a Df@sxddlZddlmZddlmZmZmZddlm Z ddl Z ddl m Z ddl mZddlZddlZddlmZz ddlZWn$eyeded dYn0zddlZdd lmZWn$eyed ed dYn0Gd ddeZddZd,ddZd-ddZd.ddZd/ddZddZddZd0d!d"Z d1d$d%Z!d2d&d'Z"Gd(d)d)ej#j$Z%Gd*d+d+ej&Z'dS)3N)UMAP)warncatch_warningsfilterwarnings) TypingError)spectral_layout)check_random_state)KDTreeaThe umap.parametric_umap package requires Tensorflow > 2.0 to be installed. You can install Tensorflow at https://www.tensorflow.org/install or you can install the CPU version of Tensorflow using pip install umap-learn[parametric_umap] z/umap.parametric_umap requires Tensorflow >= 2.0)opszEThe umap.parametric_umap package requires Keras >= 3 to be installed.z#umap.parametric_umap requires Kerasc seZdZddddddddddif fdd Zdfdd Zdfd d Zd d Zfd dZddZddZ ddZ dddZ Z S)ParametricUMAPNF?rc  stjfi| ||_||_||_||_||_||_||_||_ d|_ | |_ | |_ | |_ d|_d|_tjjddd|_|jdur|jdjd|jkrtd |jdjd|jdS) ab Parametric UMAP subclassing UMAP-learn, based on keras/tensorflow. There is also a non-parametric implementation contained within to compare with the base non-parametric implementation. Parameters ---------- batch_size : int, optional size of batch used for batch training, by default None dims : tuple, optional dimensionality of data, if not flat (e.g. (32x32x3 images for ConvNet), by default None encoder : keras.Sequential, optional The encoder Keras network decoder : keras.Sequential, optional the decoder Keras network parametric_reconstruction : bool, optional Whether the decoder is parametric or non-parametric, by default False parametric_reconstruction_loss_fcn : bool, optional What loss function to use for parametric reconstruction, by default keras.losses.BinaryCrossentropy parametric_reconstruction_loss_weight : float, optional How to weight the parametric reconstruction loss relative to umap loss, by default 1.0 autoencoder_loss : bool, optional [description], by default False reconstruction_validation : array, optional validation X data for reconstruction loss, by default None global_correlation_loss_weight : float, optional Whether to additionally train on correlation of global pairwise relationships (>0), by default 0 keras_fit_kwargs : dict, optional additional arguments for model.fit (like callbacks), by default {} NMbP?@Z clipvaluerzNDimensionality of embedder network output ({}) doesnot match n_components ({}))super__init__dimsencoderdecoderparametric_reconstruction%parametric_reconstruction_loss_weight"parametric_reconstruction_loss_fcnautoencoder_loss batch_sizeloss_report_frequencyglobal_correlation_loss_weightreconstruction_validationkeras_fit_kwargsparametric_modeln_training_epochskeras optimizersAdam optimizeroutputsshape n_components ValueErrorformat) selfrrrrrrrrrrr kwargs __class__a/nfs/NAS7/SABIOD/METHODE/ermites/ermites_venv/lib/python3.9/site-packages/umap/parametric_umap.pyr&s2. zParametricUMAP.__init__cs@|jdkr.|durtd||_t||St||SdSN precomputedzTPrecomputed distances must be supplied if metric is precomputed.)metricr*_Xrfitr,XyZprecomputed_distancesr.r0r1r6}s zParametricUMAP.fitcs@|jdkr.