AVh (*dZddlmZddlmZddlmZddlmZddlm Z ddlm Z ddl m Z dd l mZdd lmZdd lmZdd lmZd ZGddZedgddZedgddZedgddZdZdZdZdZdZy)zTensor Handle Operations.)resource_handle_pb2)pywrap_tf_session)device)dtypes)ops)tensor) array_ops)gen_data_flow_ops)compat) numpy_compat) tf_exportcztjt|jtj S)z:Encode a ResourceHandle proto as custom numpy struct type.)dtype)r np_asarray bytearraySerializeToStringr np_resource)resource_handles Q/home/dcms/DCMS/lib/python3.12/site-packages/tensorflow/python/ops/session_ops.pyencode_resource_handler!s.  1134F>rc|jS)z)The string representation of this handle.r+r(s rr"zTensorHandle.handleUs <<rc |jstd|jzt|jj |j |j\}}|jj|||j iS)z:Return the value of the tensor represented by this handle.3Persistent tensor %s may have already been deleted. feed_dict) r TypeErrorr"_get_handle_readerrgraphrrrun)r!holderreaders revalzTensorHandle.evalZsq  K $ %%' (;(;T\\(, 5NFF ==  V /E  FFrc|jstd|jzd|_t|jj d|j \}}|jj|||jiy)z-Force the deletion of this persistent tensor.r:Frr;N)r r=r"_get_handle_deleterrr?rr@)r!rAdeleters rdeletezTensorHandle.deletecso  K $ %%!D)$--*=*=q$,,OOFGMMg&$++)>?rc(d|_|jS)zReturn the raw handle of the tensor. Note that the method disables the automatic garbage collection of this persistent tensor. The caller is now responsible for managing the life time of the tensor. F)r rr(s rget_raw_handlezTensorHandle.get_raw_handlels"D <<rcztj|}tj|j ddS)z&The device name encoded in the handle.r.r/)r rpydevcanonical_namer1)r" handle_strs r_get_device_namezTensorHandle._get_device_namevs5""6*J    0 0 5b 9 ::rcRt|jd}|ddz|dzS)zThe graph key for reader.r.rr/)strr1)r" handle_partss r_get_reader_keyzTensorHandle._get_reader_key|s1v;$$S)L ?S <#3 33rc`|jjdztj|zS)zThe graph key for mover.r.)opr4rrR)feederr"s r_get_mover_keyzTensorHandle._get_mover_keys' 99>>C ,">">v"F FFrN)__name__ __module__ __qualname____doc__r$r)r,r5r7propertyr"rCrGrI staticmethodrNrRrVrrrr(s;!"7!?  G@;; 44 GGrrget_session_handle)v1Nct|tjs tdt j |5t j||cdddS#1swYyxYw)aBReturn the handle of `data`. This is EXPERIMENTAL and subject to change. Keep `data` "in-place" in the runtime and create a handle that can be used to retrieve `data` in a subsequent run(). Combined with `get_session_tensor`, we can keep a tensor produced in one run call in place, and use it as the input in a future run call. Args: data: A tensor to be stored in the session. name: Optional name prefix for the return tensor. Returns: A scalar string tensor representing a unique handle for `data`. Raises: TypeError: if `data` is not a Tensor. Example: ```python c = tf.multiply(a, b) h = tf.compat.v1.get_session_handle(c) h = sess.run(h) p, a = tf.compat.v1.get_session_tensor(h.handle, tf.float32) b = tf.multiply(a, 10) c = sess.run(b, feed_dict={p: h.handle}) ``` z`data` must be of type Tensor.r4N) isinstance tensor_libTensorr=r colocate_withr r^)datar4s rr^r^sZF D*++ , 4 55 A  / /4 @AAAs