a ;gh+@sdZddlmZddlmZddlZddlmZddl m Z m Z gdZ dd Z d d Zd d ZddZdddZdddZddZd ddZd!ddZd"ddZdS)#z` This provides a small set of utilities in NumPyro that are used to diagnose posterior samples. ) OrderedDict)productN) device_get) tree_flattentree_map)autocorrelationautocovarianceeffective_sample_size gelman_rubinhpdisplit_gelman_rubin print_summarycCsz|jdd\}}|jddd}|jdd}||d|}|jddkrn|jdd}|jddd}||}n|}||fS)NaxisZddofrr)shapevarmean)xCNZ chain_var var_within var_estimatorZ chain_meanZ var_betweenrc/mounts/lovelace/software/anaconda3/envs/metaDMG/lib/python3.9/site-packages/numpyro/diagnostics.py_compute_chain_variance_statss   rcCs~|jdksJ|jddks J|jddks2Jt|\}}tjdddt||}Wdn1sp0Y|S)aO Computes R-hat over chains of samples ``x``, where the first dimension of ``x`` is chain dimension and the second dimension of ``x`` is draw dimension. It is required that ``x.shape[0] >= 2`` and ``x.shape[1] >= 2``. :param numpy.ndarray x: the input array. :return: R-hat of ``x``. :rtype: numpy.ndarray rrrignoreinvaliddivideN)ndimrrnperrstatesqrt)rrrZrhatrrrr +s  ,r cCsn|jdksJ|jddks J|jdd}tj|ddd|f|dd| dfgdd}t|}|S)aC Computes split R-hat over chains of samples ``x``, where the first dimension of ``x`` is chain dimension and the second dimension of ``x`` is draw dimension. It is required that ``x.shape[1] >= 4``. :param numpy.ndarray x: the input array. :return: split R-hat of ``x``. :rtype: numpy.ndarray rrNrr)r"rr# concatenater )rZN_halfZ new_inputZ split_rhatrrrr >s 4r cCsl|dkr |S|}|ddkr&|d}q|ddkr<|d}q&|ddkrR|d}q<|dkr^|S|d7}q dS)Nrrrr)targetmrrr_fft_next_fast_lenQs      r,c Cs|j|}t|}d|}t||d}||jddd}tjj||dd}|t|}tjj||dd}|dd|f}|t |dd}tj d d d $||ddd f}Wdn1s0Yt||dS) z Computes the autocorrelation of samples at dimension ``axis``. :param numpy.ndarray x: the input array. :param int axis: the dimension to calculate autocorrelation. :return: autocorrelation of ``x``. :rtype: numpy.ndarray rTrZkeepdims)nr.Ngrrr) rr,r#swapaxesrZfftZrfft conjugateZirfftZaranger$) rrrMZM2Zcentered_signalZfreqvecZ freqvec_gramZautocorrrrrrcs 2rcCst|||j|ddS)z Computes the autocovariance of samples at dimension ``axis``. :param numpy.ndarray x: the input array. :param int axis: the dimension to calculate autocovariance. :return: autocovariance of ``x``. :rtype: numpy.ndarray Tr.)rr)rrrrrrs rc Cs|jdksJ|jddks Jt|dd}t|\}}d||jdd|}d|d<|ddddf|ddddf}|dd}tj|tjjtj |dddfddd ddgdd}d d |j dd}t |jdd|}|S) a Computes effective sample size of input ``x``, where the first dimension of ``x`` is chain dimension and the second dimension of ``x`` is draw dimension. **References:** 1. *Introduction to Markov Chain Monte Carlo*, Charles J. Geyer 2. *Stan Reference Manual version 2.18*, Stan Development Team :param numpy.ndarray x: the input array. :return: effective sample size of ``x``. :rtype: numpy.ndarray rrrg?rNr-.)Za_minZa_maxgg@) r"rrrrr#r'minimum accumulateZclipsumprod) rZ gamma_k_crrZrho_kZRho_kZRho_inittaun_effrrrr s"  $ &r ?c Cst||d}tj|dd}|jd}t||}|d||}||d}||}|jdd} | |} tj|| ddd} t| |d} tj|| ddd} t| |d} tj| | g|dS)a Computes "highest posterior density interval" (HPDI) which is the narrowest interval with probability mass ``prob``. :param numpy.ndarray x: the input array. :param float prob: the probability mass of samples