek1.

Using Graphs

<source lang="python">

  1. !/usr/bin/env python

from numpy import array, sum, reshape

  1. input - list of names
  2. - (3n)x 3 coordinate array
  3. output - Graph G

def make_G(names, x):

   assert x.shape[0]%3 == 0
   assert x.shape[1] == 3
   assert len(names) == len(x)
   print x.shape
   n = len(x)/3
   G = {}
   D2 = sum((reshape(x, (1,3*n,3)) - reshape(x, (3*n,1,3)))**2, -1)
   print D2
   # add bonds to G!
   for i,n in enumerate(names):
       print i,n
       G[i] = set()
       if n != 'O': continue
       for j,m in enumerate(names):
           print j,m
           if m != 'H': continue
           # all O-H distances, 1 at a time
           if D2[i,j] < 1.1:
               G[i].add(j)
               G[j].add(i) # Question for class: why does this cause an error?
   return G
  1. input - list of names (1 O per 2 H-s)
  2. - graph from atom number to atom numbers
  3. output - list of atom numbers ordered O, H, H, O, H, H, ...

def outp_graph(names, G):

   out = []
   for i,n in enumerate(names):
       if n != "O":
           continue
       out.append(i) # 'O' number
       bonds = G[i]
       assert len(bonds) == 2, "Bad number (%d) of O-bonds."%(len(bonds))
       for j in bonds: # 'H' numbers
           out.append(j)
   return out

names = ['H', 'O', 'H', 'H', 'H', 'O'] x = array([[1,0,0],[0,0,0],[-1,0,0]]) G = make_G(names[:3], x) print G

  1. G = {0:{1}, 1:{0,4, 1}, 2:{5}, 3:{5}, 4:{1}, 5:{2,3} }
  2. l = outp_graph(names, G)
  3. print l

</source>