little script help needed

abisko00 · 07-09-2005, 07:00 AM

Hi!

I have a problem (surprise

).

I have no programming skills at all, but I need to write a script that reads data from a file (tab delimited list, three entries each line), places those three paramters in a textfile, which is then fed into a program. The goal is to run each line of parameters through this program and create a resultfile for each.

My problem is with the parsing of the inputfile. I know that I may be able to solve my problem by reading sed/awk/perl/bash manuals, but I do this for several hours already without making major advances.

So if you are a programming specialist, for whom this is a task of several seconds, could you please push me into the right direction (maybe with an example script)? Thank you very much!

More info:

Input file:

Code:

1	CTTGCTATTGGTAAGGAAACTG	GTATTTGTGCTCTGGGAAGAC
2	CTTGCTATTGGTAAGGAAACTG	TATTTGTGCTCTGGGAAGAC
3	CTTGCTATTGGTAAGGAAACTG	ATTTGTGCTCTGGGAAGAC
4	TTGCTATTGGTAAGGAAACTGG	GTATTTGTGCTCTGGGAAGAC
...

Temporary text file:

Code:

PRIMER_SEQUENCE_ID=NUM
PRIMER_LEFT_INPUT=FWD
PRIMER_RIGHT_INPUT=REV
PRIMER_PICK_ANYWAY=1
PRIMER_MIN_TM=55
PRIMER_EXPLAIN_FLAG=1
...

The parameters to be replaced are NUM, FWD and REV (can be renamed).

keefaz · 07-09-2005, 07:35 AM

I would use cut for parsing the file, and no need of sed, awk nor perl

Code:

#!/bin/bash

datafile="/path/to/datafile"
prefix="config_"
savedir="/path/to/dir/where/save/configs"

while read line; do
    SEQUENCE_ID=$(echo $line | cut -d' ' -f1)
    LEFT_INPUT=$(echo $line | cut -d' ' -f2)
    RIGHT_INPUT=$(echo $line | cut -d' ' -f3)
    cat <<END > "$savedir/$prefix$SEQUENCE_ID"
PRIMER_SEQUENCE_ID=$SEQUENCE_ID
PRIMER_LEFT_INPUT=$LEFT_INPUT
PRIMER_RIGHT_INPUT=$RIGHT_INPUT
END
done < $datafile

for file in $(find $savedir -name "$prefix*"); do
cat <<END >> $file
PRIMER_PICK_ANYWAY=1
PRIMER_MIN_TM=55
PRIMER_EXPLAIN_FLAG=1
...
END
done

Of course I assume PRIMER_SEQUENCE_ID is uniq, I mean
the number corresponding to the first field in your datafile

abisko00 · 07-09-2005, 08:07 AM

Thank you very much!

It took a while, but I think I understand what this is doing.

Is it correct that this script creates a config file for each entry in my datafile? First it writes out the variable information and in the second step, it adds the fixed part to the files (cat). Correct? Great! I think this is all I needed. Thanks again!

keefaz · 07-09-2005, 08:35 AM

Yes, I opted for Keep It Simple way

eddiebaby1023 · 07-09-2005, 04:53 PM

sed is neater:

Code:

sed 's/\([^^I]*\)^I\([^^I]*\)^I\([^^I]*\)/PRIMER_SEQUENCE_ID=\1
PRIMER_LEFT_INPUT=\2
PRIMER_RIGHT_INPUT=\3
PRIMER_PICK_ANYWAY=1
PRIMER_MIN_TM=55
PRIMER_EXPLAIN_FLAG=1/'   inputfile

(The ^I represents a tab character.)

A single command will run much quicker than the cumbersome script, too.

[edit] Sorry, I've just reread your post and seen you want each line in a separate file.

abisko00 · 07-09-2005, 05:35 PM

Thanks anyway, maybe I'll learn something. sed always scares me a bit, because of the special characters and regular expressions. Very confusing!

After I got keefaz's script, everything was easy. I rewrote the code a little (now I can use a more complicated inputfile) and could check several thousand lines in a few minutes. Nice!

bigearsbilly · 07-11-2005, 08:06 AM

That looks like DNA sequences.
Are you creating a monster?

abisko00 · 07-11-2005, 08:37 AM

Quote:

Originally posted by bigearsbilly
That looks like DNA sequences.
Are you creating a monster?

You are right! My monsters are green and totally useless (in an economical sense): I am a plant researcher, working with Arabidopsis thaliana (http://www.arabidopsis.org/info/aboutarabidopsis.jsp ).