Ri_/ddZddlmZGddejeZGddeZd dZdZy ) zParsing TRANSFAC files.)motifsc2eZdZdZhdZhdZfdZxZS)MotifaStore the information for one TRANSFAC motif. This class inherits from the Bio.motifs.Motif base class, as well as from a Python dictionary. All motif information found by the parser is stored as attributes of the base class when possible; see the Bio.motifs.Motif base class for a description of these attributes. All other information associated with the motif is stored as (key, value) pairs in the dictionary, where the key is the two-letter fields as found in the TRANSFAC file. References are an exception: These are stored in the .references attribute. These fields are commonly found in TRANSFAC files:: AC: Accession number AS: Accession numbers, secondary BA: Statistical basis BF: Binding factors BS: Factor binding sites underlying the matrix [sequence; SITE accession number; start position for matrix sequence; length of sequence used; number of gaps inserted; strand orientation.] CC: Comments CO: Copyright notice DE: Short factor description DR: External databases [database name: database accession number] DT: Date created/updated HC: Subfamilies HP: Superfamilies ID: Identifier NA: Name of the binding factor OC: Taxonomic classification OS: Species/Taxon OV: Older version PV: Preferred version TY: Type XX: Empty line; these are not stored in the Record. References are stored in an .references attribute, which is a list of dictionaries with the following keys:: RN: Reference number RA: Reference authors RL: Reference data RT: Reference title RX: PubMed ID For more information, see the TRANSFAC documentation. >BFBSCCDRDTHCHPOV>RARLRTRXc t||}|S#t$r!ttj||}Y|SwxYw)N)super __getitem__ TypeErrorrr)selfkeyvalue __class__s K/home/agent/.friday_env/lib/python3.12/site-packages/Bio/motifs/transfac.pyrzMotif.__getitem__FsI ?G',E  ?&,,9#>E  ?s &>>)__name__ __module__ __qualname____doc__multiple_value_keysreference_keysr __classcell__)rs@rrr s#0dK.NrceZdZdZdZdZy)RecordzStore the information in a TRANSFAC matrix table. The record inherits from a list containing the individual motifs. Attributes: - version - The version number, corresponding to the 'VV' field in the TRANSFAC file; cd|_y)zInitialize the class.N)versionrs r__init__zRecord.__init__Ys  r"ct|S)z'Turn the TRANSFAC matrix into a string.)writer's r__str__zRecord.__str__]s T{r"N)rrrrr(r+r"rr$r$Nsr"r$c i}g}d}t}|D]}|j}|s|jdd}|dj}|rt|dk7rt d|dt|dk(r8|dj} |r#|j ddst d|d|d k(r  |_n|d vri} jdd gd k7rt d |d|d} dD]} g|| < |D]c}|j}|jdd}|dj}t|dk(r8|dj} |r#|j ddst d|d t|} | dk(r| dk(r|rt d|| dz } | | k7rt d|d|rCt|dk(rt d| dd| dd|ddt|dk7rt d|d| jdd } t| d k7rt d|dtd| D]"\} }|| jt|$f|dk(r]|dk(r j d\}}}|ddk7rt d|d |d|d!d"k7rt d|d#|dt|dd!}t||dz k7r$t d|dd$t|dzdd|dd|| i}|j||d%k(r>|6td|&}|j|||_ |j|i}g}H|tjvr |<a|tjvr||vrg||<||j  ||<|S#t$rYSwxYw)'z4Parse a transfac format handle into a Record object.NrzAThe key value of a TRANSFAC motif line should have 2 characters:"" zKA TRANSFAC motif line should have 2 spaces between key and value columns: "VV)P0PO)ACGTzA TRANSFAC matrix "z(" line should be followed by "A C G T": ACGTzaA TRANSFAC matrix should start with "01" as first row of the matrix, but this matrix uses "00": "zKThe TRANSFAC matrix row number does not match the position in the matrix: "zLA TRANSFAC matrix line should have a 2 digit key at the start of the line ("02dz"), but this matrix uses "dz": "sz".z7A TRANSFAC matrix line should have a key and a value: "zQA TRANSFAC matrix line should have a value for each nucleotide (A, C, G and T): "XXRN;[z The index "z2" in a TRANSFAC RN line should start with a "[": "]z0" in a TRANSFAC RN line should end with a "]": "zP" of the TRANSFAC RN line does not match the current number of seen references "//)alphabetcounts)r$stripsplitlen ValueError partitionr&intzipappendfloatrupdate referencesr r)handlestrict annotationsrQrFrecordline key_valuerrlengthcivaluesvindex separator accession referencemotifs rreadrbbsKJ F XFx%zz| JJtQ' l  " 3x1} vQ  y>Q aL&&(E~~d+A.$ 6$ $;"FN L F{{}Ra $88 )#/..2V5F q  0 /zz| JJtQ/ l((*y>Q&%aL..0E#~~d3A6",!DDH6!L#CAQ;16(KKO&R aKF;$,,0643x1}(??@gF667U$tAhbJ 9~*(((,vQ0r*v;!#$99=aA //DAq1I$$U1X./_0 /b 4<  D[*///#*> 'E9iQx3 !%)%%)F!-RyC !%)%%)F!-a $E:%!)+ !%+J!+A.d4(!= e I   i ( D[!vf= [)#-  e$KJ E(( ("IcN E-- -+%#% C   # #E *$K qx%r MU"s N// N=<N=c g} |j}|d|z}|j| tj}d}|D]}g}|D]}d} |D]} | dk(r|j } | dk(r|j } t|j} djdg| z}|j|t| D]s}djd g| Dcgc]}d c}zd z}|t|d zg| Dcgc]}|j||c}z| |gzz}|j|ud } nR |j| }|>| |vr |D]}| d|}|j|n| d|}|j|d } | dk(s? |j}d}|D]7}|D]0} |j| }|| d|}|j|d } 29| sd}|j|d}|j|dj|dz}|j|dj|}|S#t$rY%wxYwcc}wcc}w#t$rd}Y%wxYw#t$rY-wxYw)z7Write the representation of a motif in TRANSFAC format.Nz VV %s XX // ))ACAS)ID)r CO)NA)DE)TY)OSOC)r r )r)r3)BA)r)r)r )r PVFr3rz  z%02.dz%6.20gz %sr.Tr1rn)r?rrrrr>rD )r&rNAttributeErrorrrrXdegenerate_consensussortedrEjoinrangetuplerFgetrQ)rblocksr&blockrsectionsralinessectionblankrrXsequencelettersrVrZ_rr\rQkeysr`texts rr*r*s F!..      MM% 33H"@9 #GE4 -$;"\\F{ $99H$U^^4G#==$')9:DLL&"6] +HHgYG1Lq(1L%LM() $eUG;BCau||Aq1CD'{m,'   T* +!E% % #("55%*3*-b} % T 23'*U"UG#4D!LL. $$; -%*%5%5  >)3-I'+-(1 c(:#(=$,*-b'8 % T 2(, --[4 -j T"s9 #v T % 4' eA@D 776?D KE    `2M D*% $%*sA I I2I4I I. II I+ *I+ . I; :I; N)T) rBiorrdictlistr$rbr*r,r"rrs9>FLL$>BT(Dgr"