feat: refactor for pluggable consumers

I've broken out the OCR-specific code from the consumers and dumped it
all into its own app, `paperless_tesseract`.  This new app should serve
as a sample of how to create one's own consumer for different file
types.

Documentation for how to do this isn't ready yet, but for the impatient:

* Create a new app
    * containing a `parsers.py` for your parser modelled after
      `paperless_tesseract.parsers.RasterisedDocumentParser`
    * containing a `signals.py` with a handler moddelled after
      `paperless_tesseract.signals.ConsumerDeclaration`
    * connect the signal handler to
      `documents.signals.document_consumer_declaration` in
      `your_app.apps`
* Install the app into Paperless by declaring
  `PAPERLESS_INSTALLED_APPS=your_app`.  Additional apps should be
  separated with commas.
* Restart the consumer

src/documents/tests/test_consumer.py

@@ -1,13 +1,6 @@
-import os
-from unittest import mock, skipIf
-
-import pyocr
 from django.test import TestCase
-from pyocr.libtesseract.tesseract_raw import \
-    TesseractError as OtherTesseractError
 
 from ..models import FileInfo
-from ..consumer import image_to_string, strip_excess_whitespace
 
 
 class TestAttributes(TestCase):
@@ -310,69 +303,3 @@ class TestFieldPermutations(TestCase):
                             template.format(**spec), **spec)
 
 
-class FakeTesseract(object):
-
-    @staticmethod
-    def can_detect_orientation():
-        return True
-
-    @staticmethod
-    def detect_orientation(file_handle, lang):
-        raise OtherTesseractError("arbitrary status", "message")
-
-    @staticmethod
-    def image_to_string(file_handle, lang):
-        return "This is test text"
-
-
-class FakePyOcr(object):
-
-    @staticmethod
-    def get_available_tools():
-        return [FakeTesseract]
-
-
-class TestOCR(TestCase):
-
-    text_cases = [
-        ("simple     string", "simple string"),
-        (
-            "simple    newline\n   testing string",
-            "simple newline\ntesting string"
-        ),
-        (
-            "utf-8   строка с пробелами в конце  ",
-            "utf-8 строка с пробелами в конце"
-        )
-    ]
-
-    SAMPLE_FILES = os.path.join(os.path.dirname(__file__), "samples")
-    TESSERACT_INSTALLED = bool(pyocr.get_available_tools())
-
-    def test_strip_excess_whitespace(self):
-        for source, result in self.text_cases:
-            actual_result = strip_excess_whitespace(source)
-            self.assertEqual(
-                result,
-                actual_result,
-                "strip_exceess_whitespace({}) != '{}', but '{}'".format(
-                    source,
-                    result,
-                    actual_result
-                )
-            )
-
-    @skipIf(not TESSERACT_INSTALLED, "Tesseract not installed. Skipping")
-    @mock.patch("documents.consumer.Consumer.SCRATCH", SAMPLE_FILES)
-    @mock.patch("documents.consumer.pyocr", FakePyOcr)
-    def test_image_to_string_with_text_free_page(self):
-        """
-        This test is sort of silly, since it's really just reproducing an odd
-        exception thrown by pyocr when it encounters a page with no text.
-        Actually running this test against an installation of Tesseract results
-        in a segmentation fault rooted somewhere deep inside pyocr where I
-        don't care to dig.  Regardless, if you run the consumer normally,
-        text-free pages are now handled correctly so long as we work around
-        this weird exception.
-        """
-        image_to_string(["no-text.png", "en"])