2 in One....

seangell

Hi all.
some help on these would be greatly appreciated....

ok, i have file, the content of which is more or less like below..

IDN25900
TOP PRIORITY FOR IMMEDIATE BROADCAST
SYDNEY SEVERE THUNDERSTORM WARNING
BUREAU OF METEOROLOGY
NEW SOUTH WALES REGIONAL OFFICE Issued at 1703 on Saturday the 1st of
September 2001
This warning affects people in the following Local Government Area:

Wyong

This warning is current from 5:05pm until 6pm.

Storms are currently located to northwest of Wyong and are moving towards
the southeast. Locations that may be affected in the next hour include the
Wyong and The Entrance

Large hailstones, damaging winds and very heavy rainfall are possible.

The STATE EMERGENCY SERVICE advises that as storms approach people should:
put vehicles under cover
move indoors away from windows

During and after the storm people should:
beware of fallen trees and power lines
keep away from creeks and drains as you may be swept away

The RTA recommends motorists switch their lights & wipers on & slow down in
the wet.

If your house is damaged contact the State Emergency Service on 132 500 for
emergency assistance. Do not use the telephone during the storm.

TV CRAWL: Severe Thunderstorm Warning current for the Wyong Local
Government Area

MEDIA PLEASE NOTE: This warning will be updated within the next hour. The
Bureau and SES would appreciate it being broadcast regularly during this
period.

Now, i need to abbreviate that to 160 characters, and my current script that does it, is "clunky". It grabs the whole file, and abbreviates/deletes words etc etc.

I figure the has to be a better way, and such though it would be easier to say, parse the file for ceratin keywords, such as

"SYDNEY SEVERE THUNDERSTORM WARNING","Issued at 1703" ,"Saturday the 1st of September 2001", "Local Government Area: Wyong
"
etc etc, and these keywords would be the basis of the abbreviation. and out put would look something like

NSW 1703 1/09/ Svr TS Warning for Wyong

How can i achieve this quickly and effeciently, presently the code i do have, doesn't do a great job, and is growing by the day...

any help on getting this fixed would be great, thnx

SiMmAh
edit: removed 2nd Q

bubblenut

Firstly, could you keep seperate questions in seperate posts please? It can get a little confusing when there are two lines of conversation regarding two different problems going on in the same thread. Also (on the same line in a way) please use a relevant subject line so that people can see what the problem is about before they enter the thread. There are many people on these boards (myself included) who don't have the time to check every single post and so just filter through on titles. I actually thought this was going to be either a numeric or array problem, not sure why:/

Now, on to the problem. I'm going to tackle the first problem so if you could edit your post to remove the second and start a new thread for that one, that would be great.

This, I'm afraid, is going to be quite a difficult thing to code concisely (but we'll try 😉) as what you are basically asking the program to do is understand what is writting to a certain extent. I am going to tackle this problem with extensive use of the [man]PCRE[/man] functions so I suggest you read up on them if you're not familiar.

Before we can look further into the problem there are a couple of points which need to be cleared up.

Does the format of this file fit a standard structure?
Is the processed result to be stored or just displayed straight to the output.
If it is to be stored, how? As a flat file like a log or in a database?
Are there standard words used to measure severity? (this could be nice if you're planning to archive the information in a database as it would give you the ability to, say, list all storms in December 2003 in order of severity)
[/list=a]

seangell

Originally posted by bubblenut

Before we can look further into the problem there are a couple of points which need to be cleared up.

Does the format of this file fit a standard structure?

Is the processed result to be stored or just displayed straight to the output.

If it is to be stored, how? As a flat file like a log or in a database?

Are there standard words used to measure severity? (this could be nice if you're planning to archive the information in a database as it would give you the ability to, say, list all storms in December 2003 in order of severity)
[/list=a] [/B]

No, there are differences thorugh all diff files, using $lines = file(blah) works nicely, being able to abb and print $lines[number], just means i have to make a seperate file for just about all the diff warnings (~50 or so) that are customised to each warning..
Just displayed to the ouput
no storing
Nope, standard words just to identify the warning type, and places its for as well as date/time..etc etc...
[/list=a]

SiM
(soz for the 2Q's before, dun know what i was thinking)

bubblenut

Right, well the data we need can be easily divided up into three sections: the nature of the warning; the location of the warning; and the time of issue. Let's deal with what will probably be the easiest first, the time of issue.

Will the time format always be as in the example? ie
"Issued at [hour] on [day of the week] the [date] of
[month] [year]" followed by a new line?
If it will be we have two, very simple options. Firstly we can search for the string "Issued at" and then take it and everything after it, up untill the next newline character and that's it. If we want to display it differently we'll have to use a regular expression (have you read that [man]PCRE[/man] stuff yet because I'm going to make you do some of it as an exercise?🙂).

if(preg_match('\/Issued at ([0-9]{4}) on ([a-zA-Z]+) the ([0-9]{1,2})[a-z]+ ([0-9]{4})\n\/", $string, $matches)) {
  //we've found the data
  list($tmp, $time, $day, $date, $month, $year) = $matches;
  //we can now use the variables to structure the date how we want to.
}

After reading the php manual on PCRE can you see any areas of the above regular expression which could be improved?

Hint: there are a few points. It's quite tricky though. It's to do with them not being as specific as they could be.

seangell

Thanks for taking the time to help me with this...
Now. its 9pm, ive been up since 5am and gotta get up at 5am again tmw, so im heading off to bed now, but will go though the PCRE manual tomorrow after work.. all to confusing right now...

SiM

bubblenut

Fair enough, it's a bit of a brain-melter to start with but once you get the hang of it it's an absolute god send.

seangell

i have been busy moving house so haven't had a chance to look at it real good, will take the chance tonight hopefully and see what i can suss out 🙂