within the interval. :param int axis: the dimension to calculate hpdi. :return: quantiles of ``x`` at ``(1 - prob) / 2`` and ``(1 + prob) / 2``. :rtype: numpy.ndarray rrNN.)r#r0sortrintZargminZtake_along_axisr') rprobrZsorted_xZmassZ index_lengthZintervals_leftZintervals_rightZintervals_lengthZ index_startZ index_endZhpd_leftZ hpd_rightrrrr s     r Tc Cs|stdd|}t|ts6ddtt|dD}i}|D]\}}t|}t|d|j dd}|j dd }|j dd d }tj |dd } t ||d } t|} t|} d dd |} d dd |}td|fd|fd| f| | df|| d fd| fd| fg||<qB|S)ay Returns a summary table displaying diagnostics of ``samples`` from the posterior. The diagnostics displayed are mean, standard deviation, median, the 90% Credibility Interval :func:`~numpyro.diagnostics.hpdi`, :func:`~numpyro.diagnostics.effective_sample_size`, and :func:`~numpyro.diagnostics.split_gelman_rubin`. :param samples: a collection of input samples with left most dimension is chain dimension and second to left most dimension is draw dimension. :type samples: dict or numpy.ndarray :param float prob: the probability mass of samples within the HPDI interval. :param bool group_by_chain: If True, each variable in `samples` will be treated as having shape `num_chains x num_samples x sample_shape`. Otherwise, the corresponding shape will be `num_samples x sample_shape` (i.e. without chain dimension). cSs|dSNr:rrrrrzsummary..cSsi|]\}}d||qSzParam:{}format.0ivrrr szsummary..r)r-rNrrr)r=z{:.1f}%2rstdmedianr8r_hat)r isinstancedict enumerateritemsrr#ZreshaperrrKrLr r r rDr)samplesr=group_by_chain summary_dictnamevalueZ value_flatrrKrLZhpdr8rMZ hpd_lowerZ hpd_upperrrrsummarys:      rWc sl|stdd|}t|ts6ddtt|dD}t||dd}dd|D}tttd d| d }d t |d }|d }dgt t | d }t t |j||d} |D]\} } | dj} t| dkrt | j| g| Rqttt| D]Hddtt } t | j| | gfdd| DRqqt dS)ax Prints a summary table displaying diagnostics of ``samples`` from the posterior. The diagnostics displayed are mean, standard deviation, median, the 90% Credibility Interval :func:`~numpyro.diagnostics.hpdi`, :func:`~numpyro.diagnostics.effective_sample_size`, and :func:`~numpyro.diagnostics.split_gelman_rubin`. :param samples: a collection of input samples with left most dimension is chain dimension and second to left most dimension is draw dimension. :type samples: dict or numpy.ndarray :param float prob: the probability mass of samples within the HPDI interval. :param bool group_by_chain: If True, each variable in `samples` will be treated as having shape `num_chains x num_samples x sample_shape`. Otherwise, the corresponding shape will be `num_samples x sample_shape` (i.e. without chain dimension). cSs|dSr>rr?rrrr@#rAzprint_summary..cSsi|]\}}d||qSrBrCrErrrrI%sz!print_summary..rT)rSc Ss<i|]4\}}||ddtdd|jdddqS)[,cSs t|dS)Nr)strr?rrrr@+rAz*print_summary...rN])joinmapr)rFkrHrrrrI*scSst|S)N)lenr?rrrr@.rA z{:>}z* {:>9} {:>9} {:>9} {:>9} {:>9} {:>9} {:>9}z? {:>9.2f} {:>9.2f} {:>9.2f} {:>9.2f} {:>9.2f} {:>9.2f} {:>9.2f}rz[{}]rYcsg|] }|qSrr)rFrHidxrr @rAz!print_summary..N)rrNrOrPrrWrQmaxr]valuesrZlistkeysprintrDrr_rranger\)rRr=rSrTZ row_namesmax_lenZ name_formatZ header_formatcolumnsZ row_formatrUZ stats_dictrZidx_strrrcrr s>   r )r)r)r9r)r9T)r9T)__doc__ collectionsr itertoolsrnumpyr#ZjaxrZ jax.tree_utilrr__all__rr r r,rrr r rWr rrrrs      % /  2