I had to replace all instances of scientific notation with fixed point, and wrote a Python script to help out. Here it is:
#!/usr/bin/python2
# Originally written by Anton Golov on 29 Jan 2012.
# Feedback can be sent to <email> .
"""Provide functions for converting scientific notation to fixed point.
When imported, reads lines from stdin and prints them back to stdout with
scientific notation removed.
WARNING: The resulting fixed point values may be of a different precision
than the original values. This will usually be higher, but may be lower for
large numbers. Please don't rely on the precision of the values generated by
this script for error analysis.
"""
from re import compile
exp_regex = compile(r"(\d+(\.\d+)?)[Ee](\+|-)(\d+)")
def to_fixed_point(match):
"""Return the fixed point form of the matched number.
Parameters:
match is a MatchObject that matches exp_regex or similar.
If you wish to make match using your own regex, keep the following in mind:
group 1 should be the coefficient
group 3 should be the sign
group 4 should be the exponent
"""
sign = -1 if match.group(3) == "-" else 1
coefficient = float(match.group(1))
exponent = sign * float(match.group(4))
return "%.16f" % (coefficient * 10**exponent)
def main():
"""Read lines from stdin and print them with scientific notation removed.
"""
try:
while True:
line = raw_input('')
print exp_regex.sub(to_fixed_point, line)
except EOFError:
pass
if __name__ == "__main__":
main()
I am primarily interested about the way the regex is used (can it match something it can't replace?), but comments on code clarity and features that may be simple to add are also welcome, as well as any further nitpicking .
