From 11701391279b17995bd2becd7cc742f8ba3ff3aa Mon Sep 17 00:00:00 2001
From: Daniel Quinn <code@danielquinn.org>
Date: Mon, 14 Mar 2016 21:20:44 +0000
Subject: [PATCH] Added a consume-start and consume-finish signal

---
 src/documents/consumer.py | 19 +++++++++++++++++--
 src/documents/signals.py  |  4 ++++
 2 files changed, 21 insertions(+), 2 deletions(-)
 create mode 100644 src/documents/signals.py

diff --git a/src/documents/consumer.py b/src/documents/consumer.py
index fbdbbc276..244383211 100644
--- a/src/documents/consumer.py
+++ b/src/documents/consumer.py
@@ -26,6 +26,8 @@ from paperless.db import GnuPG
 
 from .models import Correspondent, Tag, Document, Log
 from .languages import ISO639
+from .signals import (
+    document_consumption_started, document_consumption_finished)
 
 
 class OCRError(Exception):
@@ -118,22 +120,33 @@ class Consumer(object):
 
             self.log("info", "Consuming {}".format(doc))
 
+            document_consumption_started.send(
+                sender=self.__class__, filename=doc)
+
             tempdir = tempfile.mkdtemp(prefix="paperless", dir=self.SCRATCH)
             imgs = self._get_greyscale(tempdir, doc)
             thumbnail = self._get_thumbnail(tempdir, doc)
 
             try:
-                text = self._get_ocr(imgs)
-                self._store(text, doc, thumbnail)
+
+                document = self._store(self._get_ocr(imgs), doc, thumbnail)
+
             except OCRError as e:
+
                 self._ignore.append(doc)
                 self.log("error", "OCR FAILURE for {}: {}".format(doc, e))
                 self._cleanup_tempdir(tempdir)
+
                 continue
+
             else:
+
                 self._cleanup_tempdir(tempdir)
                 self._cleanup_doc(doc)
 
+                document_consumption_finished.send(
+                    sender=self.__class__, filename=document)
+
     def _get_greyscale(self, tempdir, doc):
         """
         Greyscale images are easier for Tesseract to OCR
@@ -360,6 +373,8 @@ class Consumer(object):
 
         self.log("info", "Completed")
 
+        return document
+
     def _cleanup_tempdir(self, d):
         self.log("debug", "Deleting directory {}".format(d))
         shutil.rmtree(d)
diff --git a/src/documents/signals.py b/src/documents/signals.py
new file mode 100644
index 000000000..257a20d46
--- /dev/null
+++ b/src/documents/signals.py
@@ -0,0 +1,4 @@
+from django.dispatch import Signal
+
+document_consumption_started = Signal(providing_args=["filename"])
+document_consumption_finished = Signal(providing_args=["document"])