o
    H&i                     @   s   d dl Z d dlmZmZmZmZmZ d dlZd dlZd dlm	Z	 d dlm
Z
mZ d dlmZmZmZmZ d dlmZmZmZ ddlmZ G dd	 d	e
ZG d
d deZdS )    N)AnyCallableDictTupleOptional)TransformerProxy)ArgumentTargetNodemap_aggregate)normalize_modulenormalize_functioncreate_type_hint   )AnnotateTypesWithSchemac                       s   e Zd ZdZ	ddejjdef fddZde	de
f fd	d
Z		ddedeedf deee
f deee
df  deeee
f  f
 fddZdedeedf deee
f f fddZ  ZS )NormalizeArgsa  
    Normalize arguments to Python targets. This means that
    `args/kwargs` will be matched up to the module/functional's
    signature and rewritten to exclusively kwargs in positional order
    if `normalize_to_only_use_kwargs` is true. Also populates default
    values. Does not support positional-only parameters or varargs
    parameters (*args, **kwargs).

    If the nodes have 'type' metadata, it will use it to disambiguate
    overloads. Otherwise, it will throw an error.

    Example usage:
        m = torchvision.models.resnet18()
        traced = torch.fx.symbolic_trace(m)
        traced = NormalizeArgs(traced).transform()
    Tmodulenormalize_to_only_use_kwargsc                    s   t  | i | _|| _d S N)super__init__node_mapr   )selfr   r   	__class__ JC:\wamp64\www\opt\env\Lib\site-packages\torch/fx/experimental/normalize.pyr   $   s   
zNormalizeArgs.__init__nreturnc                    s   |  \}}fdd tj }t|tsJ tdd |D } fdd| D }jdkr>| j||||}nt	 
}jdkrX| j|< j|j_j|j_|S )	Nc                    s,   t | tjrd jv r jd S d S t| S )Ntype)
isinstancefxr   metar    )arg)r   r   r   get_type.   s   z(NormalizeArgs.run_node.<locals>.get_typec                 S   s   g | ]}t |qS r   )r   ).0ir   r   r   
<listcomp>5   s    z*NormalizeArgs.run_node.<locals>.<listcomp>c                    s   i | ]	\}}| |qS r   r   )r&   kv)r%   r   r   
<dictcomp>6   s    z*NormalizeArgs.run_node.<locals>.<dictcomp>call_functionoutput)Zfetch_args_kwargs_from_envr   argsr!   tupleitemsopr,   targetr   run_noder   r#   noder    )r   r   r.   kwargs	arg_typeskwarg_typesoutr   )r%   r   r   r3   +   s   




zNormalizeArgs.run_nodeNr2   r.   .r5   r6   r7   c           	         sN   t |sJ t|||||| j}|r|\}}| jd|||S t |||S )Nr,   )callabler   r   ZtracerZcreate_proxyr   r,   )	r   r2   r.   r5   r6   r7   new_args_and_kwargsnew_args
new_kwargsr   r   r   r,   A   s   zNormalizeArgs.call_functionc                    sN   t |tsJ t| j|||| j}|r|\}}t |||S t |||S r   )r!   strr   r   r   r   call_module)r   r2   r.   r5   r:   r;   r<   r   r   r   r>   Z   s   zNormalizeArgs.call_module)T)NN)__name__
__module____qualname____doc__torchr"   ZGraphModuleboolr   r   r   r3   r
   r   r	   r   r=   r   r,   r>   __classcell__r   r   r   r   r      s:    



r   c                       s   e Zd ZU dZejejejejejejej	ej
ejejejejejejejejejejejejejejejejiZeeeegef eeegef f ed< dedeedf deeef f fddZ  ZS )	NormalizeOperatorsa  
    Normalize callsites that are different ways of "spelling" the same
    invocation into a single, canonical call. Currently supports:

    1. Normalize operators (e.g. operator.add) to the `torch` ops they
       ultimately invoke (e.g. torch.add) when it is possible to statically
       reason that

    Example usage:

        m = torchvision.models.resnet18()

        traced = torch.fx.symbolic_trace(m)

        traced = NormalizeOperators(traced).transform()
    binary_magic_method_remapr2   r.   .r5   c                    sf   t |sJ || jv r+t|dkrt |||S |\}}t j| j| ||fi dS t |||S )N   )r2   r.   r5   )r9   rG   lenr   r,   )r   r2   r.   r5   lhsrhsr   r   r   r,      s   
z NormalizeOperators.call_function) r?   r@   rA   rB   rC   addoperatormulsubdivtruedivZfloor_dividefloordiv	remaindermodeqneltlegtgerG   r   r   r   __annotations__r
   r   r	   r=   r,   rE   r   r   r   r   rF   l   s2   
 

rF   )rM   typingr   r   r   r   r   rC   Ztorch.fxr"   r   r   Ztorch.fx.noder	   r
   r   r   Ztorch.fx.operator_schemasr   r   r   Zschema_type_annotationr   r   rF   r   r   r   r   <module>   s    Z