fix netload and shareonline, captcha subsets for better recognition

5 files changed, 22 insertions, 34 deletions

diff --git a/module/plugins/captcha/LinksaveIn.py b/module/plugins/captcha/LinksaveIn.py
index d6f61e362..22b801273 100644
--- a/module/plugins/captcha/LinksaveIn.py
+++ b/module/plugins/captcha/LinksaveIn.py
@@ -118,20 +118,6 @@ class LinksaveIn(OCR):
         self.image = new
         self.pixels = self.image.load()
     
-    def run_tesser(self):
-        self.logger.debug("create tmp tif")
-        tmp = tempfile.NamedTemporaryFile(suffix=".tif")
-        self.logger.debug("create tmp txt")
-        tmpTxt = tempfile.NamedTemporaryFile(suffix=".txt")
-        self.logger.debug("save tiff")
-        self.image.save(tmp.name, 'TIFF')
-        self.logger.debug("run tesseract")
-        self.run(['tesseract', tmp.name, tmpTxt.name.replace(".txt", ""), "nobatch", self.data_dir+"tesser_conf"])
-        self.logger.debug("read txt")
-
-        with open(tmpTxt.name, 'r') as f:
-            self.result_captcha = f.read().replace("\n", "")
-
     def get_captcha(self, image):
         self.load_image(image)
         bg = self.get_bg()
@@ -147,7 +133,7 @@ class LinksaveIn(OCR):
         for n, letter in enumerate(letters):
             self.image = letter
             self.image.save(ocr.data_dir+"letter%d.png" % n)
-            self.run_tesser()
+            self.run_tesser(True, True, False, False)
             final += self.result_captcha
         
         return final
diff --git a/module/plugins/captcha/MegauploadCom.py b/module/plugins/captcha/MegauploadCom.py
index 374bcd678..da8ab2cb9 100644
--- a/module/plugins/captcha/MegauploadCom.py
+++ b/module/plugins/captcha/MegauploadCom.py
@@ -6,7 +6,7 @@ class MegauploadCom(OCR):
         
     def get_captcha(self, image):
         self.load_image(image)
-        self.run_tesser()
+        self.run_tesser(True, True, False, True)
         return self.result_captcha
 
 if __name__ == '__main__':
diff --git a/module/plugins/captcha/NetloadIn.py b/module/plugins/captcha/NetloadIn.py
index 9799a6a2b..c99a0744c 100644
--- a/module/plugins/captcha/NetloadIn.py
+++ b/module/plugins/captcha/NetloadIn.py
@@ -9,11 +9,7 @@ class NetloadIn(OCR):
         self.to_greyscale()
         self.clean(3)
         self.clean(3)
-        self.run_tesser()
-
-        self.correct({
-        ("$", "g"): "5",
-        })
+        self.run_tesser(True, True, False, False)
 
         return self.result_captcha
 
diff --git a/module/plugins/captcha/ShareonlineBiz.py b/module/plugins/captcha/ShareonlineBiz.py
index 91124f181..7bd5d7960 100644
--- a/module/plugins/captcha/ShareonlineBiz.py
+++ b/module/plugins/captcha/ShareonlineBiz.py
@@ -37,19 +37,9 @@ class ShareonlineBiz(OCR):
         final = ""
         for letter in letters:
             self.image = letter
-            self.run_tesser()
+            self.run_tesser(True, True, False, False)
             final += self.result_captcha
 
-        #replace common errors
-        final = self.correct({
-        "A": "4",
-        "‘5": "3",
-        ("‘1", "T"): "7",
-        ("‘L", "B", "'L"): "2",
-        "b": "6",
-        ("I", "X"): "1"
-        }, final)
-
         return final
 
         #tesseract at 60%
diff --git a/module/plugins/captcha/captcha.py b/module/plugins/captcha/captcha.py
index 283b171e0..452952533 100644
--- a/module/plugins/captcha/captcha.py
+++ b/module/plugins/captcha/captcha.py
@@ -82,15 +82,31 @@ class OCR(object):
         self.image.save(tmp)
         self.result_captcha = self.run(['gocr', tmp.name]).replace("\n", "")
 
-    def run_tesser(self):
+    def run_tesser(self, subset=False, digits=True, lowercase=True, uppercase=True ):
         self.logger.debug("create tmp tif")
         tmp = tempfile.NamedTemporaryFile(suffix=".tif")
         self.logger.debug("create tmp txt")
         tmpTxt = tempfile.NamedTemporaryFile(suffix=".txt")
         self.logger.debug("save tiff")
         self.image.save(tmp.name, 'TIFF')
+
+        tessparams = ['tesseract', tmp.name, tmpTxt.name.replace(".txt", "")
+
+        if subset and (digits or lowercase or uppercase):
+            self.logger.debug("create temp subset config")
+            tmpSub = tempfile.NamedTemporaryFile(suffix=".subset")
+            tmpSub.write("tessedit_char_whitelist ")
+            if digits:
+                tmpSub.write("0123456789")
+            if lowercase:
+                tmpSub.write("abcdefghijklmnopqrstuvwxyz")
+            if uppercase:
+                tmpSub.write("ABCDEFGHIJKLMNOPQRSTUVWXYZ")
+            tessparams.append("nobatch")
+            tessparams.append(tmpSub.name)
+
         self.logger.debug("run tesseract")
-        self.run(['tesseract', tmp.name, tmpTxt.name.replace(".txt", "")])
+        self.run(tessparams)
         self.logger.debug("read txt")
 
         with open(tmpTxt.name, 'r') as f: