aliastools

This module provides tools for the efficient handling of input and output sequence aliases.

Note

We are not sure if class LazyInOutSequenceImport always works as expected under Python 3.6. Hence, function write_sequencealiases() does not use it when generating the alias modules inputs and outputs if one executes it with Python 3.6. Applying LazyInOutSequenceImport under Python 3.6 is at your own risk.

Module aliastools implements the following members:

• LazyInOutSequenceImport Import the input or output sequence the alias is referring to only when required.

• write_sequencealiases() Write the alias modules inputs and outputs.

---

class hydpy.core.aliastools.LazyInOutSequenceImport(modulename: str, classname: str, alias: str, namespace: Dict[str, Any])

Bases: object

Import the input or output sequence the alias is referring to only when required.

Importing all input and output sequences costs a significant amount of initialisation time, and one typically uses at most a few of them within a specific project. Hence, we implemented class LazyInOutSequenceImport for postponing the imports of the sequences available by modules inputs and outputs. We hope we could catch all exceptional cases so that a user will never realise to work with a LazyInOutSequenceImport object instead with an input or output class object (except when directly asking for it with function type).

Directly after importing, the alias is of type LazyInOutSequenceImport:

>>> from hydpy.core.aliastools import LazyInOutSequenceImport
>>> def get_example_alias():
...     return LazyInOutSequenceImport(
...         modulename="hydpy.models.hland.hland_inputs",
...         classname="T",
...         alias="hland_T",
...         namespace=locals(),
...     )
>>> hland_T = get_example_alias()
>>> def normalise_typestring(typestr):
...     # for handling differences between Python 3.6 and newer versions
...     if "class" in typestr:
...         typestr = typestr.split("'")[1]
...     if typestr == "typing.GenericMeta":
...         return "type"
...     return typestr
>>> normalise_typestring(str(type(hland_T)))
'hydpy.core.aliastools.LazyInOutSequenceImport'

After accessing an attribute, its type changed to type, which is the meta-class of all input and output sequences:

>>> hland_T.__name__
'T'
>>> normalise_typestring(str(type(hland_T)))
'type'

Due to Python’s special method lookup, we must explicitly define the relevant ones. The following examples list all considered methods:

>>> hland_T = get_example_alias()
>>> from hydpy.models.hland.hland_inputs import T, P
>>> hash(hland_T) == hash(T)
True

>>> hland_T = get_example_alias()
>>> hland_T == T
True
>>> hland_T = get_example_alias()
>>> T == hland_T   
True  # correct for Python 3.7 and newer but incorrect for Python 3.6

>>> hland_T = get_example_alias()
>>> hland_T != P
True
>>> hland_T = get_example_alias()
>>> P != hland_T
True

>>> hland_T = get_example_alias()
>>> normalise_typestring(str(hland_T))
'hydpy.models.hland.hland_inputs.T'

>>> hland_T = get_example_alias()
>>> normalise_typestring(repr(hland_T))
'hydpy.models.hland.hland_inputs.T'

>>> hland_T = get_example_alias()
>>> dir(hland_T)   
['INIT', 'NDIM', ...]

hydpy.core.aliastools.write_sequencealiases() → None

Write the alias modules inputs and outputs.

After extending a model or adding a new one, you can update both modules by calling write_sequencealiases():

>>> from hydpy.core.aliastools import write_sequencealiases
>>> write_sequencealiases()