Extracting information from String (regex, Beanshell)

[Mokita]

Hello,

I am working with Taverna to build a workflow. Taverna has a beanshell where I can program in java. I am having some problems in writing a script. I want to extract information from a string, separated by newline. For this i am using regex.

The String is given:

P48534
EXP value is: e-10
Q0543
EXP value is: 4e-07


My script look like this in Beanshell:

[CODE=java] import java.util.regex .Matcher;
import java.util.regex .Pattern;

Pattern pGI = Pattern.compile ("(^.*?$)");
Pattern pEvaLue = Pattern.compile ("is: (.*)$");
Matcher mGI;
Matcher mEvalue;
StringBuffer temp = new StringBuffer();
String [] line = BlastReport.spl it("\n");
int arraysize = line.length;

for (int i=0; i<(arraysize-1); i+=2){
String sGI = line[i];
String sEvalue = line[i+1];
mGI = pGI.matcher(sGI );
mEvalue = pEvalue.matcher (sEvalue);
String gi="";

if (mGI.find()){
gi =mGI.group(1);
}
if (mEvalue.find() ){
String eval = mEvalue.group(1 );
if(eval.startsW ith("e")){
eval= "1".concat(eval );
}
Double d = new Double (eval);
double Evalue = d.doubleValue() ;
if (Evalue<=0.02){
temp.append(gi + "\n");
}
}
}

String result = temp.toString() .trim();[/CODE]


The error message is: "Attempt to resolve method: matcher() on undefined variable or class name: pEvalue: at Line: 16: pEvalue .matcher(sEvalu e)"


Can someone tell me why is giving me this error and how can i fix it.

Thank you in advance,

Mokita

[prometheuzz]

Hello,

I am working with Taverna to build a workflow. Taverna has a beanshell where I can program in java. I am having some problems in writing a script. I want to extract information from a string, separated by newline. For this i am using regex.

The String is given:

P48534
EXP value is: e-10
Q0543
EXP value is: 4e-07

...

It looks like you're trying to capture the text after the is:, right?
Try this:
[CODE=java] String text = "P48534\nEX P value is: e-10\nQ0543\nEXP value is: 4e-07";
Pattern pattern = Pattern.compile ("(?<=is:\\s )[^\\n]+");
Matcher matcher = pattern.matcher (text);

while(matcher.f ind()) {
System.out.prin tln(matcher.gro up());
}[/CODE]

[r035198x]

Which one is your line 16?

[Mokita]

My line 16 is:
mEvalue = pEvalue.matcher (sEvalue);

Mokita

[r035198x]

My line 16 is:
mEvalue = pEvalue.matcher (sEvalue);

Mokita

Java is case sensitive.
Now give yourself a kick.

[prometheuzz]

My line 16 is:
mEvalue = pEvalue.matcher (sEvalue);

Mokita

That splitting of your String and increasing with 2 in your for loop looks dangerous. Perhaps I (or someone else) can suggest a better approach. But then you first need to explain what is you're trying to do.

So given the String: [CODE=java]"P48534\nEX P value is: e-10\nQ0543\nEXP value is: 4e-07"[/CODE]what is it you're trying to extract and/or group?

[prometheuzz]

My line 16 is:
mEvalue = pEvalue.matcher (sEvalue);

Mokita

Try this:
[CODE=java]String text = "P48534\nEX P value is: e-10\nQ0543\nEXP value is: 4e-07";
Pattern pattern = Pattern.compile ("([A-Z]\\d+).*\\n?.*(( ?<=is:\\s)[^\\n]+)");
Matcher matcher = pattern.matcher (text);

System.out.prin tln(text+"\n");

while(matcher.f ind()) {
String id = matcher.group(1 );
String sVal = matcher.group(2 );
sVal = sVal.startsWith ("e") ? 1+sVal : sVal;
double dVal = Double.parseDou ble(sVal);
System.out.prin tln(id+"\t"+dVa l);
}[/CODE]

[Mokita]

That splitting of your String and increasing with 2 in your for loop looks dangerous. Perhaps I (or someone else) can suggest a better approach. But then you first need to explain what is you're trying to do.

So given the String: [CODE=java]"P48534\nEX P value is: e-10\nQ0543\nEXP value is: 4e-07"[/CODE]what is it you're trying to extract and/or group?

I am trying to do a workflow with taverna, which has a beanshell, where i can write a script in it.
In my workflow i will have the output of a blast search, GI number, which are the P23234 or Q12344 or A12443 or only numbers and also a E-value, which is the EXP vaule is: e-10.
From that output i want to extract the GI numbers which have an e-value<= 0.02. The way i thought i could extract was with regular expressions in java, but the way i wrote the script it is not working.

I think it is more clear what i want to do, but if you need more explanation please ask.

Mokita

[r035198x]

I am trying to do a workflow with taverna, which has a beanshell, where i can write a script in it.
In my workflow i will have the output of a blast search, GI number, which are the P23234 or Q12344 or A12443 or only numbers and also a E-value, which is the EXP vaule is: e-10.
From that output i want to extract the GI numbers which have an e-value<= 0.02. The way i thought i could extract was with regular expressions in java, but the way i wrote the script it is not working.

I think it is more clear what i want to do, but if you need more explanation please ask.

Mokita

I don't think regex is best for this (I could be wrong of course). You are not searching for a pattern (which is where I mostly use my regex) but you are searching for numbers within some range.
P.S I hope you managed to correct the spelling mistake for that variable.

[Mokita]

Hello

I want to thank you for your help, it is working. I also want to ask you a quick question.
How can i change group 1 ([A-Z]\\d+) to catch: P00DF3 or 1234653 or Q5647GJD or A4658DF
It is not catching the ones with letters in the midle.

Thank you again,

Mokita

[prometheuzz]

...
How can i change group 1 ([A-Z]\\d+) to catch: P00DF3 or 1234653 or Q5647GJD or A4658DF
It is not catching the ones with letters in the midle.

Thank you again,

Mokita

Try replacing "([A-Z]\\d+)" with "(\\w+)"