Busca e Recorte em arquivo PDF

[3l3tr1c]

E aí galera! A "parada é simples". Eu consigo retornar, a partir de uma classe, o texto puro de um PDF para uma variável. Assim sendo, eu só preciso PROCURAR e RECORTAR o parágrafo que contém o nome que eu desejo. Um exemplo é o recorte de um trecho do Diário Oficial do Supremo Tribunal Federal, segue um link para exemplo:

DIÁRIO OFICIAL STF

Suponha que eu esteja interessado em saber se eu fui CITADO no diário do STF. Para isso a programação deve procurar no diario o meu nome completo e se achar recortar o parágrafo em que eu fui citado e enviar para o meu e-mail. O problema está SOMENTE EM RECORTAR o parágrafo que aparece o meu nome. Aqui está a classe que uso para obter o texto do PDF (Funciona!):

class.pdf2text.php

<?php
/*
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.

This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA

This code is an improved version of what can be found at:
http://www.webcheatsheet.com/php/reading_clean_text_from_pdf.php

AUTHOR:
- Webcheatsheet.com (Original code)
- Joeri Stegeman (joeri210 [at] yahoo [dot] com) (Class conversion and fixes/adjustments)

DESCRIPTION:
This is a class to convert PDF files into ASCII text or so called PDF text extraction.
It will ignore anything that is not addressed as text within the PDF and any layout.
Currently supported filters are: ASCIIHexDecode, ASCII85Decode, FlateDecode

PURPOSE(S):
Most likely for people that want their PDF to be searchable.

SYNTAX:
include('class.pdf2text.php');
$a = new PDF2Text();
$a->setFilename('test.pdf');
$a->decodePDF();
echo $a->output(); 

ALTERNATIVES:
Other excellent options to search within a PDF:
- Apache PDFbox (http://pdfbox.apache.org/). An open source Java solution
- pdflib TET (http://www.pdflib.com/products/tet/)
- Online converter: http://snowtide.com/PDFTextStream
*/


class PDF2Text {
        // Some settings
        var $multibyte = 2; // Use setUnicode(TRUE|FALSE)
        var $convertquotes = ENT_QUOTES; // ENT_COMPAT (double-quotes), ENT_QUOTES (Both), ENT_NOQUOTES (None)
        
        // Variables
        var $filename = '';
        var $decodedtext = '';
        
        function setFilename($filename) { 
                // Reset
                $this->decodedtext = '';
                $this->filename = $filename;
        }

        function output($echo = false) { 
                if($echo) echo $this->decodedtext;
                else return $this->decodedtext;
        }

        function setUnicode($input) { 
                // 4 for unicode. But 2 should work in most cases just fine
                if($input == true) $this->multibyte = 4;
                else $this->multibyte = 2;
        }

        function decodePDF() { 
                // Read the data from pdf file
                $infile = @file_get_contents($this->filename, FILE_BINARY); 
                if (empty($infile)) 
                        return ""; 
        
                // Get all text data.
                $transformations = array(); 
                $texts = array(); 
        
                // Get the list of all objects.
                preg_match_all("#obj[\n|\r](.*)endobj[\n|\r]#ismU", $infile, $objects); 
                $objects = @$objects[1]; 
        
                // Select objects with streams.
                for ($i = 0; $i < count($objects); $i++) { 
                        $currentObject = $objects[$i]; 
        
                        // Check if an object includes data stream.
                        if (preg_match("#stream[\n|\r](.*)endstream[\n|\r]#ismU", $currentObject, $stream)) { 
                                $stream = ltrim($stream[1]); 
        
                                // Check object parameters and look for text data. 
                                $options = $this->getObjectOptions($currentObject); 
        
                                if (!(empty($options["Length1"]) && empty($options["Type"]) && empty($options["Subtype"]))) 
                                        continue; 
        
                                // Hack, length doesnt always seem to be correct
                                unset($options["Length"]);
        
                                // So, we have text data. Decode it.
                                $data = $this->getDecodedStream($stream, $options);  
        
                                if (strlen($data)) { 
                        if (preg_match_all("#BT[\n|\r](.*)ET[\n|\r]#ismU", $data, $textContainers)) {
                                                $textContainers = @$textContainers[1]; 
                                                $this->getDirtyTexts($texts, $textContainers); 
                                        } else 
                                                $this->getCharTransformations($transformations, $data); 
                                } 
                        } 
                } 
        
