SDK/Components - Added naming.generate_unique_name_conversion_table (#…

…716)

generate_unique_name_conversion_table replaces _make_name_unique_by_adding_index and simplifies code in several places.

sdk/python/kfp/components/_components.py

@@ -21,7 +21,7 @@
 
 import sys
 from collections import OrderedDict
-from ._naming import _sanitize_file_name, _sanitize_python_function_name, _make_name_unique_by_adding_index
+from ._naming import _sanitize_file_name, _sanitize_python_function_name, generate_unique_name_conversion_table
 from ._yaml_utils import load_yaml
 from ._structures import ComponentSpec
 from ._structures import *
@@ -165,15 +165,11 @@ def _create_task_factory_from_component_spec(component_spec:ComponentSpec, compo
     description = component_spec.description
 
     inputs_list = component_spec.inputs or [] #List[InputSpec]
+    input_names = [input.name for input in inputs_list]
 
     #Creating the name translation tables : Original <-> Pythonic 
-    input_name_to_pythonic = {}
-    pythonic_name_to_input_name = {}
-    for io_port in inputs_list:
-        pythonic_name = _sanitize_python_function_name(io_port.name)
-        pythonic_name = _make_name_unique_by_adding_index(pythonic_name, pythonic_name_to_input_name, '_')
-        input_name_to_pythonic[io_port.name] = pythonic_name
-        pythonic_name_to_input_name[pythonic_name] = io_port.name
+    input_name_to_pythonic = generate_unique_name_conversion_table(input_names, _sanitize_python_function_name)
+    pythonic_name_to_input_name = {v: k for k, v in input_name_to_pythonic.items()}
 
     if component_ref is None:
         component_ref = ComponentReference(name=component_spec.name or component_filename or _default_component_name)

sdk/python/kfp/components/_dsl_bridge.py

@@ -140,15 +140,10 @@ def _create_container_op_from_resolved_task(name:str, container_image:str, comma
             _dummy_pipeline = dsl.Pipeline('dummy pipeline')
         _dummy_pipeline.__enter__()
 
-    from ._naming import _sanitize_kubernetes_resource_name, _make_name_unique_by_adding_index
-    output_name_to_kubernetes = {}
-    kubernetes_name_to_output_name = {}
-    for output_name in (output_paths or {}).keys():
-        kubernetes_name = _sanitize_kubernetes_resource_name(output_name)
-        kubernetes_name = _make_name_unique_by_adding_index(kubernetes_name, kubernetes_name_to_output_name, '-')
-        output_name_to_kubernetes[output_name] = kubernetes_name
-        kubernetes_name_to_output_name[kubernetes_name] = output_name
-
+    #Renaming outputs to conform with ContainerOp/Argo
+    from ._naming import _sanitize_kubernetes_resource_name, generate_unique_name_conversion_table
+    output_names = (output_paths or {}).keys()
+    output_name_to_kubernetes = generate_unique_name_conversion_table(output_names, _sanitize_kubernetes_resource_name)
     output_paths_for_container_op = {output_name_to_kubernetes[name]: path for name, path in output_paths.items()}
 
     # Construct the ComponentMeta

sdk/python/kfp/components/_naming.py

@@ -20,11 +20,14 @@
     '_convert_to_human_name',
     '_generate_unique_suffix',
     '_make_name_unique_by_adding_index',
+    '_convert_name_and_make_it_unique_by_adding_number',
+    'generate_unique_name_conversion_table',
 ]
 
 
 import re
 import sys
+from typing import Callable, Sequence, Mapping
 
 
 def _normalize_identifier_name(name):
@@ -72,3 +75,27 @@ def _make_name_unique_by_adding_index(name:str, collection, delimiter:str):
             if unique_name not in collection:
                 break
     return unique_name
+
+
+def _convert_name_and_make_it_unique_by_adding_number(name: str, used_converted_names, conversion_func: Callable[[str], str]):
+    converted_name = conversion_func(name)
+    if converted_name in used_converted_names:
+        for i in range(2, sys.maxsize ** 10): #Starting indices from 2: "Something", "Something_2", ...
+            converted_name = conversion_func(name + ' ' + str(i))
+            if converted_name not in used_converted_names:
+                break
+    return converted_name
+
+
+def generate_unique_name_conversion_table(names: Sequence[str], conversion_func: Callable[[str], str]) -> Mapping[str, str]:
+    '''Given a list of names and conversion_func, this function generates a map from original names to converted names that are made unique by adding numbers.
+    '''
+    forward_map = {}
+    reverse_map = {}
+    for name in names:
+        if name in forward_map:
+            raise ValueError('Original name {} is not unique.'.format(name))
+        converted_name = _convert_name_and_make_it_unique_by_adding_number(name, reverse_map, conversion_func)
+        forward_map[name] = converted_name
+        reverse_map[converted_name] = name
+    return forward_map