|durtd||_t||St||SdSr2)r4r*r5r fit_transformr7r.r0r1r:s zParametricUMAP.fit_transformcCs|jjt||j|jdS)awTransform X into the existing embedded space and return that transformed output. Parameters ---------- X : array, shape (n_samples, n_features) New data to be transformed. Returns ------- X_new : array, shape (n_samples, n_components) Embedding of the new data in low-dimensional space. rverbose)rpredictnp asanyarrayrr<r,r8r0r0r1 transforms zParametricUMAP.transformcs2|jr"|jjt||j|jdSt|SdS)a Transform X in the existing embedded space back into the input data space and return that transformed output. Parameters ---------- X : array, shape (n_samples, n_components) New points to be inverse transformed. Returns ------- X_new : array, shape (n_samples, n_features) Generated data points new data in data space. r;N) rrr=r>r?rr<rinverse_transformr@r.r0r1rBs z ParametricUMAP.inverse_transformc Cs:|j}t|j|j|j|j|j|j|j||j |j d |_ dS)zDefine the model in keras)negative_sample_raterr!parametric_reconstruction_loss_fnrrrrN) r UMAPModelZ_a_brCrrrrrrr!)r,Zprlwr0r0r1 _define_modelszParametricUMAP._define_modelcCs|jdkr|j}|jdur.t|dg|_n*t|jdkrXt|t|gt|j}|jrt |dkszt |dkrt dt ||j |j|j|j|j\}|_}}}|_tt|tjd|_tt|tjd|_d} t|} t|j|j|j|j| |j| \|_|_|t||j|j} |jr|j durt|jdkrzt|j t|j gt|j|_ |j t!|j fd|j if} nd} |j"j#|f|j|j$| | d |j%} | j&|_'|jj(||j)d }|ifS) Nr3rrr zMData should be scaled to the range 0-1 for cross-entropy reconstruction loss.rreconstruction)Zepochssteps_per_epochvalidation_data)r<)*r4r5rr>r(lenreshapelistrmaxminrconstruct_edge_datasetgraph_n_epochsrrZ edge_weightr array expand_dimsastypeint64headtailprepare_networksrrr)rGintrrZ zeros_liker!r6r"r historyZ_historyr=r<)r,r8rSinit random_state edge_datasetZn_edgesrXrYinit_embeddingn_datarJrKr\ embeddingr0r0r1_fit_embed_datas  "     zParametricUMAP._fit_embed_datacCstdd|jDS)Ncss,|]$\}}t||r|dvr||fVqdS))r&rrr!N) should_pickle).0kvr0r0r1 1sz.ParametricUMAP.__getstate__..)dict__dict__items)r,r0r0r1 __getstate__/szParametricUMAP.__getstate__Tc Cs2|jdur6tj|d}|j||r6td||jdurltj|d}|j||rltd||jdurtj|d}|j||rtd|t tt dtj|d}t |d  }t ||t jWdn1s0Y|rtd |Wdn1s$0YdS) N encoder.keraszKeras encoder model saved to {} decoder.keraszKeras decoder model saved to {}zparametric_model.keraszKeras full model saved to {}ignore model.pklwbz*Pickle of ParametricUMAP model saved to {})rospathjoinsaveprintr+rr!rropenpickledumpHIGHEST_PROTOCOL)r, save_locationr<encoder_outputdecoder_outputparametric_model_output model_outputoutputr0r0r1ru7s,       .zParametricUMAP.save)NN)NN)T) __name__ __module__ __qualname__rr6r:rArBrGrcrlru __classcell__r0r0r.r1r %s(W br cCs|}||jd}|dur:|jddkr6d}nd}d|j|j|jt|k<|||j}|j}|j}|j}||||||fS)a< gets elements of graphs, weights, and number of epochs per edge Parameters ---------- graph_ : scipy.sparse.csr.csr_matrix umap graph of probabilities n_epochs : int maximum number of epochs per edge Returns ------- graph scipy.sparse.csr.csr_matrix umap graph epochs_per_sample np.array number of epochs to train each sample for head np.array edge head tail np.array edge tail weight np.array edge weight n_vertices int number of vertices in graph rNr'irH) ZtocooZsum_duplicatesr(datarOfloatZeliminate_zerosrowcol)rRrSgraph n_verticesepochs_per_samplerXrYweightr0r0r1get_graph_elementsXs  rspectralcCsD|durtd}t|trF|dkrF|jdd|jd|fdtj}nt|tr|dkrt||||||d}dt | } || tj|j d |jd|gd tj}nt |} t | jd kr@tj| dd jd| jdkrfloat32rabsrOnormalrTrLuniquer querymean)Z _raw_datarr)r^r4Z _metric_kwdsr]rbZinitialisationZ expansionZ init_datatreedistindZnndistr0r0r1init_embedding_from_graphsP   rr cCst||d| S)a convert distance representation into log probability, as a function of a, b params Parameters ---------- distances : array euclidean distance between two points in embedding a : float, optional parameter based on min_dist, by default 1.0 b : float, optional parameter based on min_dist, by default 1.0 Returns ------- float log probability in embedding space r)r log1p)Z distancesabr0r0r1#convert_distance_to_log_probabilitysrrcCs>| t|}d| t|||}||}|||fS)a Compute cross entropy between low and high probability Parameters ---------- probabilities_graph : array high dimensional probabilities log_probabilities_distance : array low dimensional log probabilities EPS : float, optional offset to ensure log is taken of a positive number, by default 1e-4 repulsion_strength : float, optional strength of repulsion between negative samples, by default 1.0 Returns ------- attraction_term: float attraction term for cross entropy loss repellant_term: float repellent term for cross entropy loss cross_entropy: float cross entropy umap loss r )r Z log_sigmoid)probabilities_graphlog_probabilities_distanceZEPSrepulsion_strengthZattraction_termZrepellant_termZCEr0r0r1compute_cross_entropys rc Cs|dur^ttjj|dtjtjjdddtjjdddtjjdddtjj|ddg}|dur|rttjj|fdtjjdddtjjdddtjjdddtjjt|ddd tj|g}||fS) a` Generates a set of keras networks for the encoder and decoder if one has not already been predefined. Parameters ---------- encoder : keras.Sequential The encoder Keras network decoder : keras.Sequential the decoder Keras network n_components : int the dimensionality of the latent space dims : tuple of shape (dim1, dim2, dim3...) dimensionality of data n_data : number of elements in dataset # of elements in training dataset parametric_reconstruction : bool Whether the decoder is parametric or non-parametric init_embedding : array (optional, default None) The initial embedding, for nonparametric embeddings Returns ------- encoder: keras.Sequential encoder keras network decoder: keras.Sequential decoder keras network N)r(dZrelu)units activationz)rnameZrecon)rrr) r#Z SequentiallayersZInputFlattenZDenser>productZReshape)rrr)rrarr`r0r0r1rZ s0&     rZcsDfddjddkrdndfdd}fd d }t||\}} } } } } d urrtt| d gt| | d t| | d }}tjt t |}||tj }||tj }t j j||f}|}|d}|jdd}|j|t j jjd}|j|t j jjd}|d}|t || | | fS)a Construct a tf.data.Dataset of edges, sampled by edge weight. Parameters ---------- X : array, shape (n_samples, n_features) New data to be transformed. graph_ : scipy.sparse.csr.csr_matrix Generated UMAP graph n_epochs : int # of epochs to train each edge batch_size : int batch size parametric_reconstruction : bool Whether the decoder is parametric or non-parametric cs|SNr0)index)r8r0r1 gather_indexhsz,construct_edge_dataset..gather_indexg& .>g?TFcsVr6t|gtjgd}t|gtjgd}nt|}t|}||fS)Nr)tfZ py_functionrgather)Zedge_toZ edge_from edge_to_batchedge_from_batch)r8rgather_indices_in_pythonr0r1gather_Xos   z(construct_edge_dataset..gather_Xcs8dtdi}dkr ||d<r,||d<||f|fS)Numaprglobal_correlationrI)r repeat)rrr')rrrr0r1 get_outputsys z+construct_edge_dataset..get_outputsNir[r)Zdrop_remainder)Znum_parallel_callsr )nbytesrr[r>rPrrVrZ permutationrangerLrWrrZDatasetZfrom_tensor_slicesshufflebatchmapZ experimentalZAUTOTUNEZprefetch)r8rRrSrrrrr_rrXrYrrZ edges_to_expZedges_from_expZ shuffle_maskr_r0)r8rrrrrr1rQPs:       rQcCsz0tt|d}tt|d}Wntjtjj t tjj tjj t ttfy}ztd||WYd}~dSd}~0ty}z(td|d|d|WYd}~dSd}~00dS) a Checks if a dictionary item can be pickled Parameters ---------- key : try key