[Ready for review] Add explicit max length control for foreach value. (…

…#1733)

* add explicit max length

* make foreach value string len max a configurable thing

metaflow/flowspec.py

@@ -17,7 +17,7 @@
 )
 from .graph import FlowGraph
 from .unbounded_foreach import UnboundedForeachInput
-from .metaflow_config import INCLUDE_FOREACH_STACK
+from .metaflow_config import INCLUDE_FOREACH_STACK, MAXIMUM_FOREACH_VALUE_CHARS
 
 # For Python 3 compatibility
 try:
@@ -529,7 +529,7 @@ def _is_primitive_type(item):
             )
 
         value = item if _is_primitive_type(item) else reprlib.Repr().repr(item)
-        return basestring(value)
+        return basestring(value)[:MAXIMUM_FOREACH_VALUE_CHARS]
 
     def next(self, *dsts: Callable[..., None], **kwargs) -> None:
         """

metaflow/metaflow_config.py

@@ -205,6 +205,8 @@
 # Controls whether to include foreach stack information in metadata.
 # TODO(Darin, 05/01/24): Remove this flag once we are confident with this feature.
 INCLUDE_FOREACH_STACK = from_conf("INCLUDE_FOREACH_STACK", False)
+# Maximum length of the foreach value string to be stored in each ForeachFrame.
+MAXIMUM_FOREACH_VALUE_CHARS = from_conf("MAXIMUM_FOREACH_VALUE_CHARS", 30)
 
 ###
 # Organization customizations