AA%get_session_tensorc2tj|}tj|5t j t j}t|j||tj|||}ddd||fS#1swYfSxYw)aMGet the tensor of type `dtype` by feeding a tensor handle. This is EXPERIMENTAL and subject to change. Get the value of the tensor from a tensor handle. The tensor is produced in a previous run() and stored in the state of the session. Args: handle: The string representation of a persistent tensor handle. dtype: The type of the output tensor. name: Optional name prefix for the return tensor. Returns: A pair of tensors. The first is a placeholder for feeding a tensor handle and the second is the tensor in the session state keyed by the tensor handle. Example: ```python c = tf.multiply(a, b) h = tf.compat.v1.get_session_handle(c) h = sess.run(h) p, a = tf.compat.v1.get_session_tensor(h.handle, tf.float32) b = tf.multiply(a, 10) c = sess.run(b, feed_dict={p: h.handle}) ``` raN) rrNrrr placeholderrstring_register_handle_feederr?r rg)r"rr4 handle_devicerArs rrgrgsB//7- zz- L  " "6== 1FFLL&%8  1 1&%d KFL &  L & s AB  Bdelete_session_tensorctj|}tj|5t j t j}tj||}ddd||fS#1swYfSxYw)aDelete the tensor for the given tensor handle. This is EXPERIMENTAL and subject to change. Delete the tensor of a given tensor handle. The tensor is produced in a previous run() and stored in the state of the session. Args: handle: The string representation of a persistent tensor handle. name: Optional name prefix for the return tensor. Returns: A pair of graph elements. The first is a placeholder for feeding a tensor handle and the second is a deletion operation. raN) rrNrrr rirrjr rm)r"r4rlrArFs rrmrmss"//7- zz- I  " "6== 1F55f4HGI ' I ' s ;A22A>cJ||j|jj<yr&)_handle_feedersrTr4)r?rUrs rrkrks*/% 'rc`|jj|jjSr&)rpgetrTr4)r?rUs r_get_handle_feederrss!    " "699>> 22rctj|}|jj|}|tj |}|j 5|j |5tjtj}t|j||tj||}ddddddf}||j|<|S#1swY&xYw#1swY*xYw)z'Return a read subgraph for this handle.N)rrR_handle_readersrrrN as_defaultrr rirrjrkr?r rg)r?r"r graph_keyresultrlrArBs rr>r>s**62)  $ $Y /& ^ 11&9M    CU\\-8C$$V]]3ffllFE: 33FEBfCCf F'-E)$ - CCCCs%C,*AC ;C, C) %C,,C5ct||}|ytj|}|jj|k(rytj ||}|j j|}|t|||\}}|j5|j |jj5tj|} dddddd| f}||j |<|S#1swY&xYw#1swY*xYw)z:Return a move subgraph for this pair of feeder and handle.N) rsrrNrTrrV_handle_moversrrr>rvr r^) r?rUr"rrlrwrxrArBmovers r_get_handle_moverr| s UF +% ] //7- YY& ))&&9)    # #I .& ^'vu=NFF    ;U\\&))*:*:;;226:e;;e_F&,E# - ;;;;s$&C?8C3C?3C< 8C??Dc|jj|}|tj|}|j 5|j |5t jtj}tj|}ddddddf}||j|<|S#1swY&xYw#1swY*xYw)z+Return a deletion subgraph for this handle.N) _handle_deletersrrrrNrvrr rirrjr rm)r? deleter_keyr"rxrlrArFs rrErE!s  ! ! % %k 2& ^ 11&9M    @U\\-8@$$V]]3f!77?g@@g F*0E;' - @@@@s$B?9B3B?3B< 8B??Cr&)rZtensorflow.core.frameworkrtensorflow.python.clientrtensorflow.python.frameworkrrKrrrrctensorflow.python.opsr r tensorflow.python.utilr r tensorflow.python.util.tf_exportr rrr^rgrmrkrsr>r|rEr]rrrs  :67.+<+3)/6]G]G@ #$%'A&'AT #$%%&%P &'().03  * r