B VNb@sddlmZddlmZddlZyddlZddlmZWndZdZYnXddddgZdd dZd d ZdddZ dddZ ddZ dS))print_function)arrayN)sparsesvm_read_problem evaluationscsr_find_scale_param csr_scaleFcCstdkr2|r2td}td}tddg}td}ng}g}dg}g}d}xtt|D]\}}|dd}t|dkr|dg7}|\} } |t| tdkr"|r"d} x^| D]R} | d\} }| dkrd}t|}|dkr|t| |||| d7} qW||d | qXi}x0| D]$} | d\} }t||t| <q0W||g7}qXWtdkr|rtj |dd }tj |dd }tj |dd }tj |dd }t |||f}||fS) a% svm_read_problem(data_file_name, return_scipy=False) -> [y, x], y: list, x: list of dictionary svm_read_problem(data_file_name, return_scipy=True) -> [y, x], y: ndarray, x: csr_matrix Read LIBSVM-format data from data_file_name and return labels y and data instances x. Ndlr:0)dtype) scipyr enumerateopensplitlenappendfloatintZ frombufferr csr_matrix)Zdata_file_nameZ return_scipyZprob_yZprob_xZrow_ptrZcol_idxZ indx_startilineZlabelZfeaturesZnzeZindvalZxir0lib/python3.7/site-packages/libsvm/commonutil.pyrsP       c Cs$tdkr t|tjr t|tjs(tdt|t|kr@tdd||k}||d}t|}|}|}||}||}||} tjdd`yD|||||||||||||| ||} Wnt d} YnXWdQRXt |t |t | fS) z evaluations_scipy(ty, pv) -> (ACC, MSE, SCC) ty, pv: ndarray Calculate accuracy, mean squared error and squared correlation coefficient using the true values (ty) and predicted values (pv). Nz!type of ty and pv must be ndarrayz len(ty) must be equal to len(pv)gY@raise)allnan) r isinstanceZndarray TypeErrorr ValueErrorZmeansumZerrstater) typvACCMSEr sumvsumysumvysumvvsumyySCCrrrevaluations_scipyDs$    Dr2TcCsRtdkr"|r"tt|t|St|t|kr:tdd}}d}}}}} xnt||D]`\} } | | krz|d7}|| | | | 7}|| 7}|| 7}|| | 7}|| | 7}| | | 7} qbWt|} d|| } || }yD| | ||| | ||| |||| |||}Wntd}YnXt| t|t|fS)a; evaluations(ty, pv, useScipy) -> (ACC, MSE, SCC) ty, pv: list, tuple or ndarray useScipy: convert ty, pv to ndarray, and use scipy functions for the evaluation Calculate accuracy, mean squared error and squared correlation coefficient using the true values (ty) and predicted values (pv). Nz len(ty) must be equal to len(pv)rr gY@r#)rr2Zasarrayrr&zipr)r(r)ZuseScipyZ total_correctZ total_errorr,r-r/r0r.vyr r*r+r1rrrr_s.    Drr c Cst|tjst||kst|j\}}|jdd}|jdd}||||}d||dk||dk<| ||}d||dk<t |dk|d| krt dt j d||dS)Nr)Zaxisg?zWARNING: The #nonzeros of the scaled data is at least 2 times larger than the original one. If feature values are non-negative and sparse, set lower=0 rather than the default lower=-1.)file)coefoffset)r$rrAssertionErrorshapeminZtoarrayZflattenmaxr'getnnzprintsysstderr) xlowerupperr nZfeat_minZfeat_maxr8r9rrrrs    cCst|tjst|d}|d}t|t|ks4t|j\}}|t|ksltdtjdt ||}t ||}t| d|}tj |g|d|j d}| tj|d||fd |}||krtd |d |d tjd|S) Nr9r8zMWARNING: The dimension of scaling parameters and feature number do not match.)r7r Zcsr)formatrr)r;zWARNING: original #nonzeros %d z > new #nonzeros %d zsIf feature values are non-negative and sparse, get scale_param by setting lower=0 rather than the default lower=-1.)r$rrr:rr;r?r@rAZresizeZreshapeZvstackrdotZdiagsr>)rBZ scale_paramr9r8r rEZscaled_xrrrrs$     )F)T)rr ) Z __future__rrr@rr__all__rr2rrrrrrrs     3 !