Fwd: Extract value and average

[Francesco Pietra]

Sorry, I missed the last part of the message

This makes a more generic parser (comment/uncomment the corresponding
"# 1" or "# 2" code based on whether a new block is found by a first
line containing "NSTEP" or a last line containing "EWALD"). This >
yields a dictionary for each item in the input file. You can pull out
whichever value(s) you want to manipulate.


I recognize now that I did not define "block" correctly. The EWALD
line, although separated by a blank line, makes part of the block. The
layout of all other blocks in the output file is the same.

francesco


---------- Forwarded message ----------
From: Francesco Pietra <[email protected]>
Date: Mon, Jun 8, 2009 at 11:55 PM
Subject: Re: Extract value and average
To: Tim Chase <[email protected]>

Since Python considers whitespace valueable, indentation needs to match.
 When I transfer code into email, I tend to use two-space indentation so
lines don't wrap (in my regular coding, I tend to use hard-tabs with a
tabstop-du-jour of either 2/3/4 spaces-per-tab) and indent it one level to
offset it from the email body text.

If you use Vim, you can copy/paste my emailed code and simply use

 :%s/  /\t/g
 :%s/^\t

to expand the two-space indents to actual tabs and then the second one
strips off the leading indentation.

That done, previous code (for DIHED only) works fine with both the
example file.txt

francesco@deb32:~/tmp$ python tim.py
Total: 4660.165
Count: 1
Average: 4660.165

and a file of 102 instances of DIHED:

francesco@deb32:~/tmp1$ python tim.simple.edited.prod1.py
Total: 465628.4416
Count: 102
Average: 4564.98472157
francesco@deb32:~/tmp1$

Incidentally, that solves my current problems. Thanks indeed for your
generous help.
====================

The more complex parser also worked fine with file.txt:

francesco@deb32:~/tmp$ python tim.py
Total: 4660.165
Count: 1
Average: 4660.165


while (replacing in the parser 'file.txt' with 'prod1.out')
with the 102 count file above it reported:

francesco@deb32:~/tmp$ python tim.prod1.py
Traceback (most recent call last):
 File "tim.prod1.py", line 27, in <module>
   for thing   in      builder():
 File "tim.prod1.py", line 15, in builder
   for k,v     in      pair_re.findall(line)
 File "tim.prod1.py", line 15, in <genexpr>
   for k,v     in      pair_re.findall(line)
ValueError: invalid literal for float(): E

I did the filename replacement with vim, though it is likely that I
did some other
mistake. If not, and you are interested in the prod1.out file, I can
send it (82 kB).

francesco

From the code you sent, it looks like whitespace didn't get copied into your
script file correctly or your un-indentation didn't take place on every line
(the last 4 lines look like they have an extra space).  Once you fix the
whitespace, it should run.  For convenience, I've repasted below with
4-spaces-per-tab and no offsetting indentation.

-tkc

----------------------------
import re
find_dihed_re = re.compile(r'\bDIHED\s*=\s*([.-e\d]+)', re.I)
total = count = 0
for line in file('file.txt'):
   m = find_dihed_re.search(line)
   if m:
       str_value = m.group(1)
       try:
           f = float(str_value)
           total += f
           count += 1
       except:
           print "Not a float: %s" % str_value
print "Total:", total
print "Count:", count
if count:
   print "Average:", total/count