Chore rebase on main (#1679)

* chore(deps): update graphene-django

Update to graphene-django 2.12.1 and fix connection field, schema.

* chore(deps): add pytest-xfaillist plugin

* chore(deps): switch to graphql 3.0

* chore: fix ResolveInfo fixture

* chore(test): ensure name is valid identifier

* chore(deps): remove XPASSing tests from xfails

* chore(deps): handle enum in JSONValueFilter

* chore(deps): remove passing tests from xfails.list

* fix(tests): hand over UUID as string

* chore(deps): get ordering value from Enum if any

* chore(deps): use graphene's updated data types

issue #814

* chore(deps): update and sort xfaillist

* chore(deps): remove passing test from xfaillist

#814

* chore(deps): handle enum in filter

#814

* check if these tests are correct

if so fixup commit with previous

* chore(deps): upgrade pytest-xfaillist

* chore(deps): pass kwargs to data source resolver (#1643)

* fix(graphene): make more tests work (#1644)

* Introduce our own choices field (for now): Graphene seems to already
  instantiate Enum instances from the request data, which causes the
  regular DRF choice field to fail, as the enum object isn't handled
* Fix a few resolvers which now receive additional (unused by use)
  parameters

* Graphene django v3 (#1645)

* fix(datasource): accept new additional parameters from graphene

* fix(filters): overload resolve_queryset() to support multi order_by

Graphene does not really support order_by with multiple values like we need it.

There's a bug in graphene that we need to work around:
 graphql-python/graphene-django#1280

While we're at it, also clean up some more data.

* fix(historical): fix uuid and resolver params (#1646)

* fix(historical): fix uuid and resolver params

* fix(form): accept new additional resolver args

* Graphene django v3 (#1647)

* fix(tests): adjust test to new error message style

graphene 3.0 returns a text fragment of the offending query

* fix(middleware): use new visitor and ast conventions

The node types are not CamelCase named, but snake_cased. Also,
the operation is an AST object, not a bare string.

In queries without variables, the variables MUST NOT be set to an empty
string. (None works however...)

* fix(cases): various fixes in csae filter and tests

* fix(schema): make schema tests work again (sort of)

The schema test relies on stringification of the root schema object.
This currently fails due to a function within the graphql core package
being unable to handle django lazy translation objects.

The test setup is rigged such that once upstream fixes the problem, it
will trigger a failure here and we can remove the hacky workaround.

* fix(tests): more tests fixed with correct (non-uuid) input objects

Also remove passing some info objects to schema_executor(), it doesn't
accept them anymore (just pass context_value as that's the interesting
bit)

* fix(tests): stringify localized fields when passing to gql

* chore(coverage): improve coverage

Most of the lines that were "not covered" will never happen or are error
cases such as raising validation errors, or other non-problematic parts.

* fix(tests): sort the schema type map (#1648)

A sorted type map (as it used to be in older graphene versions)
helps better comparing / diffing the schema

* Graphene django v3 (#1649)

* chore: coverage back to 100%

* chore: fix conflict in setup.py

* fix(schema): use explicit types for case status and workitem status

Graphene would otherwise use an automatic naming scheme that breaks our clients

* fix: coverage in new status fields (#1650)

Those status fields are only used for the type system, but not as actual
input types. Thus, they're never actually instantiated

* rebase graphene update branch on main (#1664)

* feat(form): enable hint messages on questions (#1655)

Add a hint text field on all question types except for
form, static and action button questions. This functionality
requires a db migration.

* feat(workflow): implement WorkItem redo pattern (#1656)

This commit implements a possibility to redo an already finished
WorkItem and all the ones that were finished in between.

Closes #1510

* chore: update dependencies (#1660)

* chore: cleanup `WorkItem.get_redoable()`

* chore: update dependencies

* chore: release 7.15 (#1661)

* chore: add a code of conduct (#1651)

* chore(deps): update graphene-django

Update to graphene-django 2.12.1 and fix connection field, schema.

