o Nrf)@sddlZddlZddlZddlZddlZddlZddl m m Z ddl mZmZddlmZdZdZdZdZeeZiZddded<dd d Zd d Z    dd dZidfddZ ddZdS)N) abs2genomicget_quadtree_depth)ThreadPoolExecutor significantZ Significant)namevaluecCst||}t|d<|S)N range_modes)hgbi tileset_infor )bbpath chromsizestir]/var/www/html/software/conda/envs/catlas/lib/python3.10/site-packages/clodius/tiles/bigbed.pyr s r c sl|\ }}}}}}}}} z$|j|} ttj|| || } |dkr(| | } } n| | } } Wnty=d\} } Yn tyId\} } Ynwd}d}i}|D]}|||j|<||7}|d7}qRg}d}g}| sn|S| D]"}z |t|dWnt tfy|t Ynw|d7}qp||kr||kr| D]+}zt|d}||| ||dWqt tfy||| t |dYqw|S||kr4g||}tt |d|}| D]5}zt|d}||kr|| ||dWqt tfyt |kr|| t |dYqwt }||kr4tt||}fddt|D}|S) Nr)NNrr)Z chrOffsetZ importancefieldscsg|]}|qSrr).0iZthresholded_intervalsrr zfetch_data..)indexftpartialbbiZfetch_intervals IndexErrorKeyErrorappendint ValueError DEFAULT_SCOREnpZquantilelenrandomsamplerangesorted)ar binsizer range_mode min_elements max_elementsZcidstartendZchromZ fetch_factoryZ intervalsZ intervals2offsetZ offsetIdxZ chrOffsetsZchrSizeZfinal_intervalsZintervals_lengthZscoresintervalZscoreZ desired_percZthresholded_scoreZthresholded_intervals_lengthindicesrrr fetch_datas *        &    r3c sdur tdurtdurtt}||tt||} tdd} t| t fdd| D} Wdn1sJwYdd| D} dd| DS)N) max_workersc s(g|]}tgt|qSr)tuplelistrcr r*rr-r,r+rrrs z#get_bigbed_tile..cSsg|]}|gkr|qSrr)rxrrrrcSsg|] }|D]}|qqSrr)rZsublistitemrrrrs) r get_chromsizes MIN_ELEMENTS MAX_ELEMENTSZget_zoom_resolutionsr7rrmapr3) r zoom_level start_posend_posrr+r,r- resolutionsZcids_starts_endseZarraysresultsrr:rget_bigbed_tiles*    rHc Csd}d}g}|D]}|ddd}|dd} | d} ttt| dd} t| dkr4| dnt} | tvr<| nd} tdd |D}d |vrQt|d }d |vr[t|d }||krf|}|}|}n ||krp|}|d}|dkrxt}t }|rd d |D}d d |D}t j ||d}nd}d|vr|d}||vr||}nd}| d}| d}|durt |}t|t j}t jd||}||}||}t|||||| ||d}|||fg7}q|S)a  Generate tiles from a bigbed file. Parameters ---------- tileset: tilesets.models.Tileset object The tileset that the tile ids should be retrieved from tile_ids: [str,...] A list of tile_ids (e.g. xyx.0.0) identifying the tiles to be retrieved chromsizes_map: {uid: []} A set of chromsizes listings corresponding to the parameters of the tile_ids. To be used if a chromsizes id is passed in with the tile id with the `|cos:id` tag in the tile id chromsizes: [[chrom, size],...] A 2d array containing chromosome names and sizes. Overrides the chromsizes in chromsizes_map Returns ------- tile_list: [(tile_id, tile_data),...] A list of tile_id, tile_data tuples |rNr.cSsg|]}|dqS):)split)rorrrr sztiles..minmaxcSsg|]}|dqS)rrr8rrrr rcSsg|]}t|dqS)r)r r8rrrr!r<)rcos)r+r,r-)rNr7rAr r$DEFAULT_RANGE_MODEr dictr?r@pdZSeriesr r>rZ TILE_SIZErH)r Ztile_idsZchromsizes_maprr,r-Zgenerated_tilesZtile_idZtile_option_partsZtile_no_optionsZ tile_id_partsZ tile_position return_valuer+Z tile_optionsZtemp_min_elementsZ chromnamesZ chromlengthsZchromsizes_to_useZ chromsizes_idrBZtile_pos max_depthZ tile_sizerCrDZ tile_valuerrrtilessp       rYcCs t|SN)r r)filenamerrrrIs rrZ)NNNN)r functoolsrloggingnumpyr#ZpandasrVr%Zclodius.tiles.bigwigrYZbigwigr utilsrrconcurrent.futuresrrTr?r@r" getLogger__name__loggerr r r3rHrrrrrs4    1 e