Merge branch 'feature/pdf-ids' into develop

 [SVCS-699]
 Closes: #338

mfr/core/extension.py

@@ -5,7 +5,7 @@
 
 class BaseExporter(metaclass=abc.ABCMeta):
 
-    def __init__(self, ext, source_file_path, output_file_path, format):
+    def __init__(self, ext, source_file_path, output_file_path, format, metadata):
 
         """Initialize the base exporter.
 
@@ -19,6 +19,7 @@ def __init__(self, ext, source_file_path, output_file_path, format):
         self.source_file_path = source_file_path
         self.output_file_path = output_file_path
         self.format = format
+        self.metadata = metadata
         self.exporter_metrics = MetricsRecord('exporter')
         if self._get_module_name():
             self.metrics = self.exporter_metrics.new_subrecord(self._get_module_name())

mfr/core/utils.py

@@ -33,7 +33,7 @@ def make_provider(name, request, url):
         )
 
 
-def make_exporter(name, source_file_path, output_file_path, format):
+def make_exporter(name, source_file_path, output_file_path, format, metadata):
     """Returns an instance of :class:`mfr.core.extension.BaseExporter`
 
     :param str name: The name of the extension to instantiate. (.jpg, .docx, etc)
@@ -49,7 +49,7 @@ def make_exporter(name, source_file_path, output_file_path, format):
             namespace='mfr.exporters',
             name=normalized_name,
             invoke_on_load=True,
-            invoke_args=(normalized_name, source_file_path, output_file_path, format),
+            invoke_args=(normalized_name, source_file_path, output_file_path, format, metadata),
         ).driver
     except RuntimeError:
         raise exceptions.MakeExporterError(

mfr/extensions/unoconv/export.py

@@ -1,6 +1,12 @@
 import os
 import subprocess
 
+from pdfrw import (
+    PdfReader,
+    PdfWriter
+)
+
+
 from mfr.core import extension
 from mfr.core import exceptions
 
@@ -20,6 +26,7 @@ def export(self):
                 '-vvv',
                 self.source_file_path
             ], check=True, timeout=settings.UNOCONV_TIMEOUT)
+
         except subprocess.CalledProcessError as err:
             name, extension = os.path.splitext(os.path.split(self.source_file_path)[-1])
             raise exceptions.SubprocessError(
@@ -32,3 +39,8 @@ def export(self):
                 extension=extension or '',
                 exporter_class='unoconv',
             )
+
+        pdf = PdfReader(self.output_file_path)
+        pdf.ID[0] = self.metadata.unique_key
+        pdf.ID[1] = self.metadata.unique_key
+        PdfWriter(self.output_file_path, trailer=pdf).write()

mfr/server/handlers/export.py

@@ -31,7 +31,7 @@ async def prepare(self):
         self.format = format[0].decode('utf-8')
         self.exporter_name = utils.get_exporter_name(self.metadata.ext)
 
-        self.cache_file_id = '{}.{}'.format(self.metadata.unique_key, self.format)
+        self.cache_file_id = '{}.{}'.format(self.self.metadata.unique_key, self.format)
 
         if self.exporter_name:
             cache_file_path_str = '/export/{}.{}'.format(self.cache_file_id, self.exporter_name)
@@ -87,7 +87,8 @@ async def get(self):
             self.metadata.ext,
             self.source_file_path.full_path,
             self.output_file_path.full_path,
-            self.format
+            self.format,
+            self.metadata,
         )
 
         self.extension_metrics.add('class', exporter._get_module_name())

requirements.txt

@@ -37,6 +37,7 @@ mistune==0.7
 
 # Pdf
 reportlab==3.4.0
+pdfrw==0.4.0
 
 # Pptx
 # python-pptx==0.5.7

tests/extensions/image/test_exporter.py

@@ -35,7 +35,8 @@ def test_jpg(self, directory, file_name, tolerance):
         output_file_path = os.path.join(directory, 'test.{}'.format(settings.EXPORT_TYPE))
         format = '{}.{}'.format(settings.EXPORT_MAXIMUM_SIZE, settings.EXPORT_TYPE)
         exporter = ImageExporter(source_file_path=source_file_path, ext='.jpg',
-                                 output_file_path=output_file_path, format=format)
+                                 output_file_path=output_file_path, format=format,
+                                 metadata={})
 
         assert not os.path.exists(output_file_path)
 