* chore(deps): add pytest-xfaillist plugin

* chore(deps): switch to graphql 3.0

* chore: fix ResolveInfo fixture

* chore(test): ensure name is valid identifier

* chore(deps): remove XPASSing tests from xfails

* chore(deps): handle enum in JSONValueFilter

* chore(deps): remove passing tests from xfails.list

* fix(tests): hand over UUID as string

* chore(deps): get ordering value from Enum if any

* chore(deps): use graphene's updated data types

issue #814

* chore(deps): update and sort xfaillist

* chore(deps): remove passing test from xfaillist

#814

* chore(deps): handle enum in filter

#814

* check if these tests are correct

if so fixup commit with previous

* chore(deps): upgrade pytest-xfaillist

* chore(deps): pass kwargs to data source resolver (#1643)

* fix(graphene): make more tests work (#1644)

* Introduce our own choices field (for now): Graphene seems to already
  instantiate Enum instances from the request data, which causes the
  regular DRF choice field to fail, as the enum object isn't handled
* Fix a few resolvers which now receive additional (unused by use)
  parameters

* Graphene django v3 (#1645)

* fix(datasource): accept new additional parameters from graphene

* fix(filters): overload resolve_queryset() to support multi order_by

Graphene does not really support order_by with multiple values like we need it.

There's a bug in graphene that we need to work around:
 graphql-python/graphene-django#1280

While we're at it, also clean up some more data.

* fix(historical): fix uuid and resolver params (#1646)

* fix(historical): fix uuid and resolver params

* fix(form): accept new additional resolver args

* Graphene django v3 (#1647)

* fix(tests): adjust test to new error message style

graphene 3.0 returns a text fragment of the offending query

* fix(middleware): use new visitor and ast conventions

The node types are not CamelCase named, but snake_cased. Also,
the operation is an AST object, not a bare string.

In queries without variables, the variables MUST NOT be set to an empty
string. (None works however...)

* fix(cases): various fixes in csae filter and tests

* fix(schema): make schema tests work again (sort of)

The schema test relies on stringification of the root schema object.
This currently fails due to a function within the graphql core package
being unable to handle django lazy translation objects.

The test setup is rigged such that once upstream fixes the problem, it
will trigger a failure here and we can remove the hacky workaround.

* fix(tests): more tests fixed with correct (non-uuid) input objects

Also remove passing some info objects to schema_executor(), it doesn't
accept them anymore (just pass context_value as that's the interesting
bit)

* fix(tests): stringify localized fields when passing to gql

* chore(coverage): improve coverage

Most of the lines that were "not covered" will never happen or are error
cases such as raising validation errors, or other non-problematic parts.

* fix(tests): sort the schema type map (#1648)

A sorted type map (as it used to be in older graphene versions)
helps better comparing / diffing the schema

* Graphene django v3 (#1649)

* chore: coverage back to 100%

* chore: fix conflict in setup.py

* fix(schema): use explicit types for case status and workitem status

Graphene would otherwise use an automatic naming scheme that breaks our clients

* fix: coverage in new status fields (#1650)

Those status fields are only used for the type system, but not as actual
input types. Thus, they're never actually instantiated

* fix: pin graphql-relay to <3.1.1

the minor version includes deprecation of backwards compatibility
measures.

See graphql-python/graphql-relay-py#45

Might include other changes that are unrelated to the above
that break a number of our tests.

* update snapshots for test_question

* fix(test): fix UUID

* fix(tests): update test_workflow snapshot

* fix: remove schema workaround and update snapshot

Co-authored-by: luytena <[email protected]>
Co-authored-by: Fabio Ambauen <[email protected]>
Co-authored-by: David Vogt <[email protected]>
Co-authored-by: Christian Zosel <[email protected]>
Co-authored-by: Stefan Borer <[email protected]>
Co-authored-by: Jean-Louis Fuchs <[email protected]>

* chore: cleanup schema test workaround

* fix(coverage): raise cov and fix a typo

* chore(deps): update graphene-django

 * chore(deps): Update to graphene-django 2.12.1 and fix connection field, schema.

