From 1744840c2f94cfb7372e0959039e9d4f7f15800e Mon Sep 17 00:00:00 2001
From: Vincent Vanwaelscappel <vincent@cubedesigners.com>
Date: Wed, 20 Oct 2021 09:52:53 +0200
Subject: [PATCH] wip #4804 @0.5

---
 src/Document.php | 4 ++--
 src/PDFTools.php | 6 +++++-
 2 files changed, 7 insertions(+), 3 deletions(-)

diff --git a/src/Document.php b/src/Document.php
index b6683e1..c3a006e 100644
--- a/src/Document.php
+++ b/src/Document.php
@@ -157,9 +157,9 @@ class Document
         PDFTools::extractHighlightsData($this->getPDFInput(), $this->getConvertPath());
     }
 
-    public function processFullTexts()
+    public function processFullTexts($force = true)
     {
-        PDFTools::extractTexts($this->getPDFInput(), $this->getConvertPath());
+        PDFTools::extractTexts($this->getPDFInput(), $this->getConvertPath(), '', $force);
     }
 
     public function getPageNumbers()
diff --git a/src/PDFTools.php b/src/PDFTools.php
index 8469b0c..4c88e3a 100644
--- a/src/PDFTools.php
+++ b/src/PDFTools.php
@@ -326,11 +326,15 @@ class PDFTools
         $fwstk->execute();
     }
 
-    public static function extractTexts($pdf, $out, $ignoreSeparators = '')
+    public static function extractTexts($pdf, $out, $ignoreSeparators = '', $force = true)
     {
         $out .= 'texts/';
         Files::mkdir($out);
 
+        if (!$force && file_exists($out . 'fp1.txt')) {
+            return;
+        }
+
         $fwstk = new FWSTK();
         $fwstk->setArg('--input ' . $pdf);
         $fwstk->setArg('--extractTexts ' . $out . '%s%d.txt');
-- 
2.39.5