pulling headlines out of text products

[WXMikey]

I am working on some code to pull headlines out of National Weather Services text products, more specifically Winter Storm Warnings (WSW) and Non Precipitation Warnings (NPW). I am running into a bit of problems with this.

Each WSW/NPW, most of the time, contains headlines giving more specific weather information and I am trying to pull this out consistantly if the headlines exists. They start out with 3 periods....and ends with 3 periods, but could go from one line to up to 3 lines.

Here is an example of both products...

WWUS44 KEPZ 042222
WSWEPZ

URGENT - WINTER WEATHER MESSAGE
NATIONAL WEATHER SERVICE EL PASO TX
319 PM MST FRI FEB 4 2005

...SNOW ADVISORY FOR THE SOUTHWEST AND SOUTHERN SACRAMENTO MOUNTAINS
FOR THIS EVENING THROUGH SATURDAY...
.AN UPPER LEVEL LOW PRESSURE SYSTEM JUST WEST OF THE BAJA PENINSULA
HAS BEGUN TO PULL SUB TROPICAL MOISTURE OVER SOUTHERN NEW MEXICO.
AREAS OF RAIN WILL DEVELOP LATE THIS AFTERNOON OVER SOUTHWEST AND
SOUTHCENTRAL NEW MEXICO. SNOW WILL OCCUR ABOVE 7000 FEET AND LOWER
TO 5000 FEET BY SATURDAY MORNING. SEVERAL INCHES OF SNOW
ACCUMULATION ARE POSSIBLE FOR MOUNTAIN LOCATIONS BY THE TIME THE
SYSTEMS MOVES OUT OF THE AREA LATE SATURDAY AFTERNOON.

NMZ022-050402-
SOUTHWEST MOUNTAINS/LOWER GILA REGION NM-
INCLUDING THE CITIES OF...SILVER CITY
319 PM MST FRI FEB 4 2005

...SNOW ADVISORY REMAINS IN EFFECT UNTIL SATURDAY AFTERNOON...

SNOW WILL DEVELOP THIS EVENING FOR MUCH OF THE GILA MOUNTAINS. THE
SNOW LEVEL WILL INITIALLY BEGIN AT AROUND 7000 FEET BUT LOWER TO 5000
FEET BY SATURDAY MORNING. SNOW WILL CONTINUE THROUGH MOST OF SATURDAY
ALTHOUGH THE SNOW WILL BECOME SCATTERED SHOWERS. AREAS BELOW 7000
FEET CAN EXPECT BETWEEN 1 TO 3 INCHES OF SNOW ACCUMULATION WITH 3
TO 6 INCHES ABOVE 7000 FEET POSSIBLE. SNOW SHOWERS WILL TAPER OFF
SUNDAY MORNING.

A SNOW ADVISORY IS ISSUED WHEN LIGHT SNOW IS EXPECTED TO DEVELOP, BUT
ONLY ENOUGH TO CAUSE MINOR INCONVENIENCES SUCH AS TRAVEL DELAYS. NO
SEVERE WINTER WEATHER IS EXPECTED AT THIS TIME.

STAY TUNED TO NOAA WEATHER RADIO AND OTHER LOCAL MEDIA FOR FURTHER
DETAILS OR UPDATES.

$$

NMZ025-050415-
SOUTHERN SACRAMENTO MOUNTAINS NM-
319 PM MST FRI FEB 4 2005

...SNOW ADVISORY REMAINS IN EFFECT UNTIL 6 PM MST SATURDAY EVENING...

SNOW WILL DEVELOP THIS EVENING FOR ELEVATIONS AT 7000 FEET AND ABOVE.
THE SNOW LEVEL WILL FALL TO 5000 FEET BY SATURDAY MORNING. SNOW WILL
CONTINUE THROUGH THE NIGHT WITH THE SNOW TURNING TO SHOWERS ON
SATURDAY. THE MOUNTAINS CAN EXPECT FROM 3 TO 5 INCHES TO FALL THROUGH
SATURDAY AFTERNOON.

A SNOW ADVISORY IS ISSUED WHEN LIGHT SNOW IS EXPECTED TO DEVELOP, BUT
ONLY ENOUGH TO CAUSE MINOR INCONVENIENCES SUCH AS TRAVEL DELAYS. NO
SEVERE WINTER WEATHER IS EXPECTED AT THIS TIME.

