DUf^ddlZddlZddlZddlZddlZddlZdZdZGddeZ dS)Nc(tjS)N)timec/var/www/html/software/conda/lib/python3.11/site-packages/pybedtools/contrib/intersection_matrix.pynowr s 9;;rctjtj|dS)Nr)ospathsplitextbasename)fnames rget_namer s- 7  BG,,U33 4 4Q 77rcHeZdZdZd dZd dZdZdZdZd Z d d Z d Z dS)IntersectionMatrixzE Class to handle many pairwise comparisons of interval files NFc||_||_||_||_|jrd||t j||_t j|j_ |j |_ dSdS)a Class to handle and keep track of many pairwise comparisons of interval files. A lightweight database approach is used to minimize computational time. The database stores filenames and calculation timestamps; re-calculating a matrix using the same interval files will only re-calculate values for those files whose modification times are newer than the timestamp in the database. `beds` is a list of bed files. `genome` is the string assembly name, e.g., "hg19" or "dm3". `dbfn` is the filename of the database you'd like to use to track what's been completed. Example usage: First, get a list of bed files to use: #>>> beds = [ #... pybedtools.example_filename(i) for i in [ #... 'Cp190_Kc_Bushey_2009.bed', #... 'CTCF_Kc_Bushey_2009.bed', #... 'SuHw_Kc_Bushey_2009.bed', #... 'BEAF_Kc_Bushey_2009.bed' #... ]] Set some parameters. "dm3" is the genome to use; info will be stored in "ex.db". `force=True` means to overwrite what's in the database #>>> # In practice, you'll want many more iterations... #>>> im = IntersectionMatrix(beds, 'dm3', #... dbfn='ex.db', iterations=3, force=True) #>>> # Use 4 CPUs for randomization #>>> matrix = im.create_matrix(verbose=True, processes=4) N) bedsgenomedbfn iterations_init_dbsqlite3connectconnRow row_factorycursorc)selfrrrrforces r__init__zIntersectionMatrix.__init__s}L   $ 9 ( MM% --DI$+KDI !Y%%''DFFF  ( (rcB|jdStj|jr|sdSt j|j}|}|r|d|d| dS)zB Prepare the database if it doesn't already exist Nz#DROP TABLE IF EXISTS intersections;a CREATE TABLE intersections ( filea TEXT, fileb TEXT, timestamp FLOAT, actual FLOAT, median FLOAT, iterations INT, self INT, other INT, fractionabove FLOAT, fractionbelow FLOAT, percentile FLOAT, PRIMARY KEY (filea, fileb, iterations)); ) rr r existsrrrexecute executescriptcommit)rr rrs rrzIntersectionMatrix._init_dbGs 9  F 7>>$) $ $ U  Fty)) KKMM  = II; < < <     " rc|jdSt|jdt }t |dkrdSt |dksJ|dS)z Return the sqlite3.Row from the database corresponding to files `fa` and `fb`; returns None if not found. Nz SELECT * FROM intersections WHERE filea=:fa AND fileb=:fb AND iterations=:iterations r)rlistrr$localslen)rfafbrresultss rget_rowzIntersectionMatrix.get_rowfsz 9  F FNN        w<<1   F7||q    qzrc||||}|rXtj|}tj|}|d|kr|d|krdSdS)z Retrieves row from db and only returns True if there's something in there and the timestamp is newer than the input files. timestampTF)r/r r getmtime)rr,r-rrowtfatfbs rdonezIntersectionMatrix.done}st ll2r:..  '""2&&C'""2&&CK 3&&S-=-C-Cturc tj||j}|j|d<|j|fi|}||dS)Nr) pybedtoolsBedToolset_chromsizesrr randomstatsadd_row)rr,r-kwargsar.s rrun_and_insertz!IntersectionMatrix.run_and_inserts_  r " " 1 1$+ > >#|!---f-- Wrcgd}|D]\}}||||<t|d<d}|j|||jdS)a Inserts data into db. `results` is a dictionary as returned by BedTool.randomstats with keys like:: 'iterations' 'actual' 'file_a' 'file_b' self.fn other.fn 'self' 'other' 'frac randomized above actual' 'frac randomized below actual' 'median randomized' 'normalized' 'percentile' 'lower_%sth' % lower_thresh 'upper_%sth' % upper_thresh ))file_afilea)file_bfileb)zmedian randomizedmedian)zfrac randomized above actual fractionabove)zfrac randomized below actual fractionbelowr1a< INSERT OR REPLACE INTO intersections ( filea, fileb, timestamp, actual, median, iterations, self, other, fractionabove, fractionbelow, percentile) VALUES ( :filea, :fileb, :timestamp, :actual, :median, :iterations, :self, :other, :fractionabove, :fractionbelow, :percentile) N)rrr$rr&)rr. translationsorignewsqls rr<zIntersectionMatrix.add_rows~,   & ) )ID#"4=GCLL"uu  < sG$$$ rc t|j}|dz}d}tjt}|jD]}|jD]}|dz }|rLt jdtzt j | |||j s|j ||fi|| |||j |t|t|<|S)a Matrix (implemented as a dictionary), where the final values are sqlite3.ROW objects from the database:: { filea: { filea: ROW, fileb: ROW, ...}, fileb: { filea: ROW, fileb: ROW, ...}, } } rr(z$%(i)s of %(total)s: %(fa)s + %(fb)s )r+r collections defaultdictdictsysstderrwriter*flushr6rr?r/r) rverboser=nfilestotalimatrixr,r-s r create_matrixz IntersectionMatrix.create_matrixs$TY!  (..)  Bi  Q'J$$%Lvxx%WXXXJ$$&&&yyR99:'D'B99&99959\\DO66x||$Xb\\22  rcdS)a: Prints a pairwise matrix of values. `matrix` is a dict-of-dicts from create_matrix(), and `key` is a field name from the database -- one of: ['filea', 'fileb', 'timestamp', 'actual', 'median', 'iterations', 'self', 'other', 'fractionabove', 'fractionbelow', 'percentile'] Nr)rrYkeys r print_matrixzIntersectionMatrix.print_matrixsr)NF)F) __name__ __module__ __qualname____doc__r!rr/r6r?r<rZr]rrrrrs/(/(/(/(b>.    AAAF%%%%N     rr) r rQrr8rrNrrobjectrrrrrcs  888p p p p p p p p p p r