 * chore(deps): add pytest-xfaillist plugin

 * chore(deps): switch to graphql 3.0

* chore: fix ResolveInfo fixture

* chore(test): ensure name is valid identifier

* chore(deps): handle enum in JSONValueFilter

* fix(tests): hand over UUID as string

* chore(deps): get ordering value from Enum if any

* chore(deps): use graphene's updated data types

issue #814

* chore(deps): handle enum in filter

#814

* chore(deps): pass kwargs to data source resolver (#1643)

* fix(graphene): make more tests work (#1644)

* Introduce our own choices field (for now): Graphene seems to already
  instantiate Enum instances from the request data, which causes the
  regular DRF choice field to fail, as the enum object isn't handled
* Fix a few resolvers which now receive additional (unused by use)
  parameters

* Graphene django v3 (#1645)

* fix(datasource): accept new additional parameters from graphene

* fix(filters): overload resolve_queryset() to support multi order_by

Graphene does not really support order_by with multiple values like we need it.

There's a bug in graphene that we need to work around:
 graphql-python/graphene-django#1280

While we're at it, also clean up some more data.

* fix(historical): fix uuid and resolver params (#1646)

* fix(historical): fix uuid and resolver params

* fix(form): accept new additional resolver args

* Graphene django v3 (#1647)

* fix(tests): adjust test to new error message style

graphene 3.0 returns a text fragment of the offending query

* fix(middleware): use new visitor and ast conventions

The node types are not CamelCase named, but snake_cased. Also,
the operation is an AST object, not a bare string.

In queries without variables, the variables MUST NOT be set to an empty
string. (None works however...)

* fix(cases): various fixes in csae filter and tests

* fix(schema): make schema tests work again (sort of)

The schema test relies on stringification of the root schema object.
This currently fails due to a function within the graphql core package
being unable to handle django lazy translation objects.

The test setup is rigged such that once upstream fixes the problem, it
will trigger a failure here and we can remove the hacky workaround.

* fix(tests): more tests fixed with correct (non-uuid) input objects

Also remove passing some info objects to schema_executor(), it doesn't
accept them anymore (just pass context_value as that's the interesting
bit)

* fix(tests): stringify localized fields when passing to gql

* chore(coverage): improve coverage

Most of the lines that were "not covered" will never happen or are error
cases such as raising validation errors, or other non-problematic parts.

* fix(tests): sort the schema type map (#1648)

A sorted type map (as it used to be in older graphene versions)
helps better comparing / diffing the schema

* Graphene django v3 (#1649)

* chore: coverage back to 100%

* chore: fix conflict in setup.py

* fix(schema): use explicit types for case status and workitem status

Graphene would otherwise use an automatic naming scheme that breaks our clients

* fix: coverage in new status fields (#1650)

Those status fields are only used for the type system, but not as actual
input types. Thus, they're never actually instantiated

* rebase graphene update branch on main (#1664)

* feat(form): enable hint messages on questions (#1655)

Add a hint text field on all question types except for
form, static and action button questions. This functionality
requires a db migration.

* feat(workflow): implement WorkItem redo pattern (#1656)

This commit implements a possibility to redo an already finished
WorkItem and all the ones that were finished in between.

Closes #1510

* chore: update dependencies (#1660)

* chore: cleanup `WorkItem.get_redoable()`

* chore: update dependencies

* chore: release 7.15 (#1661)

* chore: add a code of conduct (#1651)

* chore(deps): update graphene-django

Update to graphene-django 2.12.1 and fix connection field, schema.