                // Analyze text blocks taking into account character transformations and return results. 
                $this->decodedtext = $this->getTextUsingTransformations($texts, $transformations); 
        }


        function decodeAsciiHex($input) {
                $output = "";
        
                $isOdd = true;
                $isComment = false;
        
                for($i = 0, $codeHigh = -1; $i < strlen($input) && $input[$i] != '>'; $i++) {
                        $c = $input[$i];
        
                        if($isComment) {
                                if ($c == '\r' || $c == '\n')
                                        $isComment = false;
                                continue;
                        }
        
                        switch($c) {
                                case '': case '\t': case '\r': case '\f': case '\n': case ' ': break;
                                case '%': 
                                        $isComment = true;
                                break;
        
                                default:
                                        $code = hexdec($c);
                                        if($code === 0 && $c != '0')
                                                return "";
        
                                        if($isOdd)
                                                $codeHigh = $code;
                                        else
                                                $output .= chr($codeHigh * 16 + $code);
        
                                        $isOdd = !$isOdd;
                                break;
                        }
                }
        
                if($input[$i] != '>')
                        return "";
        
                if($isOdd)
                        $output .= chr($codeHigh * 16);
        
                return $output;
        }
        
        function decodeAscii85($input) {
                $output = "";
        
                $isComment = false;
                $ords = array();
                
                for($i = 0, $state = 0; $i < strlen($input) && $input[$i] != '~'; $i++) {
                        $c = $input[$i];
        
                        if($isComment) {
                                if ($c == '\r' || $c == '\n')
                                        $isComment = false;
                                continue;
                        }
        
                        if ($c == '' || $c == '\t' || $c == '\r' || $c == '\f' || $c == '\n' || $c == ' ')
                                continue;
                        if ($c == '%') {
                                $isComment = true;
                                continue;
                        }
                        if ($c == 'z' && $state === 0) {
                                $output .= str_repeat(chr(0), 4);
                                continue;
                        }
                        if ($c < '!' || $c > 'u')
                                return "";
        
                        $code = ord($input[$i]) & 0xff;
                        $ords[$state++] = $code - ord('!');
        
                        if ($state == 5) {
                                $state = 0;
                                for ($sum = 0, $j = 0; $j < 5; $j++)
                                        $sum = $sum * 85 + $ords[$j];
                                for ($j = 3; $j >= 0; $j--)
                                        $output .= chr($sum >> ($j * 8));
                        }
                }
                if ($state === 1)
                        return "";
                elseif ($state > 1) {
                        for ($i = 0, $sum = 0; $i < $state; $i++)
                                $sum += ($ords[$i] + ($i == $state - 1)) * pow(85, 4 - $i);
                        for ($i = 0; $i < $state - 1; $i++)
                                $ouput .= chr($sum >> ((3 - $i) * 8));
                }
        
                return $output;
        }
        
        function decodeFlate($input) {
                return gzuncompress($input);
        }
        
        function getObjectOptions($object) {
                $options = array();

                if (preg_match("#<<(.*)>>#ismU", $object, $options)) {
                        $options = explode("/", $options[1]);
                        @array_shift($options);
        
                        $o = array();
                        for ($j = 0; $j < @count($options); $j++) {
                                $options[$j] = preg_replace("#\s+#", " ", trim($options[$j]));
                                if (strpos($options[$j], " ") !== false) {
                                        $parts = explode(" ", $options[$j]);
                                        $o[$parts[0]] = $parts[1];
                                } else
                                        $o[$options[$j]] = true;
                        }
                        $options = $o;
                        unset($o);
                }
        
                return $options;
        }
        
        function getDecodedStream($stream, $options) {
                $data = "";
                if (empty($options["Filter"]))
                        $data = $stream;
                else {
                        $length = !empty($options["Length"]) ? $options["Length"] : strlen($stream);
                        $_stream = substr($stream, 0, $length);
        
