This is needed for synchronisation with a file system. Then we need to address
each bucket by key (e.g. filename). But we also need to address by the order of
ingestion for processing with progress indicator.

call reindexObject on temporary object doesn't triggered catalog method: https://lab.nexedi.com/nexedi/erp5/blob/master/product/ERP5Type/mixin/temporary.py#L50, so call directly catalogObjectList on temporary object

example usage: data_array.activate().reindexRows(0, 10000, ('z_catalog_dynamic_duo_raw_array_list',))

/reviewed-on !7

Ingestion operations which write to data stream do not
have bucket_reference parameter. Now we do only pass
this parameter if we get it from the tag parsing script.

Data ingestion batch does not have validation workflow, so
we cannot validate it.

/reviewed-on klaus/wendelin!6

when using excludemin=True, the result was not as expected if
no key > start_key was in the tree

this way, a transformation can define which items to create in an analyis

enable data analysis as causality for other data analysis and adding input and output parameter as dict with additional information

this enabes support for resolution as parameter for resampling script

without this work around aggregate from any portal type would be returned, because getAggregateValue(portal_type=[]) does not return None as expected. This must be fixed properly in generic erp5

ingestion: also copy use supply line to ingestion line, it is needed for batch/stream case distinction