* chore(deps): add pytest-xfaillist plugin

* chore(deps): switch to graphql 3.0

* chore: fix ResolveInfo fixture

* chore(test): ensure name is valid identifier

* chore(deps): remove XPASSing tests from xfails

* chore(deps): handle enum in JSONValueFilter

* chore(deps): remove passing tests from xfails.list

* fix(tests): hand over UUID as string

* chore(deps): get ordering value from Enum if any

* chore(deps): use graphene's updated data types

issue #814

* chore(deps): update and sort xfaillist

* chore(deps): remove passing test from xfaillist

#814

* chore(deps): handle enum in filter

#814

* check if these tests are correct

if so fixup commit with previous

* chore(deps): upgrade pytest-xfaillist

* chore(deps): pass kwargs to data source resolver (#1643)

* fix(graphene): make more tests work (#1644)

* Introduce our own choices field (for now): Graphene seems to already
  instantiate Enum instances from the request data, which causes the
  regular DRF choice field to fail, as the enum object isn't handled
* Fix a few resolvers which now receive additional (unused by use)
  parameters

* Graphene django v3 (#1645)

* fix(datasource): accept new additional parameters from graphene

* fix(filters): overload resolve_queryset() to support multi order_by

Graphene does not really support order_by with multiple values like we need it.

There's a bug in graphene that we need to work around:
 graphql-python/graphene-django#1280

While we're at it, also clean up some more data.

* fix(historical): fix uuid and resolver params (#1646)

* fix(historical): fix uuid and resolver params

* fix(form): accept new additional resolver args

* Graphene django v3 (#1647)

* fix(tests): adjust test to new error message style

graphene 3.0 returns a text fragment of the offending query

* fix(middleware): use new visitor and ast conventions

The node types are not CamelCase named, but snake_cased. Also,
the operation is an AST object, not a bare string.

In queries without variables, the variables MUST NOT be set to an empty
string. (None works however...)

* fix(cases): various fixes in csae filter and tests

* fix(schema): make schema tests work again (sort of)

The schema test relies on stringification of the root schema object.
This currently fails due to a function within the graphql core package
being unable to handle django lazy translation objects.

The test setup is rigged such that once upstream fixes the problem, it
will trigger a failure here and we can remove the hacky workaround.

* fix(tests): more tests fixed with correct (non-uuid) input objects

Also remove passing some info objects to schema_executor(), it doesn't
accept them anymore (just pass context_value as that's the interesting
bit)

* fix(tests): stringify localized fields when passing to gql

* chore(coverage): improve coverage

Most of the lines that were "not covered" will never happen or are error
cases such as raising validation errors, or other non-problematic parts.

* fix(tests): sort the schema type map (#1648)

A sorted type map (as it used to be in older graphene versions)
helps better comparing / diffing the schema

* Graphene django v3 (#1649)

* chore: coverage back to 100%

* chore: fix conflict in setup.py

* fix(schema): use explicit types for case status and workitem status

Graphene would otherwise use an automatic naming scheme that breaks our clients

* fix: coverage in new status fields (#1650)

Those status fields are only used for the type system, but not as actual
input types. Thus, they're never actually instantiated

* fix: pin graphql-relay to <3.1.1

the minor version includes deprecation of backwards compatibility
measures.

See graphql-python/graphql-relay-py#45

Might include other changes that are unrelated to the above
that break a number of our tests.

* update snapshots for test_question

* fix(test): fix UUID

* fix(tests): update test_workflow snapshot

* fix: remove schema workaround and update snapshot

Co-authored-by: luytena <[email protected]>
Co-authored-by: Fabio Ambauen <[email protected]>
Co-authored-by: David Vogt <[email protected]>
Co-authored-by: Christian Zosel <[email protected]>
Co-authored-by: Stefan Borer <[email protected]>
Co-authored-by: Jean-Louis Fuchs <[email protected]>

* chore: cleanup schema test workaround

* feat(form): enable hint messages on questions (#1655)

Add a hint text field on all question types except for
form, static and action button questions. This functionality
requires a db migration.

