DUf ~ddlZddlZddlmZddlmZddZdZ dd Z ede d d dd d d e fdZ dS)N)is_cooler_balanced)pool_decoratorbalancedct|d}||d||dz|dz|d|<|S)z Multiply raw pixel counts by the balancing bias and return a modified chunk with an additional column named balanced_column_name pixelsbin1_idbin2_idcount)chunkbiasbalanced_column_namers S/var/www/html/software/conda/lib/python3.11/site-packages/cooltools/api/coverage.py_apply_balancingrsG 8_F,0 1B,Cd6R[K\F],]`fgn`o,oE(O() Lc|dkrDtj|dd|ddz |k}d|dd|<|S)Nrrr r r )npabs)r n_diagsmasks r _zero_diagsrsO{{veHoi05?93MMNNQXX)*h & Lrr cD|d}|d}t|d}tjd|f}||}|d|d|d|dk}|dxxtj|d||z|z cc<|dxxtj|d||z|z cc<|d xxtj|d||z cc<|d xxtj|d||z cc<|S) a0 Compute cis and total coverages of a cooler chunk. Every interaction is contributing to the "coverage" twice: at its row coordinate bin1_id, and at its column coordinate bin2_id Parameters ---------- chunk : dict of dict/pd.DataFrame A cooler chunk produced by the cooler split-apply-combine pipeline. pixel_weight_key: str The key of a pixel chunk to retrieve pixel weights. Returns ------- covs : np.array 2 x n_bins A numpy array with cis (the first row) and total (the 4nd) coverages. binsrchromrr r r)weights minlength)lenrzerosbincount)r pixel_weight_keyrrn_binscovs pixel_weightscis_masks r_get_chunk_coverager's@& =D 8_F g  F 8QK D+,MG}VI./4= AR3SSHGGGr{y=8#;vGGG GGGr{y=8#;vGGG GGGr{6),mvVVVVGGGGGGr{6),mvVVVVGGG KrgcAFcovrc 6 ||n5||jddzd}n#tdxYw|d} npt ||rM||dd} t j| } t j| } d} ntd|d tj |||| } |r| t| } || t| | } |jd } | t | t jt j| }|%|t*}|d|dg}n!t j|dd| f<|d||d|g}|r|d5}|t*nt0}t3||D]C\}}||dvr |d|=t5dd}|dj|fd|i|d|iD|=|jdt j|ddzt*dddn #1swxYwY|S)a Calculate the sums of cis and genome-wide contacts (aka coverage aka marginals) for a sparse Hi-C contact map in Cooler HDF5 format. Note that for raw coverage (i.e. clr_weight_name=None) the sum(tot_cov) from this function is two times the number of reads contributing to the cooler, as each side contributes to the coverage. Parameters ---------- clr : cooler.Cooler Cooler object ignore_diags : int, optional Drop elements occurring on the first ``ignore_diags`` diagonals of the matrix (including the main diagonal). If None, equals the number of diagonals ignored during IC balancing. chunksize : int, optional Split the contact matrix pixel records into equally sized chunks to save memory and/or parallelize. Default is 10^7 clr_weight_name : str Name of the weight column. Specify to calculate coverage of balanced cooler. store : bool, optional If True, store the results in the input cooler file when finished. If clr_weight_name=None, also stores total cis counts in the cooler info. Default is False. store_prefix : str, optional Name prefix of the columns of the bin table to save cis and total coverages. Will add suffixes _cis and _tot, as well as _raw in the default case or _clr_weight_name if specified. nproc : int, optional How many processes to use for calculation. Ignored if map_functor is passed. map_functor : callable, optional Map function to dispatch the matrix chunks to workers. If left unspecified, pool_decorator applies the following defaults: if nproc>1 this defaults to multiprocess.Pool; If nproc=1 this defaults the builtin map. Returns ------- cis_cov : 1D array, whose shape is the number of bins in ``h5``. Vector of bin sums in cis. tot_cov : 1D array, whose shape is the number of bins in ``h5``. Vector of bin sums. N/z /bins/weight ignore_diagszjPlease, specify ignore_diags and/or IC balance this cooler! Cannot access the value used in IC balancing. r rz+cooler is not balanced, orbalancing weight z is not available in the cooler.) chunksizemapuse_lock)r)rrnbins)r"_cis_raw_tot_raw_cis__tot_zr+rgzip) compressioncompression_optsdatadtypecisrr)r9) _load_attrsrootrstrip ValueErrorrrrisnan nan_to_numcoolerparallelsplitpiperrinfor'reduceaddr astypeintnanopenfloatzipdictcreate_datasetattrscreatesum)clrr+r,r.clr_weight_namestore store_prefixnproc map_functorr"r bias_na_maskchunksr#r$ store_namesgrpr9 store_namecov_arrh5optss rcoverager`As`j  ' L!5!5!FGGW    x   " C 1 1  xxzz/*111-x~~ }T""% R / R R R   _ " "3)W_ " ` `F@[,??"-DO_`` Xg F ;;*=M; N N U UVXV\^`^fgm^n^n o oD{{3&000\2K2K2KM "QQQ _&>>_>><@g@gVe@g@gi  I XXd^^ Is*2CCE'*;'='=  # GV,,F J/&1EEE*F *%,06>C&   tAw1(res++++++''''''$$$$Lc#hh   kkkkkkr