                        foreach ($options as $key => $value) {
                                if ($key == "ASCIIHexDecode")
                                        $_stream = $this->decodeAsciiHex($_stream);
                                if ($key == "ASCII85Decode")
                                        $_stream = $this->decodeAscii85($_stream);
                                if ($key == "FlateDecode")
                                        $_stream = $this->decodeFlate($_stream);
                                if ($key == "Crypt") { // TO DO
                                }
                        }
                        $data = $_stream;
                }
                return $data;
        }
        function getDirtyTexts(&$texts, $textContainers) {
                
                for ($j = 0; $j < count($textContainers); $j++) {
                        if (preg_match_all("#\[(.*)\]\s*TJ[\n|\r]#ismU", $textContainers[$j], $parts))
                                $texts = array_merge($texts, @$parts[1]);
                        elseif(preg_match_all("#T[d|w|m|f]\s*(\(.*\))\s*Tj[\n|\r]#ismU", $textContainers[$j], $parts))
                                $texts = array_merge($texts, @$parts[1]);
                        elseif(preg_match_all("#T[d|w|m|f]\s*(\[.*\])\s*Tj[\n|\r]#ismU", $textContainers[$j], $parts))
                                $texts = array_merge($texts, @$parts[1]);
                }
        }
        function getCharTransformations(&$transformations, $stream) {
                preg_match_all("#([0-9]+)\s+beginbfchar(.*)endbfchar#ismU", $stream, $chars, PREG_SET_ORDER);
                preg_match_all("#([0-9]+)\s+beginbfrange(.*)endbfrange#ismU", $stream, $ranges, PREG_SET_ORDER);
        
                for ($j = 0; $j < count($chars); $j++) {
                        $count = $chars[$j][1];
                        $current = explode("\n", trim($chars[$j][2]));
                        for ($k = 0; $k < $count && $k < count($current); $k++) {
                                if (preg_match("#<([0-9a-f]{2,4})>\s+<([0-9a-f]{4,512})>#is", trim($current[$k]), $map))
                                        $transformations[str_pad($map[1], 4, "0")] = $map[2];
                        }
                }
                for ($j = 0; $j < count($ranges); $j++) {
                        $count = $ranges[$j][1];
                        $current = explode("\n", trim($ranges[$j][2]));
                        for ($k = 0; $k < $count && $k < count($current); $k++) {
                                if (preg_match("#<([0-9a-f]{4})>\s+<([0-9a-f]{4})>\s+<([0-9a-f]{4})>#is", trim($current[$k]), $map)) {
                                        $from = hexdec($map[1]);
                                        $to = hexdec($map[2]);
                                        $_from = hexdec($map[3]);
        
                                        for ($m = $from, $n = 0; $m <= $to; $m++, $n++)
                                                $transformations[sprintf("%04X", $m)] = sprintf("%04X", $_from + $n);
                                } elseif (preg_match("#<([0-9a-f]{4})>\s+<([0-9a-f]{4})>\s+\[(.*)\]#ismU", trim($current[$k]), $map)) {
                                        $from = hexdec($map[1]);
                                        $to = hexdec($map[2]);
                                        $parts = preg_split("#\s+#", trim($map[3]));
                                        
                                        for ($m = $from, $n = 0; $m <= $to && $n < count($parts); $m++, $n++)
                                                $transformations[sprintf("%04X", $m)] = sprintf("%04X", hexdec($parts[$n]));
                                }
                        }
                }
        }
        function getTextUsingTransformations($texts, $transformations) {
                $document = "";
                for ($i = 0; $i < count($texts); $i++) {
                        $isHex = false;
                        $isPlain = false;
        
                        $hex = "";
                        $plain = "";
                        for ($j = 0; $j < strlen($texts[$i]); $j++) {
                                $c = $texts[$i][$j];
                                switch($c) {
                                        case "<":
                                                $hex = "";
                                                $isHex = true;
                                        break;
                                        case ">":
                                                $hexs = str_split($hex, $this->multibyte); // 2 or 4 (UTF8 or ISO)
                                                for ($k = 0; $k < count($hexs); $k++) {
                                                        $chex = str_pad($hexs[$k], 4, "0"); // Add tailing zero
                                                        if (isset($transformations[$chex]))
                                                                $chex = $transformations[$chex];
                                                        $document .= html_entity_decode("&#x".$chex.";");
                                                }
                                                $isHex = false;
                                        break;
                                        case "(":
                                                $plain = "";
                                                $isPlain = true;
                                        break;
                                        case ")":
                                                $document .= $plain;
                                                $isPlain = false;
                                        break;
                                        case "\\":
                                                $c2 = $texts[$i][$j + 1];
                                                if (in_array($c2, array("\\", "(", ")"))) $plain .= $c2;
                                                elseif ($c2 == "n") $plain .= '\n';
                                                elseif ($c2 == "r") $plain .= '\r';
                                                elseif ($c2 == "t") $plain .= '\t';
                                                elseif ($c2 == "b") $plain .= '\b';
                                                elseif ($c2 == "f") $plain .= '\f';
                                                elseif ($c2 >= '0' && $c2 <= '9') {
                                                        $oct = preg_replace("#[^0-9]#", "", substr($texts[$i], $j + 1, 3));
                                                        $j += strlen($oct) - 1;
                                                        $plain .= html_entity_decode("&#".octdec($oct).";", $this->convertquotes);
                                                }
                                                $j++;
                                        break;
        
