tf)ddlZddlmZddlZddlmZmZddl Z ddl m Z m Z ddlmZddZdZdZdZd Zdd ZGd d e ZGd de ZdZdZGdde ZdZdZdZdZy)N)ListOptional) BaseModel validator)load_chromsizesc g}t|fdD]}|jd}ttt|dd}d}|D]}|D]z}|jd} ttt| ddz} d} t || D]+\} } t t | t | z dkDs*d} -| rr||gz }d}n|sn|r||ggz }|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 cj|jdddzDcgc] }t|c}Scc}w)N.)splitint)xp dimensions ]/var/www/html/software/conda/envs/higlass/lib/python3.12/site-packages/clodius/tiles/utils.pyz-partition_by_adjacent_tiles..$s,a!i-1P QAQ Q Qs0keyr r FT)sortedr listmapr zipabs)tile_idsr tile_id_liststile_id tile_id_parts tile_positionadded tile_id_list ct_tile_idct_tile_id_partsct_tile_position far_apartp1p2s ` rpartition_by_adjacent_tilesr) s*MQ") c* SmAa&89: ) L*  #-#3#3C#8 #'C1A!a)m1T(U#V ! "-1AB)FB3r7SW,-1$( ) ! WI-L E ") * wi[ (ME")H cptj|\}}|jdk(s|jdk(ry|jdk(s|jdk(ry|jdk(ry|jd k(ry |jd k(ry |jd k(ryy)Nz.bwz.bigwigbigwigz.mcoolz.coolcoolerz.htimetime-interval-jsonz.hitilehitilez.beddbbeddbz.mv5multivec)opsplitextlower)filename_exts rinfer_filetyper8Js [[ "FAs yy{esyy{i7  CIIK7$:  #  !     r*c@|dk(ry|dk(ry|dk(ry|dk(ry|dk(ry y) Nr-matrixr,vectorr.z time-intervalr/r0bedlike)filetypes rinfer_datatyper?]s?88''87r*cg}|D]\}|jd}t|dkr tdtt|d|d|dg\}}}||||||fgz }^|S)Nr rzNot enough tile info presentrr )r len IndexErrorrr )rtiles_function tile_valuesrpartszrys rtiles_wrapper_2drIjsK< c" u:>;< <cE!HeAha9:1a.Aq"9:;; < r*cg}t|}|D]D}t|djdd}|djdd}|Dcgc]-}|jdddDcgc] }t|c}/} }}t| Dcgc]}|d c}} t | Dcgc]}|d c}} t| Dcgc]}|d c}} t | Dcgc]}|d c}} ||| | | | z dz| | z dz}||Dcgc]6\}}dj |dj tt||f8c}}z }G|Scc}wcc}}wcc}wcc}wcc}wcc}wcc}}w)aF Bundle adjacent tile requests so that they can be processed concurrently. This is helpful for function that require scanning a dataset. It's faster to filter a large region and then break it down into individual tiles than to go over the entire dataset and filter individual tiles multiple times. rr rr r)widthheightz{}.{}) r)r r minmaxformatjoinrstr)rrDrEpartitioned_tile_lists tile_group zoom_level tileset_idtrtile_positionsminxmaxxminymaxytfr datas rbundled_tiles_wrapper_2dr^zs|K8B, A,,S1!45 ]((-a0 GQR!1773<!+<=a3q6=RR.1QAaD12.1QAaD12.1QAaD12.1QAaD12  d$+/$+PQ/  )+ %^^JS-1H(I JD Q  ! * #>R1111  s6E+E=E E% * E*  E/ " E4 ;E9 Ec|d}|d}t|d|dz |d|dz }|d|zz } |d|| zz} |d||z| zz} |d|| zz} |d||z| zz} | | | | gS)a Get the coordinate boundaries for the given tile. Parameters: ----------- tsinfo: { min_pos: [], max_pos [] } Tileset info containing the bounds of the dataset z: int The zoom level x: int The x position y: int The y position width: int Return bounds for a region encompassing multiple tiles height: int Return bounds for a region encompassing multiple tiles min_posmax_posrrr )rN)tsinforGrrHrKrLr`ra max_width tile_widthfrom_xto_xfrom_yto_ys r tile_boundsris&YGYGGAJ+WQZ'!*-DEIQ!V#J QZ!j. (F 1:Uj0 0D QZ!j. (F 1:Vz1 1D FD$ ''r*czeZdZUeed<eed<eeed<eeed<eddZeddZy) TilesetInfomax_zoomrcrar`c8|dkr tdt|S)2Check to make sure the zoom level is 0 or greater.rz1The zoom level must be greater than or equal to 0 ValueErrorr clsvs rmax_zoom_zero_or_greaterz$TilesetInfo.max_zoom_zero_or_greaters  q5PQ Q1v r*c8|dkr tdt|S)z2Check to make sure the max_width is greater than 0rz$The max_width must be greater than 0rorqs rmax_width_greater_than_zeroz'TilesetInfo.max_width_greater_than_zeros  6CD D1v r*N) __name__ __module__ __qualname__r __annotations__rrrtrvr=r*rrkrksPMN #Y #Yz {r*rkcteZdZUeed<eeed<eeed<eeed<eeed<eddZy)TileInfozoompositionrKstartendc8|dkr tdt|S)rnrz%The zoom level must be greater than 0rorqs rzoom_zero_or_greaterzTileInfo.zoom_zero_or_greaters  q5DE E1v r*N) rwrxryr rzrrrrr=r*rr|r|sE I3i C= 9 cNvr*r|c(|jddjd}ttt|dd}t|d}|jdt|dzz }t |ddDcgc]*\}}||jd|zz z|j |z,}}}t |ddDcgc]-\}}||jd|zz z|j |z|z/} }}t||dd||| Scc}}wcc}}w)N|rr rrAr )r}r~rKrr)r rrr rc enumerater`r|) rrbrr rTrdiposstartsendss r parse_tile_idrs<MM#&q)//4MS-!"456M]1%&J!!A]1-=)>$>>J"-"34 Q v!z/12V^^A5FFF "-"34  Q   1 ?2 3fnnQ6G G* T D   qr"    s =/D2Dc#,Ktjdtj|f}tj|||gddz \}}|||z }|||z }|}t ||D]} | ||| fd}|t |t |fyw)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)siderN)npr_cumsum searchsortedranger ) chromsizes start_posend_posabs_chrom_offsetscid_locid_hi rel_pos_lo rel_pos_hircids r abs2genomicrsa:!667 )Iw+?gNQRRFF.v66J,V44J EVV$5*S/)) #e*c*o --sBBc6eZdZUeed<eed<eed<eed<y)ChromosomeIntervalrnamerrN)rwrxryr rzrQr=r*rrrs H I J Hr*rc#Kt|\}}}t|||D]\}}} t|||||y#t$rYywxYww)zConvert an absolute genomic range to sections of genomic ranges. E.g. (1000,2000) => [('chr1', 1000, 1500), ('chr2', 1500, 2000)] )rrrrN)rrrrC)chromsizes_filenamerr chrom_info chrom_names chrom_sizesrs r abs2genome_fnrsk .==P-Q*Zk&{E3?UC $k#.e     s%$A>A A A A  Actjt||z }ttjtj|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 )rceilsumr log2)r tile_size_bpmin_tile_covers rget_quadtree_depthr0s8WWS_|;rs~ !)/;|&   "J(D)* y 2.4  "1"%P7r*