AztecOO: error: identifier “sswap_” is undefined

By K4Paul 0 Development, Linux , , , ,

Error
I’ve been trying to compile the AztecOO library outside of Trilinous. In doing so, I pulled the package Aztec source code from the Trilinous source code and opted to write my own Makefile to then compile it. My Makefile included the same settings reported in the Trilinous Makefile.export.AztecOO file.

To prepare this build, I needed to fetch the AztecOO_config.h file from a built version of Trilinous and place it in the source directory of the Aztec directory I was compiling (this file is generated upon ‘make’ in Trilinous). With that in place, I had the files I needed to compile Aztec, however, I ran into the following error.

error: identifier "sswap_" is undefined

Solution
SWAP_F77 is a defined module that points to the F77_BLAS_MANGLE module at the top of az_c_util.c. This second module is defined in AztecOO_config.h. All this module does, is add a “_” to the end of the name provided in the parameters and defines it with the provided value. This works when you are using the Fortran Blas library, but I am using the C version. To fix this, I edited all F77 functions in the AztecOO_config.h file to just return “name” instead of “name ## _”.

I also added this include statement to the file that errors src/az_c_util.c which required that I add –I to the compile statement in order for the compiler to find the right header files.

#include "mkl_blas.h"

Python URL Term Scrapper (Kind of Like a Very Dumb/Simple Watson)

By K4Paul 0 Development, Python, Web , , , , ,

I wrote this a while back to scrape a given URL page for all links it contains and then search those links for term relations. The Python scrapper first finds all links in a given url. It then searches all the links found for a list of search terms provided. It will return stats on the number of times specific provided terms show up.

This may come in handy while trying to find more information on a given topic. I’ve used it on Google searches (be careful, you can only scrape google once ever 8 or so seconds before you are locked out) and wikipedia pages to gather correlation statistics between topics.

It is old code… so there might be some errors. Keep me posted!

#!/usr/bin/env python
#ENSURE permissions are 755 in order to have script run as executable

import os, sys, re, datetime
from optparse import OptionParser
import logging, urllib2

def parsePage(link, list):
    searchList = {}
    try:
        f = urllib2.urlopen(link)
        data = f.read()
        for item in list:
            if (item.title() in data) or (item.upper() in data) or (item.lower() in data):
                searchList[item]=searchList[item]+1
                searchList["count"]=searchList["count"]+1
        return searchList
    except Exception, e:
        print "An error has occurred while parsing page " +str(link)+"."
        log.error(str(datetime.datetime.now())+" "+str(e))

def searchUrl(search):
    try:
        f = urllib2.urlopen(search)
        data = f.read()
        pattern = r"/wiki(/\S*)?$" #regular expression to find url
        links = re.findall(pattern, data)
        return links
    except Exception, e:
        print "An error has occurred while trying to reach the site."
        log.error(str(datetime.datetime.now())+" "+str(e))

def main():
    try:
        parser = OptionParser() #Help menu options
        parser.add_option("-u", "--url", dest="search", help="String containing URL to search.")
        parser.add_option("-f", "--file", dest="file", help="File containing search terms.")
        (options, args) = parser.parse_args()
        if not options.search or not options.file:
            parser.error('Term file or URL to scrape not given')
        else:
            urls = searchUrl(options.search)
            f = open(options.file, 'r')
            terms = f.readlines()
            for url in urls:
                parsePage(url, terms)
            print "Results:"
            print searchList
    except Exception, e:
        log.error(str(datetime.datetime.now())+" "+str(e))

if __name__ == "__main__":
    log = logging.getLogger("error") #create error log
    log.setLevel(logging.ERROR)
    formatter = logging.Formatter('[%(levelname)s] %(message)s')
    handler = logging.FileHandler('error.log')
    handler.setFormatter(formatter)
    log.addHandler(handler)
    try:
        main()
    except Exception, e:
        print "An error has occurred, please review the error.log for more details."
        log.error(str(datetime.datetime.now())+" "+str(e))