Fix doc for slicing and transformations

Author: Hartorn

giskard/core/core.py

@@ -77,6 +77,7 @@ def _get_plugin_method_full_name(func):
 
 def create_test_function_id(func):
     from giskard.ml_worker.testing.registry.registry import plugins_root
+
     is_relative = Path(inspect.getfile(func)).is_relative_to(plugins_root)
     if is_relative:
         full_name = _get_plugin_method_full_name(func)
@@ -273,6 +274,13 @@ def extract_code(self, callable_obj):
             logger.info(f"Failed to extract test function code {self.full_name}: %s" % e)
         return code
 
+    @staticmethod
+    def default_doc(description: str) -> str:
+        doc = CallableDocumentation()
+        doc.description = description
+        doc.parameters = {}
+        return json.dumps(doc.to_dict())
+
     @staticmethod
     def extract_doc(func) -> Optional[str]:
         if not func.__doc__:

giskard/scanner/robustness/text_transformations.py

@@ -26,7 +26,7 @@ def __init__(self, column):
         self.meta.name = self.name
         self.meta.display_name = self.name
         self.meta.tags = ["pickle", "scan"]
-        self.meta.doc = "Automatically generated transformation function"
+        self.meta.doc = self.meta.default_doc("Automatically generated transformation function")
 
     def __str__(self):
         return self.name

giskard/slicing/slice.py

@@ -1,4 +1,6 @@
 # @TODO: simplify this module, don’t need this complexity.
+from typing import Callable, Dict, List, Sequence
+
 import itertools
 import operator
 import re
@@ -8,7 +10,6 @@
 
 import numpy as np
 import pandas as pd
-from typing import Callable, Sequence, List, Dict
 
 from ..core.core import DatasetProcessFunctionMeta, DatasetProcessFunctionType
 from ..ml_worker.testing.registry.slicing_function import SlicingFunction
@@ -304,7 +305,7 @@ def __init__(self, query: Query):
         self.meta.name = str(self)
         self.meta.display_name = str(self)
         self.meta.tags = ["pickle", "scan"]
-        self.meta.doc = '{"description": "Automatically generated slicing function.", "parameters": {}}'
+        self.meta.doc = self.meta.default_doc("Automatically generated slicing function")
         self.meta.uuid = str(uuid.uuid5(uuid.NAMESPACE_OID, self.meta.name))
 
     def execute(self, data: pd.DataFrame):

giskard/slicing/text_slicer.py

@@ -1,21 +1,20 @@
-import os
-import copy
 from typing import Optional, Sequence
 
+import copy
+import os
+
 import numpy as np
 import pandas as pd
 
-from ..ml_worker.testing.registry.registry import get_object_uuid
-
-from ..core.core import DatasetProcessFunctionMeta
-
-from .base import BaseSlicer
-from .slice import Query, QueryBasedSliceFunction, ContainsWord
-from .utils import get_slicer
 from ..client.python_utils import warning
+from ..core.core import DatasetProcessFunctionMeta
 from ..datasets.base import Dataset
+from ..ml_worker.testing.registry.registry import get_object_uuid
 from ..ml_worker.testing.registry.slicing_function import SlicingFunction
 from ..slicing.category_slicer import CategorySlicer
+from .base import BaseSlicer
+from .slice import ContainsWord, Query, QueryBasedSliceFunction
+from .utils import get_slicer
 
 
 class TextSlicer(BaseSlicer):
@@ -162,7 +161,8 @@ class VectorizerError(ValueError):
 
 
 def _make_vectorizer(data: pd.Series, tfidf=False, **kwargs):
-    from sklearn.feature_extraction.text import TfidfVectorizer, CountVectorizer
+    from sklearn.feature_extraction.text import CountVectorizer, TfidfVectorizer
+
     from .stop_words import sw_en, sw_fr
 
     raw_stopwords = sw_en + sw_fr
@@ -195,7 +195,7 @@ def __init__(self, query: Query, feature: str, provider: str):
         self.meta.name = str(self)
         self.meta.display_name = str(self)
         self.meta.tags = ["pickle", "scan"]
-        self.meta.doc = "Automatically generated slicing function"
+        self.meta.doc = self.meta.default_doc("Automatically generated slicing function")
 
     def execute(self, dataset: Dataset) -> pd.DataFrame:
         metadata = dataset.column_meta[self.feature, self.provider]