* feat(workflow): implement WorkItem redo pattern (#1656)

This commit implements a possibility to redo an already finished
WorkItem and all the ones that were finished in between.

Closes #1510

* fix(coverage): raise cov and fix a typo

* fixme: question snapshot differs

* fixme: caluma analytics snapshot differs and tests fail

* fix: incomplete merge

* chore(analytics): compatibility cleanups

Co-authored-by: Stefan Borer <[email protected]>
Co-authored-by: Jean-Louis Fuchs <[email protected]>
Co-authored-by: David Vogt <[email protected]>
Co-authored-by: luytena <[email protected]>
Co-authored-by: Fabio Ambauen <[email protected]>
Co-authored-by: Christian Zosel <[email protected]>

Author: 7 people

.reuse/dep5

@@ -5,6 +5,8 @@ Source: https://github.com/projectcaluma/caluma
 
 # Files below the threshold of originality (metadata, config and project-information)
 Files:
+    deps/
+    xfails.list
     .dockerignore
     .editorconfig
     .github/dependabot.yml

caluma/caluma_analytics/schema.py

@@ -1,6 +1,7 @@
 import graphene
 from graphene import ConnectionField, String, relay
 from graphene.types import ObjectType, generic
+from graphene_django.rest_framework import serializer_converter
 
 from ..caluma_core.filters import (
     CollectionFilterSetFactory,
@@ -74,6 +75,16 @@ def resolve_records(table, info, *args, **kwargs):
         return rows
 
 
+StartingObject = graphene.Enum(
+    "StartingObject",
+    [(key.upper(), key) for key, _ in models.AnalyticsTable.STARTING_OBJECT_CHOICES],
+)
+
+serializer_converter.get_graphene_type_from_serializer_field.register(
+    serializers.StartingObjectField, lambda field: StartingObject
+)
+
+
 class AnalyticsTable(DjangoObjectType):
     available_fields = ConnectionField(
         AvailableFieldConnection,
@@ -83,10 +94,10 @@ class AnalyticsTable(DjangoObjectType):
     result_data = graphene.Field(AnalyticsOutput)
 
     @staticmethod
-    def resolve_available_fields(obj, info, prefix=None, depth=0):
+    def resolve_available_fields(obj, info, prefix=None, depth=None, **kwargs):
         start = obj.get_starting_object(info)
 
-        depth = depth if depth > 0 else 1
+        depth = depth if depth and depth > 0 else 1
         prefix = prefix.split(".") if prefix else []
 
         return [
@@ -117,7 +128,6 @@ class AnalyticsField(DjangoObjectType):
 
     class Meta:
         model = models.AnalyticsField
-        # exclude = ("cases", "task_flows")
         interfaces = (relay.Node,)
         connection_class = CountableConnectionBase
 

caluma/caluma_analytics/serializers.py

@@ -5,7 +5,14 @@
 from . import models
 
 
+class StartingObjectField(serializers.CalumaChoiceField):
+    def __init__(self, **kwargs):
+        super().__init__(models.AnalyticsTable.STARTING_OBJECT_CHOICES, **kwargs)
+
+
 class SaveAnalyticsTableSerializer(serializers.ModelSerializer):
+    starting_object = StartingObjectField(required=True)
+
     class Meta:
         model = models.AnalyticsTable
         fields = [