                                        default:
                                                if ($isHex)
                                                        $hex .= $c;
                                                if ($isPlain)
                                                        $plain .= $c;
                                        break;
                                }
                        }
                        $document .= "\n";
                }
        
                return $document;
        }
}
?>

Uso:

<?php
//Arindo o PDF do diário da justiça e mostrando
include('class.pdf2text.php');
$a = new PDF2Text();
$a->setFilename('http://www.stf.jus.br/portal/diarioJusticaEletronico/verDiarioEletronico.asp?seq=1746305&data=16/02/2012&ano=2012&numero=35');
$a->decodePDF();
echo $a->output(); 

?>

E aí? Como será que eu recorto somente o parágrafo que contém o meu nome? Antemão agradeço as colaborações.

[Markus Magnus]

Vééééi do céu te juro que eu tentei fazer essa função funcionar na minha máquina, mas com pdfs meus da erro e com o link que você mandou demora de mais, mas enfim... tava olhando a classe e pelo que entendi ele troca as quebras de linha por "\n" e se você procurar por "\n" ou "\n\n" antes e depois da palavra achada

tipo assim:

$pdf = $a->output();
$pdf2 = explode($meunome, $pdf);
if(is_array($pdf2)) {
    foreach($pdf2 as $i => $pdf3) {
        if(isset($pdf2[$i+1])) {
            $iniciop = strrpos($pdf2[$i], "\n\n");
            $fimp = strpos($pdf, "\n\n", $iniciop);
            $textoemail .= "...\n\n...substr($pdf, $iniciop, $fimp-$inciop);
        }
    }
}

Espero ter ajudado

Abrass

[3l3tr1c]

E ai cara! Obrigado pela participação. A demora é para baixar o arquivo. E eu pensei em quebrar nos "\n", mas ele quebra todo final de linha, como será que o PDF processa um "parágrafo"? Valeu!

Outra coisa, eu preciso "cortar o parágrafo" e não a próxima palavra depois do meu nome, entende? Se eu achar o meu nome no parágrafo eu começo a cortá-lo do início e termino no final!

[Markus Magnus]

A intenção da rotina que eu tinha feito era justamente cortar o paragrafo

$pdf = $a->output();
$pdf2 = explode($meunome, $pdf);//corto o pdf a cada vez que acho $meunome
if(is_array($pdf2)) {//se tiver achado pelo menos uma vez $meunome
    foreach($pdf2 as $i => $pdf3) {
        if(isset($pdf2[$i+1])) {//verifico se não é o último registro que estou acessando
            $iniciop = strrpos($pdf2[$i], "\n\n");//encontro a ultima ocorrência de "\n\n" (considerando que seja isso que divida os paragrafos) antes da posição de $meunome
            $fimp = strpos($pdf, "\n\n", $iniciop);//encontro a primeira ocorrência de "\n\n" depois de $meunome
            $textoemail .= "...\n\n...substr($pdf, $iniciop, $fimp-$inciop);//corto $pdf nas posições que achei acima
        }
    }
}

Puis o script pra rodar agora quando terminar vou tentar analizar a saida se tiver novidades volto a postar

Abrass

[3l3tr1c]

Olá! Eu usei a sua rotina, mas não deu certo. Tem algumas coisas estranhas nelas, e.g., se usamos a função EXPLODE para quebrar o texto quando acho o Nome como iremos saber onde estava o início do parágrafo?