STAY TUNED TO NOAA WEATHER RADIO AND OTHER LOCAL MEDIA FOR FURTHER
DETAILS OR UPDATES.

$$

WWUS74 KSHV 041220
NPWSHV

URGENT - WEATHER MESSAGE
NATIONAL WEATHER SERVICE SHREVEPORT LA
620 AM CST FRI FEB 4 2005

ARZ050-051-059>061-070>073-OKZ077-TXZ096-097-108>112-041600-
BOWIE-CAMP-CASS-COLUMBIA-FRANKLIN-HEMPSTEAD-HOWARD-LAFAYETTE-
LITTLE RIVER-MCCURTAIN-MILLER-MORRIS-NEVADA-RED RIVER TX-SEVIER-
TITUS-UNION AR-
INCLUDING THE CITIES OF...ASHDOWN...ATLANTA...CLARKSVILLE...
DAINGERFIELD...DE QUEEN...EL DORADO...HOPE...IDABEL...LEWISVILLE...
MAGNOLIA...MT PLEASANT...MT VERNON...NASHVILLE...NEW BOSTON...
PITTSBURG...PRESCOTT...TEXARKANA
620 AM CST FRI FEB 4 2005

...FREEZING FOG ADVISORY IN EFFECT UNTIL 10 AM CST THIS MORNING...

THE NATIONAL WEATHER SERVICE IN SHREVEPORT HAS ISSUED A FREEZING FOG
ADVISORY. A MOSTLY CLEAR SKY...LIGHT WINDS...AND A WET GROUND FROM
RECENT RAINFALL...HAVE ALLOWED FOR AREAS OF DENSE FOG TO DEVELOP
ACROSS ALL OF SOUTHWEST ARKANSAS...SOUTHEAST OKLAHOMA...AND
PORTIONS OF NORTHEAST TEXAS...GENERALLY NORTH OF A PITTSBURG...TO
LINDEN LINE. VISIBILITIES WILL BE REDUCED BELOW ONE QUARTER OF A
MILE...TO NEAR ZERO AT TIMES THROUGH MIDMORNING ACROSS THE
ADVISORY AREA. WITH TEMPERATURES BELOW FREEZING...THE FREEZING FOG
WILL CREATE SLICK SPOTS ON BRIDGES AND OVERPASSES...AND POSSIBLY ON
SOME ROADS AS WELL.

MOTORISTS COMMUTING THIS MORNING SHOULD DRIVE WITH EXTREME CAUTION
AND ALLOW ADDITIONAL TRAVEL TIME TO REACH THEIR DESTINATION. USE LOW
BEAM HEADLIGHTS...REDUCE DRIVING SPEED...AND ALLOW FOR PLENTY OF
ROOM BETWEEN YOU AND OTHER CARS. YOU CAN OVERTAKE ANOTHER VEHICLE
QUICKLY IN POOR VISIBILITIES...SO SLOW DOWN AND DRIVE DEFENSIVELY.

$$

15

Can anyone give me some help on this?

Thanks in advance!
Michael

[rehfeld]

assuming the headline(s) will ALWAYS start and end with 3 periods, this should pull all headlines out of some text

$text = 'your text here';
$pattern = '#\\s\\.\\.\\.([^.]{15,})\\.\\.\\.#';

preg_match_all($pattern, $text, $matches);

print_r($matches);

it assumes a headline will always be more than 15 characters long. you can adjust that value. it also assumes the headline will start on its own line.

[WXMikey]

Thanks but they are not always the same....and could go from one line to three lines long...starting with ... on the first line and ending with ... on the second or third. Also these products contain ... in it alot! Other ideas?

I do see that there is a pattern. The headlines start after the date.

[rehfeld]

i made some changes. give it a shot now.

[Weedpacket]

Starts with ..., ends with ... and may run across lines.

/^\r?\n\.\.\.(.+?)\.\.\.\r?$/ms

I'm guessing the \r characters are unnecessary, but they may be; so I've made it optional (use \r if you know you need them, take \r? out completely if you know you don't, leave it as \r? if you're not sure). I've assumed that every headline is preceded by a blank line (to prevent lines starting with ... buried in the reports from being flagged as headlines.