caluma/caluma_analytics/tests/__snapshots__/test_run_analytics_cmdline.ambr

@@ -1,38 +1,35 @@
 # name: test_cmdline_output[output_mode0-False-expect_output1]
   <class 'dict'> {
-    'data': '
-      case_id                                blablub          
-      ------------------------------------   ---------------  
-      case1pk   Shelly Watson    
-      case2pk   Ruben Ibarra     
-      case3pk   Katelyn Mcgrath  
-  
-    ',
+    'data': <class 'list'> [
+      '------------------------------------   ---------------  ',
+      'case1pk   Shelly Watson    ',
+      'case2pk   Ruben Ibarra     ',
+      'case3pk   Katelyn Mcgrath  ',
+      'case_id                                blablub          ',
+    ],
     'stderr': '',
   }
 ---
 # name: test_cmdline_output[output_mode0-True-expect_output0]
   <class 'dict'> {
-    'data': '
-      case_id                                blablub        
-      ------------------------------------   -------------  
-      case1pk   Shelly Watson  
-      case2pk   Ruben Ibarra   
-  
-    ',
+    'data': <class 'list'> [
+      '------------------------------------   -------------  ',
+      'case1pk   Shelly Watson  ',
+      'case2pk   Ruben Ibarra   ',
+      'case_id                                blablub        ',
+    ],
     'stderr': '',
   }
 ---
 # name: test_cmdline_output[output_mode1-False-expect_output1]
   <class 'dict'> {
-    'data': '
-      case_id                                blablub          
-      ------------------------------------   ---------------  
-      case1pk   Shelly Watson    
-      case2pk   Ruben Ibarra     
-      case3pk   Katelyn Mcgrath  
-  
-    ',
+    'data': <class 'list'> [
+      '------------------------------------   ---------------  ',
+      'case1pk   Shelly Watson    ',
+      'case2pk   Ruben Ibarra     ',
+      'case3pk   Katelyn Mcgrath  ',
+      'case_id                                blablub          ',
+    ],
     'stderr': '
       -- SQL: 
       WITH
@@ -115,13 +112,12 @@
 ---
 # name: test_cmdline_output[output_mode1-True-expect_output0]
   <class 'dict'> {
-    'data': '
-      case_id                                blablub        
-      ------------------------------------   -------------  
-      case1pk   Shelly Watson  
-      case2pk   Ruben Ibarra   
-  
-    ',
+    'data': <class 'list'> [
+      '------------------------------------   -------------  ',
+      'case1pk   Shelly Watson  ',
+      'case2pk   Ruben Ibarra   ',
+      'case_id                                blablub        ',
+    ],
     'stderr': '
       -- SQL: 
       WITH
@@ -207,7 +203,8 @@
 ---
 # name: test_cmdline_output[output_mode2-False-expect_output1]
   <class 'dict'> {
-    'data': '',
+    'data': <class 'list'> [
+    ],
     'stderr': '
       -- SQL: 
       WITH
@@ -290,7 +287,8 @@
 ---
 # name: test_cmdline_output[output_mode2-True-expect_output0]
   <class 'dict'> {
-    'data': '',
+    'data': <class 'list'> [
+    ],
     'stderr': '
       -- SQL: 
       WITH
@@ -378,16 +376,16 @@
   <class 'dict'> {
     'data': <class 'list'> [
       <class 'dict'> {
-        'blablub': 'Shelly Watson',
-        'case_id': 'case1pk',
+        'blablub': 'Katelyn Mcgrath',
+        'case_id': 'case3pk',
       },
       <class 'dict'> {
         'blablub': 'Ruben Ibarra',
         'case_id': 'case2pk',
       },
       <class 'dict'> {
-        'blablub': 'Katelyn Mcgrath',
-        'case_id': 'case3pk',
+        'blablub': 'Shelly Watson',
+        'case_id': 'case1pk',
       },
     ],
     'stderr': '',
@@ -396,14 +394,14 @@
 # name: test_cmdline_output[output_mode3-True-expect_output0]
   <class 'dict'> {
     'data': <class 'list'> [
-      <class 'dict'> {
-        'blablub': 'Shelly Watson',
-        'case_id': 'case1pk',
-      },
       <class 'dict'> {
         'blablub': 'Ruben Ibarra',
         'case_id': 'case2pk',
       },
+      <class 'dict'> {
+        'blablub': 'Shelly Watson',
+        'case_id': 'case1pk',
+      },
     ],
     'stderr': '',
   }