for dictionary element val : None element of dictionary Returns ------- picklable: bool whether the dictionary item can be pickled base64zDid not pickle {}: {}NFzFailed at pickling :z due to T)codecsencoderxdumpsdecodeloads PicklingErrorrerrorsZInvalidArgumentError TypeErrorZ InternalErrorZ NotFoundError OverflowErrorrAttributeErrorrr+r*)keyvalZpickledrer0r0r1rds& rdTcCstj|d}tt|d}|r0td|tj|d}tj|rjt j ||_ |rjtd|tj|d}tj|rt j ||_ td|tj|d}tj|rt j ||_td ||S) a Load a parametric UMAP model consisting of a umap-learn UMAP object and corresponding keras models. Parameters ---------- save_location : str the folder that the model was saved in verbose : bool, optional Whether to print the loading steps, by default True Returns ------- parametric_umap.ParametricUMAP Parametric UMAP objects rprbz-Pickle of ParametricUMAP model loaded from {}rmz"Keras encoder model loaded from {}rnz"Keras decoder model loaded from {}r!zKeras full model loaded from {})rrrsrtrxloadrwrvr+existsr#modelsZ load_modelrrr!)r{r<rmodelr|r}r~r0r0r1load_ParametricUMAPs(   rFc Cst|}|tj|ddd}|durF|}tj|t|d|d}n4tj||jd}|tj|ddd}t|jdg}dg}tj|dd}tj|dd}t|}t|}t|d}t|d}t |||f} t |||f} t |||f} t |||f} t | t| t||j} t | |ddf} t| } t | t| dd||f} |stj | dd } | S) zAdapted from TF Probability.rTrkeepdimsNdtyperint32rr) r Zconvert_to_tensorrZconjrrLr(castZ transposerMmatmulsqueeze) xr9rZ event_axisZ sample_axisZx_permedZy_permedZn_eventsZ n_samplesZ x_permed_flatZ y_permed_flatZcovr0r0r1 covariances`      rcCs>|tj|ddd}|dur0|tj|ddd}t|||dS)NrTrrr9r)r stdrrr0r0r1 correlationDsrc@seZdZddZdS) StopGradientcCs t|Sr)r stop_gradient)r,rr0r0r1callQszStopGradient.callN)rrrrr0r0r0r1rPsrcsJeZdZdfdd Zdd Zdd d Zdd d ZddZddZZ S)rENFr rH umap_modelc stj| d||_||_||_| |_| |_||_||_||_ | |_ |pVt j j ddd}|j|dt j|_t j|_|durt jjdd|_n||_dS)N)rrrr)r&T)Z from_logits)rrrrrrrrC umap_loss_a umap_loss_brr#r$r%compilerrflattenrZ SeedGeneratorseed_generatorlossesZBinaryCrossentropyrD) r,rrrCrrr&rDrrrrrr.r0r1rVs*     zUMAPModel.__init__cCsZ|\}}||}||}||d}|jrV|jr>||}n|t|}||d<|S)N) embedding_toembedding_fromrI)rrrrr r)r,inputsZto_xZfrom_xrry_predZembedding_to_reconr0r0r1rs   zUMAPModel.callcKsxg}|jD]}|tj|tjdq ||||jdkrV|| |||j rn|| ||t |S)Nrr) rappendr rr#backendZfloatx _umap_lossr_global_correlation_lossr_parametric_reconstruction_losssum)r,rr9rZ sample_weightr-rlossr0r0r1 compute_losss  zUMAPModel.compute_losscCs|d}|d}tj||jdd}tj||jdd}t|d}tjjt||jd}tj dkrvt ||} n||} tj tj||ddtj|| ddgdd} t| |j|j} t|d} tj t| ft| |jfgdd} t| | |d\}}}t|S) Nrrrrseed tensorflowr)r)r rrCr(r#rrZarangerconfigrrrZ concatenatenormrrrZoneszerosrr)r,rrrrZembedding_neg_toZ repeat_negZrepeat_neg_batch_dimZshuffled_indicesZembedding_neg_fromZdistance_embeddingrrrZattraction_lossZrepellant_lossZce_lossr0r0r1rsF      zUMAPModel._umap_lossc Cs||d}||d}dd}||}||}t|dd}t|dd}tj|dd|dddd }tj|dd|dddd }|tjj|j|jd d }t t t |dt |dd }| |j S) NrrcSs|t|t|Sr)r rr)rr0r0r1z_scoresz3UMAPModel._global_correlation_loss..z_scoreir rrrrg|=)rr9) rr Zcliprr#rrr(rrrrUr) r,r9rrZz_xrZdxZdzZcorr_dr0r0r1rs ""z"UMAPModel._global_correlation_losscCs||d|d}||jS)NrI)rDr)r,r9rrr0r0r1rs z)UMAPModel._parametric_reconstruction_loss)NNFr rHFr)NNNN)r ) rrrrrrrrrrr0r0r.r1rEUs)  5rE)r)r r )rr )N)T)NF)NF)(numpyr>rrwarningsrrrZnumbarrrZ umap.spectralrZ sklearn.utilsrrrxZsklearn.neighborsr rr ImportErrorr#r r rrrrrZrQrdrrrrZLayerrZModelrEr0r0r0r1sZ         57 :  5 DY( 2 C