A scraper for tables on SPEC website. Attempts to parse various processor descriptions into more meaningful fields.  It generally does a good job, but still requires some manual cleanup.  To use it, do something like

$ python scrapeSpec.py <URL> <PARSELIB> <TEST>

The arguments following scrapeSpec.py are defined in more detail below.

When the script finishes, it will write a CSV file (the name depends on the arguments ... see scrapeSpec.py for details).

If the arguments are not consistent, the script is likely to error out, or provide incomplete results. It should be possible to write a wrapper that makes using this script convenient, since many of the args depend on each other. For example, something that accepts "--year=1992 --type=int" implies the following arguments:
http://performance.netlib.org/performance/html/new.spec.cint92.col0.html Spec1992Data SpecDataElemInt1992

Argument Descriptions:

<URL>: path to the results on the SPEC website. The list of current URLs is

SPEC1992
http://performance.netlib.org/performance/html/new.spec.cint92.col0.html
http://performance.netlib.org/performance/html/new.spec.cfp92.col0.html

SPEC1995
http://www.spec.org/cpu95/results/cint95.html
http://www.spec.org/cpu95/results/cfp95.html

SPEC2000
http://www.spec.org/cpu2000/results/cint2000.html
http://www.spec.org/cpu2000/results/cfp2000.html

SPEC2006
http://www.spec.org/cpu2006/results/cint2006.html
http://www.spec.org/cpu2006/results/cfp2006.html

<PARSELIB> A class that parses the HTML tables. Each URL will only work with one class
The list of current parsing classes are:
Spec1992Data
Spec1995Data
Spec2000Data
Spec2006Data

<TEST> A class that defines the information about the tests (usually just names). Again, each URL will only work with one of these classes. The list of classes are
SpecDataElemInt1992
SpecDataElemInt1995
SpecDataElemInt2000
SpecDataElemInt2006
SpecDataElemFp1992
SpecDataElemFp1995
SpecDataElemFp2000
SpecDataElemFp2006