o Uݢg<@sddlmZddlZddlZddlZddlmZddlmZddl Z ddl Z ddl Z ddlmZmZddlmmZddlmmZddlmZddlmmZddlmm m!Z"ergddlm#Z#ddl$m%Z%m&Z&m'Z'm(Z(dZ)e*d gd Z+d*d+ddZ, d,d+ddZ-GdddZ.d+ddZ/d+ddZ0ddZ1ddZ2ddZ3ddZ4d d!Z5d-d$d%Z6d&d'Z7Gd(d)d)Z8dS).) annotationsN)Sequence) TYPE_CHECKING) ensure_binary ensure_str) Projection)AggrCountSamplePropertiesCountSamplePropertiesExtendedCountSamplePropertiesSampleProperties filtered_bcs_transcriptome_unionBarcodeRankPlotSegmentstartend cell_densitylegendF cell_barcodes set[bytes]csN||ksJtfddt||D}t|t||}t||||dS)z(Helper function to build a plot segment.c3s |] }|vrdVqdS)N).0ir sorted_bcrf/oak/stanford/groups/akundaje/marinovg/programs/cellranger-9.0.1/lib/python/cellranger/webshim/data.py /sz#get_plot_segment..r)sumrangefloatr ) start_index end_indexrrr num_cellsdensityrrrget_plot_segment,s r$c Cs|r&tt|ddd|}||dd|}|ddd|}n||dd}|ddd}t|||}||}||} ~t|| |} | | fS)aA helper function to generate the data required to generate the barcode rank. plot for RNA counter. The barcode rank plot consists of multiple plot segments If the barcodes are ordered by UMI counts - All the barcodes until the first non cell barcodes go into the `Cell` segment. - All the barcodes beyond the last cell barcode goes into the `Background` segment. - The remaining barcodes are further divided into multiple plot segments Input: - cell_barcodes: A set of bytes with the cell barcodes - barcode_summary: The barcode summary from the h5. - key: Specifies the entry to look up in barcode summary for getting the UMI counts - restrict_barcodes: Optional list of cell barcodes to restrict to Output: - sorted_counts: UMI counts sorted in descending order - plot_segments: List of BarcodeRankPlotSegment bc_sequenceN)npnonzeroisincompute_sort_ordercompute_plot_segments) rbarcode_summarykeyrestrict_barcodesZrestrict_indices counts_per_bcZbarcode_sequences srt_orderr sorted_counts plot_segmentsrrr'generate_counter_barcode_rank_plot_data6s  r2c@s>eZdZdddZddZddZd d Zd d Zd dZdS) SampleData projectionsSequence[Projection]cCs|r t|ts Jt|j|_|jd|_t|j|||\|_ |_ t |j |_ t|do1|jdu|_|jd|_t|trB|jnd|_t|trM|jnd|_|jr[tt|jnd|_|jrvtj|jdkrvtjt|jdtidnd|_|j rt!"|j nd|_#|j$r|j#dusJdd t%&|j$D|_#t'|j(|_)t*|j+|_,t-|j.|_/t*|j0|_1t*|j2|_3t*|j4|_5t*|j6|_7dS) Nr target_settargeting_methodFrbarcode) converterscSsh|]}t|qSr)r)rbcrrr sz&SampleData.__init__..)8 isinstancer load_metrics_summary summary_pathsummarygetr" load_analyses analysis_pathanalysesoriginal_cluster_sizesload_barcode_summarybarcode_summary_pathr+hasattrr6has_target_setr7r Z is_visium_hd filter_probesZvdj_clonotype_summary_pathpdread_csvrvdj_clonotype_summaryZvdj_barcode_support_pathospathgetsizervdj_barcode_supportfiltered_barcodes_pathcr_utilsget_cell_associated_barcode_setrZvdj_cell_barcodes_path vdj_utilsload_cell_barcodes_jsonload_feature_countsfeature_metrics_pathZfeature_metricsload_antibody_dataantibody_histograms_pathantibody_histogramsload_treemap_dataantibody_treemap_pathantibody_treemapantigen_histograms_pathantigen_histogramsZraw_normalized_heatmap_pathZraw_normalized_heatmapZisotype_scatter_pathZisotype_scatterZ gex_fbc_correlation_heatmap_pathZgex_fbc_correlation_heatmap)selfsample_propertiessample_data_pathsplot_preprocess_funcr4rrr__init__bsl           zSampleData.