U md;@szddlmZddlmZddlmZddlZddlZddl Z ddZ ddZ d d d gddd fej eeeeed ddZdS))norm)check_nonnegative_integers)productNc2Csdd}d}t|jddf}t|jd}|jd}|dk rJ||nd}||}t|d}tj|dd } tj|dd } t| ddd f} t| ddd| f} t| t | } }t| t | }}i}i}t |D]}|dd|f}t | ddd|f|kd}t | ddd f|k}||}||}||||}||| |}|||<|||<qi}d}t t |t |D]\}}| ddd f|k}| ddd f|k} || @}!t |!dkrqt |!d}"d t|t|g}#||t |!d<|#||<|d7}||}$||}%||!}&ttj|&dd|ff|%dd |%d |%d d |}'||}(ttj|&dd|f|(d |(d d |})ttj|&dd|f|$d |$d d |}*ttj|&dd|f|$d |$d d |}+tj |*|+gdd },tj |*|'gdd }-tj |)|'gdd }.|,|-|.g}/t|/D]\}0}1|1||"|0f<qZq|||fS)a Calculate log likelihoods for each hypothesis, negative, singlet, doublet Parameters ---------- data : np.ndarray cells by hashing counts matrix number_of_noise_barcodes : int, number of barcodes to used to calculated noise distribution Returns ------- log_likelihoods_for_each_hypothesis : np.ndarray a 2d np.array log likelihood of each hypothesis all_indices counter_to_barcode_combo cSsd|d}t|}t|dkr.dt|n|}|||}t|dkrht||||||n|}|d||ddfS)a Update parameters of your gaussian https://www.cs.ubc.ca/~murphyk/Papers/bayesGauss.pdf Parameters ---------- data : np.array 1-d array of counts mu_o : float, global mean for hashing count distribution std_o : float, global std for hashing count distribution Returns ------- float mean of gaussian float std of gaussian rg?)lennpvarmean)dataZmu_oZstd_oZlam_onZlamZlam_nZmu_nr U/home/sam/Atlas/atlas_env/lib/python3.8/site-packages/scanpy/external/pp/_hashsolo.pygaussian_updates.s  ,z4_calculate_log_likelihoods..gaussian_updatesgV瞯`__. Parameters ---------- adata Anndata object with cell hashes in .obs columns cell_hashing_columns list specifying which columns in adata.obs are cell hashing counts priors a list of your prior for each hypothesis first element is your prior for the negative hypothesis second element is your prior for the singlet hypothesis third element is your prior for the doublet hypothesis We use [0.01, 0.8, 0.19] by default because we assume the barcodes in your cell hashing matrix are those cells which have passed QC in the transcriptome space, e.g. UMI counts, pct mito reads, etc. pre_existing_clusters column in adata.obs for how to break up demultiplexing for example leiden or cell types, not batches though number_of_noise_barcodes Use this if you wish change the number of barcodes used to create the noise distribution. The default is number of cell hashes - 2. inplace To do operation in place Returns ------- if inplace is False returns AnnData with demultiplexing results in .obs attribute otherwise does is in place Examples ------- >>> import anndata >>> import scanpy.external as sce >>> data = anndata.read("data.h5ad") >>> sce.pp.hashsolo(data, ['Hash1', 'Hash2', 'Hash3']) >>> data.obs.head() z\Please cite HashSolo paper: https://www.cell.com/cell-systems/fulltext/S2405-4712(20)30195-2z(Cell hashing counts must be non-negativeNznumber_of_noise_barcodes must be at least one less than the number of samples you have as determined by the number of cell_hashing_columns you've given as input rr'r(cluster_featureZnegative_hypothesis_probabilityZsinglet_hypothesis_probabilityZdoublet_hypothesis_probability)columnsindexrrZClassificationZDoubletZNegativer)printZobsvaluesr ValueErrorrrpdZ DataFramerrZ obs_namesuniquer.rr,r5)r/r0r-r1r!r2r Z num_of_cellsresultsZcluster_featuresZunique_cluster_featuresr4Zcluster_feature_bool_vectorZposterior_dictZ all_singsZsinglet_sample_indexr r rhashsolos3           r=)Z scipy.statsrZ scanpy._utilsr itertoolsrZanndatanumpyrZpandasr:r&r.ZAnnDatalistrintboolr=r r r rs(   ;1