From eea34bdfaa325b43a98ccbf592acf1092a9f6325 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Denis=20Krienb=C3=BChl?= <denis.krienbuehl@seantis.ch>
Date: Wed, 3 Jun 2015 15:38:08 +0200
Subject: [PATCH] Stores a checksum with each form definition and submission.

---
 HISTORY.rst                          |  3 +++
 onegov/form/models.py                | 22 ++++++++++++++++++++++
 onegov/form/tests/test_collection.py | 11 ++++++++++-
 setup.py                             |  2 ++
 4 files changed, 37 insertions(+), 1 deletion(-)

diff --git a/HISTORY.rst b/HISTORY.rst
index 2323817..6cc951b 100644
--- a/HISTORY.rst
+++ b/HISTORY.rst
@@ -4,6 +4,9 @@ Changelog
 Unreleased
 ~~~~~~~~~~
 
+- Stores a checksum with each form definition and submission.
+  [href]
+
 - Adds the ability to filter out submissions older than one hour.
   [href]
 
diff --git a/onegov/form/models.py b/onegov/form/models.py
index 6e713cc..cd28812 100644
--- a/onegov/form/models.py
+++ b/onegov/form/models.py
@@ -1,3 +1,4 @@
+from hashlib import md5
 from onegov.core.orm import Base
 from onegov.core.orm.mixins import TimestampMixin
 from onegov.core.orm.types import JSON, UUID
@@ -8,9 +9,14 @@
     deferred,
     relationship,
 )
+from sqlalchemy_utils import observes
 from uuid import uuid4
 
 
+def hash_definition(definition):
+    return md5(definition.encode('utf-8')).hexdigest()
+
+
 class FormDefinition(Base, TimestampMixin):
     """ Defines a form stored in the database. """
 
@@ -31,6 +37,10 @@ class FormDefinition(Base, TimestampMixin):
     #: the form as parsable string
     definition = Column(Text, nullable=False)
 
+    #: the checksum of the definition, forms and submissions with matching
+    #: checksums are guaranteed to have the exact same definition
+    checksum = Column(Text, nullable=False)
+
     #: the type of the form, this can be used to create custom polymorphic
     #: subclasses. See `<http://docs.sqlalchemy.org/en/improve_toc/\
     #: orm/extensions/declarative/inheritance.html>`_.
@@ -49,6 +59,10 @@ def form_class(self):
 
         return parse_form(self.definition)
 
+    @observes('definition')
+    def definition_observer(self, definition):
+        self.checksum = hash_definition(definition)
+
 
 class FormSubmission(Base, TimestampMixin):
     """ Defines a submitted form in the database. """
@@ -66,6 +80,10 @@ class FormSubmission(Base, TimestampMixin):
     #: want to keep the old form around just in case.
     definition = Column(Text, nullable=False)
 
+    #: the checksum of the definition, forms and submissions with matching
+    #: checksums are guaranteed to have the exact same definition
+    checksum = Column(Text, nullable=False)
+
     #: the submission data
     data = Column(JSON, nullable=False)
 
@@ -92,6 +110,10 @@ def form_class(self):
 
         return parse_form(self.definition)
 
+    @observes('definition')
+    def definition_observer(self, definition):
+        self.checksum = hash_definition(definition)
+
     def complete(self):
         """ Changes the state to 'complete', if the data is valid. """
 
diff --git a/onegov/form/tests/test_collection.py b/onegov/form/tests/test_collection.py
index 961a926..36ae2eb 100644
--- a/onegov/form/tests/test_collection.py
+++ b/onegov/form/tests/test_collection.py
@@ -4,7 +4,7 @@
 from delorean import Delorean
 from onegov.core.compat import BytesIO
 from onegov.form import FormCollection, PendingFormSubmission
-from onegov.form.models import FormSubmissionFile
+from onegov.form.models import FormSubmissionFile, hash_definition
 from onegov.form.errors import UnableToComplete
 from sqlalchemy.exc import IntegrityError
 from sqlalchemy.orm.exc import FlushError
@@ -14,6 +14,10 @@
 from wtforms.csrf.core import CSRF
 
 
+def test_form_checksum():
+    assert hash_definition('abc') == '900150983cd24fb0d6963f7d28e17f72'
+
+
 def test_add_form(session):
     collection = FormCollection(session)
 
@@ -36,6 +40,8 @@ def test_add_form(session):
             Last Name * = ___
         """))
 
+    assert form.checksum and form.checksum == hash_definition(form.definition)
+
 
 def test_submit_form(session):
     collection = FormCollection(session)
@@ -58,6 +64,9 @@ def test_submit_form(session):
     form = collection.definitions.by_name('tps-report')
     submission = form.submissions[0]
 
+    assert submission.checksum and submission.checksum == hash_definition(
+        form.definition) == hash_definition(submission.definition)
+
     stored_form = submission.form_class(data=submission.data)
 
     assert submitted_form.data == stored_form.data
diff --git a/setup.py b/setup.py
index 525248d..a7f37a4 100644
--- a/setup.py
+++ b/setup.py
@@ -34,6 +34,7 @@ def get_long_description():
     zip_safe=False,
     platforms='any',
     install_requires=[
+        'delorean',
         'humanize',
         'jsonpickle',
         'onegov.core>=0.3.5',
@@ -41,6 +42,7 @@ def get_long_description():
         'pyyaml',
         'python-magic',
         'python-stdnum',
+        'sqlalchemy_utils',
         'wtforms',
         'wtforms-components[color]',
         'unidecode'