# Challenge, return number of matches within a page
import mpi
import urllib

# what we are searching for
inputs = [
    "seidel",
    "beck",
    "loni",
    "eastman",
    "brandt"]

# collection of web pages our search engine knows about
pages = [
  'http://www.cct.lsu.edu/about/overview/',
  'http://www.cct.lsu.edu/projects/Coastal_Modeling/',
  'http://www.cct.lsu.edu/projects/SURASCOOP/',
  'http://www.cct.lsu.edu/projects/CactusCode/',
  'http://www.cct.lsu.edu/about/employment/employment.php',
  'http://www.cct.lsu.edu/projects/GridChemCCG/',
  'http://www.cct.lsu.edu/about/people/faculty/all.php',
  'http://www.cct.lsu.edu/about/focus/',
  'http://www.cct.lsu.edu/projects/Enlightened/',
  'http://www.loni.org/',
  'http://www.loni.org/plan/',
  'http://www.cct.lsu.edu/news/news/289']

n = len(pages)/mpi.size
ilo = mpi.rank*n
ihi = (mpi.rank+1)*n-1

# each mpi proc searches a subset of pages from ilo to ihi
c = range(ihi+1)
for i in range(ilo,ihi+1):
    c[i] = urllib.urlopen(pages[i]).read().lower()

for input in inputs:
    matches = []

    for i in range(ilo,ihi+1):
        if c[i].find(input) >= 0:
            matches.append(pages[i])

    # proc zero receives the results of all searches
    if mpi.rank == 0:
        for i in range(1,mpi.size):
            other_matches = mpi.recv(mpi.ANY_SOURCE)[0]
            for match in other_matches:
                matches.append(match)
    else:
        mpi.send(matches,0)

    if mpi.rank == 0:
        print input,len(matches)