o Nrf)@sddlZddlmZddlZddlmZmZddl Z ddl m Z m Z ddlmZd%ddZdd Zd d Zd d ZddZd&ddZGddde ZGddde ZddZddZGddde ZddZdd Zd!d"Zd#d$ZdS)'N)ListOptional) BaseModel validator)load_chromsizesc sg}t|fdddD]a}|d}ttt|dd}d}|D]B}|D]9}|d} ttt| dd} d} t|| D]\} } tt| t| dkrSd } qA| s_||g7}d }nq&|rdnq"|sm||gg7}q |S) a Partition a set of tile ids into sets of adjacent tiles. For example, if we're requesting a set of four tiles that form a rectangle, then those four tiles will become one set of adjacent tiles. Non-contiguous tiles are not grouped together. Parameters ---------- tile_ids: [str,...] A list of tile_ids (e.g. xyx.0.0.1) identifying the tiles to be retrieved dimension: int The dimensionality of the tiles Returns ------- tile_lists: [tile_ids, tile_ids] A list of tile lists, all of which have tiles that are within 1 position of another tile in the list cs dd|dddDS)NcSg|]}t|qSint).0pr r \/var/www/html/software/conda/envs/catlas/lib/python3.10/site-packages/clodius/tiles/utils.py $zApartition_by_adjacent_tiles.....rsplit)x dimensionr r$s z-partition_by_adjacent_tiles..keyrrFT)sortedrlistmapr zipabs)tile_idsrZ tile_id_liststile_id tile_id_parts tile_positionaddedZ tile_id_listZ ct_tile_idZct_tile_id_partsZct_tile_positionZ far_apartp1p2r rrpartition_by_adjacent_tiles s8     r(cCst|\}}|dks|dkrdS|dks!|dkr#dS|dkr+dS|d kr3d S|d kr;d S|d krCdSdS)Nz.bwz.bigwigbigwigz.mcoolz.coolcoolerz.htimetime-interval-jsonz.hitilehitilez.beddbbeddbz.mv5Zmultivec)opsplitextlower)filename_extr r rinfer_filetypeJs    r4cCs@|dkrdS|dkr dS|dkrdS|dkrdS|dkrd SdS) Nr*matrixr)Zvectorr+z time-intervalr,r-Zbedliker )filetyper r rinfer_datatype]sr7cCshg}|D]-}|d}t|dkrtdtt|d|d|dg\}}}||||||fg7}q|S)NrrzNot enough tile info presentrr)rlen IndexErrorrr )r!tiles_function tile_valuesr"partszryr r rtiles_wrapper_2djs  "r@c sg}t|}|D]]}t|ddd}|ddddd|D}tdd|D}tdd|D}tdd|D} td d|D} |||| ||d| | dd } |fd d| D7}q|S) aF Bundle adjacent tile requests so that they can be processed concurrently. This is helpful for function that require scanning a dataset. It's faster to filter a large region and then break it down into individual tiles than to go over the entire dataset and filter individual tiles multiple times. rrrcSs(g|]}dd|dddDqS)cSrr r )r rr r rrrz7bundled_tiles_wrapper_2d...rrrrr tr r rrs(z,bundled_tiles_wrapper_2d..cSg|]}|dqSrr rAr r rrrcSrCrDr rAr r rrrcSrCrr rAr r rrrcSrCrEr rAr r rrr)widthheightc s,g|]\}}ddtt||fqS)z{}.{}r)formatjoinrstr)r r$dataZ tileset_idr rr)r(r rminmax) r!r;r<Zpartitioned_tile_listsZ tile_group zoom_levelZtile_positionsZminxZmaxxZminyZmaxytfr rLrbundled_tiles_wrapper_2dzs"   rRrcCs|d}|d}t|d|d|d|d}|d|} |d|| } |d||| } |d|| } |d||| } | | | | gS)a Get the coordinate boundaries for the given tile. Parameters: ----------- tsinfo: { min_pos: [], max_pos [] } Tileset info containing the bounds of the dataset z: int The zoom level x: int The x position y: int The y position width: int Return bounds for a region