__init__csV|jdurdSfdd|jD}t|dkst|dksJt|dkr'dS|dS)Ncsg|] }t|r|qSr)r<)ra analysis_typerr sz+SampleData.get_analysis..rr)rClen)r`rgrrrfr get_analysiss  zSampleData.get_analysiscCs|jdusJt|j|j|S)a!Generate the data required to generate the barcode rank. plot for RNA counter. The barcode rank plot consists of multiple plot segments If the barcodes are ordered by UMI counts - All the barcodes until the first non cell barcodes go into the `Cell` segment. - All the barcodes beyond the last cell barcode goes into the `Background` segment. - The remaining barcodes are further divided into multiple plot segments Input: - key: Specifies the entry to look up in barcode summary for getting the UMI counts Output: - sorted_counts: UMI counts sorted in descending order - plot_segments: List of BarcodeRankPlotSegment N)rr2r+)r`r,rrrcounter_barcode_rank_plot_datas z)SampleData.counter_barcode_rank_plot_datacCsb|jdusJ|jd}t||jd|j}|jd|}||}~t|||j}||fS)zEGenerate the data required to generate the barcode rank plot for VDJ.Ncountr8)rrPto_numpyr)r*)r`r.r/rr0r1rrrvdj_barcode_rank_plot_datasz%SampleData.vdj_barcode_rank_plot_datacC|jSN)rHr`rrr is_targetedzSampleData.is_targetedcCs|jduo t|jvS)zReturns true if data is antibody only, currently determined by the presence of a metric. Returns: bool: Whether or not the data is antibody-only. N)r? FILTERED_BCS_TRANSCRIPTOME_UNIONrrrrris_antibody_onlyszSampleData.is_antibody_onlyNr4r5) __name__ __module__ __qualname__rdrkrlrorsrvrrrrr3as P  r3csftt|ks Jttdt|D] \}}||v|<qtttfdddd}|S)NFcs||fSrqr)xr.is_cellrrsz$compute_sort_order..T)r,reverse)rir&full enumeratesortedr)r.Z bc_sequencesrrr:r/rr|rr)sr)c Cst|}t|D] \}}||vr|}nqd}ttt|D] }|||vr+|}nqd||dt|g}g}|td|dddd|t|d|ddddt||d|d} tt| dD]}|t| || |d||d d qf|S) Nrrg?TrF)r)rirreversedrappendr segment_log_plot_by_lengthr$) rr0rZfirst_non_cellrr:Z last_cellrangesr1Zmixed_segmentsrrrr*s8 r*c Cs||krgSd}d}tt|}tt|}d}|g}t||D]I} | dkr)q"t|| d} t| t| |} t|| t|| |} |tj| | g7}||krk| |d|krk|| dd}q"|d|krw|||S)zGiven the extends of the mixed region [x_start, x_end), compute. the x-indices that would divide the plot into segments of a prescribed length (in pixel coordinates) with a minimum number of barcodes in each segment g{Gz?rrr)r&logrimaxrlinalgnormr) Zy_dataZx_startZx_endZSEGMENT_NORMALIZED_MAX_LENZ MIN_X_SPANZ log_max_xZ log_max_yZthis_segment_lenZ segment_idxrZlast_idxdyrrrrs,   rcCB|durdSt| }t|WdS1swYdSrqopenjsonload)r>frrrr=C  $r=cCs|durdSt|Srq)SummaryFeatureCounts from_file)rWrrrrVJs rVcCrrqr)rYrrrrrXPrrXcCrrqr)r\rrrrr[Wrr[r4r5cCs|durdSt|tr,t|jdkrtjj|d|dg}ntjj|d|dtj|g}|ddur4dSd}|d}|duret|tjre| }t|toP|j }|j j t jkre|s_t|tre|t j|||fS)z2Returns (analysis_object, original_cluster_sizes).N)NNrpca)methodr4r)r<r rigenomescr_sg_analysisSingleGenomeAnalysisload_default_formatcr_mg_analysisMultiGenomeAnalysisget_cluster_sizesZ is_spatialmatrixbcs_dim gex_constantsMAX_WEBSHIM_BCS_DIMr subsample_bcs)base_dirrcrar4rCrDanalysisZis_single_cellrrrrA^s:      rAcCs|durdSt|dS)Nrj)h5pyFile)h5_pathrrrrEs rEc@sDeZdZdZddddZeddd Zd d Zdd d ZddZ dS)rzJStores and queries a dataframe that summarizes relevant features per gene.Ndf pd.DataFramecCs$||_|djdd|_||_dS)a Stores and queries a dataframe that summarizes relevant features per gene. There should be one row per gene and any number of columns. Args: df (pandas dataframe): Dataframe with n_genes rows and columns with features per genes. feature_idindex)orientN)r set_indexto_dict feature_dictfilepath)r`rrrrrrds zSummaryFeatureCounts.__init__returncCs,t|}tj|s J|tt||S)z#Loads the features from a csv file.)rrMrNisfilerJrK)clsrNrrrrszSummaryFeatureCounts.from_filecCst|}|jj|dddS)NF)r)rrto_csv)r`rNrrrrszSummaryFeatureCounts.to_csvcCrprq)rrrrrrget_dfrtzSummaryFeatureCounts.get_dfcCs.||jvrdS||j|vrdS|j||S)z(Get the given key for the given feature.N)r)r`rr,rrrget_value_for_features z*SummaryFeatureCounts.get_value_for_featurerq)rr)rr)rr) rxryrz__doc__rd classmethodrrrrrrrrrs    r)F)rrrqrw)9 __future__r collectionsrrMcollections.abcrtypingrrnumpyr&pandasrJsixrrcellranger.analysis.multigenomer multigenomer cellranger.analysis.singlegenome singlegenomercellranger.utilsutilsrRcellranger.vdj.utilsvdjrT cellranger.webshim.constants.gexwebshim constantsgexrr'cellranger.websummary.sample_propertiesrr r r ru namedtupler r$r2r3r)r*rr=rVrXr[rArErrrrrsJ       +  %$ /