Handle kfp types (#108)

This PR enables the user to pass different `bool`, `lists` and `dict` to
a component. Kubeflow typically handles those arguments by serializing
them as a string . For this reason, they need to be de-serialized again
within the component in order for them to be properly handled.

This might go away once we move to V2. 

References to the issue: 
kubeflow/pipelines#7457
kubeflow/pipelines#7719

examples/pipelines/simple_pipeline/components/load_from_hub/fondant_component.yaml

@@ -19,4 +19,10 @@ output_subsets:
 args:
   dataset_name:
     description: Name of dataset on the hub
-    type: str
+    type: str
+  bool_name:
+    description: Bool a
+    type: bool
+  dict_name:
+    description: List a
+    type: dict

examples/pipelines/simple_pipeline/components/load_from_hub/requirements.txt

@@ -1,5 +1,5 @@
 huggingface_hub==0.14.1
-git+https://github.com/ml6team/fondant.git@910331f3f2dc09030ed32a6578dcc2107c020590
+git+https://github.com/ml6team/fondant.git
 pyarrow>=7.0
 Pillow==9.4.0
 gcsfs==2023.4.0

examples/pipelines/simple_pipeline/components/load_from_hub/src/main.py

@@ -41,6 +41,7 @@ def load(self, *, dataset_name: str) -> dd.DataFrame:
         Returns:
             Dataset: HF dataset
         """
+
         # 1) Load data, read as Dask dataframe
         logger.info("Loading dataset from the hub...")
         dask_df = dd.read_parquet(f"hf://datasets/{dataset_name}")

fondant/component_spec.py

@@ -1,4 +1,5 @@
 """This module defines classes to represent an Fondant component specification."""
+import ast
 import copy
 import json
 import pkgutil
@@ -15,25 +16,24 @@
 from fondant.exceptions import InvalidComponentSpec
 from fondant.schema import Field, KubeflowCommandArguments, Type
 
+# TODO: remove after upgrading to kfpv2
+kubeflow2python_type = {
+    "String": str,
+    "Integer": int,
+    "Float": float,
+    "Boolean": ast.literal_eval,
+    "JsonObject": json.loads,
+    "JsonArray": json.loads,
+}
 # TODO: Change after upgrading to kfp v2
 # :https://www.kubeflow.org/docs/components/pipelines/v2/data-types/parameters/
 python2kubeflow_type = {
     "str": "String",
     "int": "Integer",
     "float": "Float",
     "bool": "Boolean",
-    "dict": "Map",
-    "list": "List",
-    "tuple": "List",
-    "set": "Set",
-}
-
-# TODO: remove after upgrading to kfpv2
-kubeflow2python_type = {
-    "String": str,
-    "Integer": int,
-    "Float": float,
-    "Boolean": bool,
+    "dict": "JsonObject",
+    "list": "JsonArray",
 }