encompassing multiple tiles height: int Return bounds for a region encompassing multiple tiles min_posmax_posrrr)rO)tsinfor>rr?rFrGrSrT max_width tile_widthZfrom_xZto_xZfrom_yZto_yr r r tile_boundss"  rXc@sVeZdZUeed<eed<eeed<eeed<edddZedddZd S) TilesetInfoZmax_zoomrVrTrScC|dkrtdt|S)2Check to make sure the zoom level is 0 or greater.rz1The zoom level must be greater than or equal to 0 ValueErrorr clsvr r rmax_zoom_zero_or_greaterz$TilesetInfo.max_zoom_zero_or_greatercCs|dkrtdt|S)z2Check to make sure the max_width is greater than 0rz$The max_width must be greater than 0r\r^r r rmax_width_greater_than_zerorbz'TilesetInfo.max_width_greater_than_zeroN) __name__ __module__ __qualname__r __annotations__rrrarcr r r rrYs    rYc@sVeZdZUeed<eeed<eeed<eeed<eeed<edddZdS) TileInfozoompositionrFstartendcCrZ)r[rz%The zoom level must be greater than 0r\r^r r rzoom_zero_or_greaterrbzTileInfo.zoom_zero_or_greaterN) rdrerfr rgrrrrmr r r rrhs     rhcs|ddd}ttt|dd}t|djdt|dfddt|ddD}fd dt|ddD}t|dd||d S) N|rrrr8rcs,g|]\}}|jdj|qSrrVrSr ipos)rUrPr rrrMz!parse_tile_id..cs0g|]\}}|jdj|qSrorprqrWrUrPr rrs)rirjrFrkrl)rrrr rV enumeraterh)r"rUr#r$ZstartsZendsr rtr parse_tile_ids"   rvc cstjdt|f}tj|||gddd\}}|||}|||}|}t||D] } | ||| fVd}q,|t|t|fVdS)a0 Convert absolute coordinates to genomic coordinates Parameters: ----------- chromsizes: [[chrom, size],...] A list of chromosome sizes associated with this tileset start_pos: int The absolute start coordinate end_pos: int The absolute end coordinate rright)ZsiderN)npZr_ZcumsumZ searchsortedranger ) chromsizes start_posend_posZabs_chrom_offsetsZcid_loZcid_hiZ rel_pos_loZ rel_pos_hirkcidr r r abs2genomics   r~c@s.eZdZUeed<eed<eed<eed<dS)ChromosomeIntervalr}namerkrlN)rdrerfr rgrJr r r rrs  rc cs\t|\}}}t|||D]\}}}z t|||||dVWqty+YdSwdS)zConvert an absolute genomic range to sections of genomic ranges. E.g. (1000,2000) => [('chr1', 1000, 1500), ('chr2', 1500, 2000)] )r}rrkrlN)rr~rr:)Zchromsizes_filenamerkrlZ chrom_infoZ chrom_namesZ chrom_sizesr}r r r abs2genome_fns  rcCs&tt||}ttt|S)au Depth of quad tree necessary to tesselate the concatenated genome with quad tiles such that linear dimension of the tiles is a preset multiple of the genomic resolution. Parameters: ----------- chromsizes: pandas.Series A series representation of the chromosome sizes tile_size_bp: int The size of each tile in the tileset )rxceilsumr log2)rzZ tile_size_bpZmin_tile_coverr r rget_quadtree_depth0s rcCs|ddkr#|d}|ddkr!|d}t|d|dSdS|ddkr,dStdtj}tdd||D}tdd||D}d D]}||vrXdS||vradSqMz||krjWdS||krqWdSWdSty}YdSw) Nr2rrz(\d+)cS$g|]}|r|rt|n|qSr isdigitr r ar r rrS$znatcmp..cSrr rrr r rrTr)mr?r) findrnatcmprecompileUtupler0 TypeError)rr?Zx_partsZy_partsZ _NS_REGEXrr r rrAs4     rcCst|ttdS)z3 Sort an iterable by natural genomic order r)rft cmp_to_keyr)iterabler r r natsortedisrro)rr) functoolsrZos.pathpathr.rtypingrrnumpyrxZpydanticrrZclodius.chromosomesrr(r4r7r@rRrXrYrhrvr~rrrrrr r r rs,   >  %" (