@@ -64,7 +65,8 @@ def test_png_with_transparency(self, directory):
         output_file_path = os.path.join(directory, 'test.{}'.format(settings.EXPORT_TYPE))
         format = '{}.{}'.format(settings.EXPORT_MAXIMUM_SIZE, settings.EXPORT_TYPE)
         exporter = ImageExporter(source_file_path=source_file_path, ext='.png',
-                                 output_file_path=output_file_path, format=format)
+                                 output_file_path=output_file_path, format=format,
+                                 metadata={})
 
         assert not os.path.exists(output_file_path)
 
@@ -98,7 +100,8 @@ def test_bmp(self, directory):
         output_file_path = os.path.join(directory, 'test.{}'.format(settings.EXPORT_TYPE))
         format = '{}.{}'.format(settings.EXPORT_MAXIMUM_SIZE, settings.EXPORT_TYPE)
         exporter = ImageExporter(source_file_path=source_file_path, ext='.bmp',
-                                 output_file_path=output_file_path, format=format)
+                                 output_file_path=output_file_path, format=format,
+                                 metadata={})
 
         assert not os.path.exists(output_file_path)
 
@@ -123,7 +126,8 @@ def test_ratio(self, directory):
         output_file_path = os.path.join(directory, 'test.{}'.format(settings.EXPORT_TYPE))
         format = '{}.{}'.format(settings.EXPORT_MAXIMUM_SIZE, settings.EXPORT_TYPE)
         exporter = ImageExporter(source_file_path=source_file_path, ext='.png',
-                                 output_file_path=output_file_path, format=format)
+                                 output_file_path=output_file_path, format=format,
+                                 metadata={})
 
         assert not os.path.exists(output_file_path)
 
@@ -146,7 +150,8 @@ def test_exception_file_not_found(self, directory):
                                         'test.{}'.format(settings.EXPORT_TYPE))
         format = '{}.{}'.format(settings.EXPORT_MAXIMUM_SIZE, settings.EXPORT_TYPE)
         exporter = ImageExporter(source_file_path=source_file_path, ext='.jpg',
-                                 output_file_path=output_file_path, format=format)
+                                 output_file_path=output_file_path, format=format,
+                                 metadata={})
 
         assert not os.path.exists(output_file_path)
         with pytest.raises(exceptions.PillowImageError) as e:
@@ -159,7 +164,8 @@ def test_exception_courrupt_file(self, directory):
         output_file_path = os.path.join(directory, 'test.{}'.format(settings.EXPORT_TYPE))
         format = '{}.{}'.format(settings.EXPORT_MAXIMUM_SIZE, settings.EXPORT_TYPE)
         exporter = ImageExporter(source_file_path=source_file_path, ext='.jpg',
-                                 output_file_path=output_file_path, format=format)
+                                 output_file_path=output_file_path, format=format,
+                                 metadata={})
 
         assert not os.path.exists(output_file_path)
         with pytest.raises(exceptions.PillowImageError) as e:

tests/extensions/pdf/test_exporter.py

@@ -45,7 +45,8 @@ def test_single_page_tiff(self, directory, file_name):
         output_file_path = os.path.join(directory, 'test.{}'.format(settings.EXPORT_TYPE))
         format = '{}.{}'.format(settings.EXPORT_MAXIMUM_SIZE, settings.EXPORT_TYPE)
         exporter = PdfExporter(source_file_path=source_file_path, ext='.tif',
-                               output_file_path=output_file_path, format=format)
+                               output_file_path=output_file_path, format=format,
+                               metadata={'unique_key': 'moo moo moo'})
 
         assert not os.path.exists(output_file_path)
 
@@ -62,7 +63,8 @@ def test_bad_tiff(self, directory):
         output_file_path = os.path.join(directory, 'test.{}'.format(settings.EXPORT_TYPE))
         format = '{}.{}'.format(settings.EXPORT_MAXIMUM_SIZE, settings.EXPORT_TYPE)
         exporter = PdfExporter(source_file_path=source_file_path, ext='.tif',
-                               output_file_path=output_file_path, format=format)
+                               output_file_path=output_file_path, format=format,
+                               metadata={'unique_key': 'moo moo moo'})
 
         assert not os.path.exists(output_file_path)