EiSrSSKrSSKrSSKJr Sr"SS\\5r"SS \ \ 5r S r S r SS jrSS jrSrSrSrSS.SjrSrSrSrSSjrSrg)zKAssorted utilities, which do not need anything other then torch and stdlib.N) _dtypes_implch[U[5(ag[U5 g![a gf=f)NFT) isinstancestrlen Exception)seqs O/home/mande/repo/quber/.venv/lib/python3.13/site-packages/torch/_numpy/_util.py is_sequencer s7#s C  s $ 11c\rSrSrSrg) AxisErrorN__name__ __module__ __qualname____firstlineno____static_attributes__rr rrrrc\rSrSrSrg)UFuncTypeErrorrNrrrr rrrrrcNUb!URU:waURU5nU$N)dtypeto)tensorrs r cast_if_neededr!s& V\\U25! Mrc[R"UR5S:a.UR[R"5R 5nU$)N)r _categoryrrdefault_dtypes float_dtype)xs r cast_int_to_floatr(&s;agg&* DD,,.:: ; Hrc\U*Us=::aU:dO [SUSU35eUS:aX- nU$)Nzaxis z) is out of bounds for array of dimension r)r)axndimargnames r normalize_axis_indexr-.s= ER $ %t#LTFSTT Av  Ircf^^[U5[[4;a[R"U5/n[UU4SjU55nU(dP[ [[[U555[ U5:wa!T(a[STS35e[S5eU$![ a N{f=f)ar Normalizes an axis argument into a tuple of non-negative integer axes. This handles shorthands such as ``1`` and converts them to ``(1,)``, as well as performing the handling of negative indices covered by `normalize_axis_index`. By default, this forbids axes from being specified multiple times. Used internally by multi-axis-checking logic. Parameters ---------- axis : int, iterable of int The un-normalized index or indices of the axis. ndim : int The number of dimensions of the array that `axis` should be normalized against. argname : str, optional A prefix to put before the error message, typically the name of the argument. allow_duplicate : bool, optional If False, the default, disallow an axis from being specified twice. Returns ------- normalized_axes : tuple of int The normalized axis index, such that `0 <= normalized_axis < ndim` c3>># UHn[UTT5v M g7fr)r-).0r*r,r+s r 'normalize_axis_tuple..[sH4R%b$884zrepeated axis in `z ` argumentz repeated axis) typetuplelistoperatorindex TypeErrorrsetmapint ValueError)axisr+r,allow_duplicates `` r normalize_axis_tupler@7s< Dz%& NN4()D H4H HD s3s3~#673t9D 1'*EF F_- - K   sB## B0/B0cJUcU$[U5S:wa [S5eUS$)Nrzdoes not handle tuple axisr)rNotImplementedErrorr>s r allow_only_single_axisrDds, |  4yA~!">?? 7Nrc[U5[[4;aU4n[U5[U5-n[ X5n[ U5n[ U5Vs/sHoDU;aSO [U5PM nnU$s snf)Nr)r4r6r5rr@iterrangenext) arr_shaper>out_ndimshape_itr*shapes r expand_shaperMlsn Dz$&w4y3y>)H  /DIH;@? K?R*Q$x. 0?E K L LsA7cUc&SU-nURU5R5nU$[URU5nUR U5nU$)Nr)expand contiguousrMrLreshape)r r>r+rLs r apply_keepdimsrSwsS |t u%002 MV\\40& MrrCc:Uc[SU55nUS4$X4$)z#Flatten the arrays if axis is None.c3@# UHoR5v M g7fr)flatten)r0ars r r1$axis_none_flatten..s7w wsrr5)r>tensorss r axis_none_flattenr[s) |7w77z}rc [RnU"URXS9(d[SURSUSUS35e[ X5$)aDtype-cast tensor to target_dtype. Parameters ---------- t : torch.Tensor The tensor to cast target_dtype : torch dtype object The array dtype to cast all tensors to casting : str The casting mode, see `np.can_cast` Returns ------- `torch.Tensor` of the `target_dtype` dtype Raises ------ ValueError if the argument cannot be cast according to the `casting` rule )castingzCannot cast array data from z to z according to the rule '')r can_cast_implrr9r!)t target_dtyper]can_casts r typecast_tensorrcs[,))H AGG\ ;*177)4~5gYa A   ! **rc2^^[UU4SjU55$)Nc3>># UHn[UTT5v M g7fr)rc)r0r`r]ras r r1#typecast_tensors..sLGqL'::Gr3rY)rZrar]s ``r typecast_tensorsrgs LGL LLrc[R"U5nU$![a"nSUS[U5S3n[ U5eSnAff=f)Nzfailed to convert z! to ndarray. Internal error is: .)torch as_tensorr rrB)objr emesgs r _try_convert_to_tensorrosS(% M (#C5(J3q6(RST!$''(s AAAcH[U[R5(aUnOo[R"5n[R"[ R "[R55 [U5n[R"U5 [XA5nX4R- nUS:a!URSU-UR-5n[U5nU(aUR5nU$![R"U5 f=f![a SnNBf=f)adThe core logic of the array(...) function. Parameters ---------- obj : tensor_like The thing to coerce dtype : torch.dtype object or None Coerce to this torch dtype copy : bool Copy or not ndmin : int The results as least this many dimensions is_weak : bool Whether obj is a weakly typed python scalar. Returns ------- tensor : torch.Tensor a tensor object with requested dtype, ndim and copy semantics. Notes ----- This is almost a "tensor_like" coercive function. Does not handle wrapper ndarrays (those should be handled in the ndarray-aware layer prior to invoking this function). rrOF)rrjTensorget_default_dtypeset_default_dtyperget_default_dtype_forfloat32ror!r+viewrLboolr=clone)rlrcopyndminr default_dtype ndim_extras r _coerce_to_tensorr}s6#u||$$ //1   B B5== QR 3+C0F  # #M 2F *F$JA~TJ.=>Dz   M)  # #M 2 s1 C7 D7D D! D!cSSKJn [U5S:Xa [5$[U5S:XajUSn[ X!5(a UR $[ U[ 5(a2/nUHn[U5nURU5 M! [ U5$U$[ U[ 5(d![S[U5R35e[U5$)zHConvert all ndarrays from `inputs` to tensors. (other things are intact)r)ndarrayrz#Expected inputs to be a tuple, got ) _ndarrayrrr=rr r5ndarrays_to_tensorsappendAssertionErrorr4r)inputsrinput_result sub_input sub_results r rrs! 6{a| V  f & &==  & &F# 0;  j)$= M&%(( 5d6l6K6K5LM #6**rr)NF)NFr)__doc__r7rjrr r= IndexErrorrr9 RuntimeErrorrr!r(r-r@rDrMrSr[rcrgror}rrrr rsR  J  Y   *Z&*+@M;|+r