ViK$6SrSSKJr SSKrSSKJr SSKJrJrJ r J r J r J r J r JrJr \"S5rS\l\"S 5rS \lS r"S S \5rSr"SS\5r"SS\5r"SS\5r"SS\5r"SS\5r"SS\5rg)z Group mode )divisionN) make_sentinel) glomMODESKIPSTOPTargetRegistryPathTBadSpec_MISSINGACC_TREEz tree of accumulators for aggregation; structure roughly corresponds to the result, but is not 1:1; instead the main purpose is to ensure data is kept until the Group() finishes executing CUR_AGGaq the spec which is currently performing aggregation -- useful for specs that want to work in either "aggregate" mode, or "spec" mode depending on if they are in Group mode or not; this sentinel in the Scope allows a spec to decide if it is "closest" to the Group and so should behave like an aggregate, or if it is further away and so should have normal spec behavior. c U[RSX[S9nU"U5nU$![a>n[ SUR R <S[U[6<SU<S35eSnAff=f)Niterate)pathz&failed to iterate on instance of type z at z (got ))r get_handlerr Exception TypeError __class____name__)targetscoperiteratores `/home/james-whalen/.local/share/pipx/venvs/semgrep/lib/python3.13/site-packages/glom/grouping.py target_iterr!s{N#// 6d /TGN6? O N!++44dE$K6H!MN NNs* A29A--A2c*\rSrSrSrSrSrSrSrg)Group,aVsupports nesting grouping operations -- think of a glom-style recursive boltons.iterutils.bucketize the "branches" of a Group spec are dicts; the leaves are lists, or an Aggregation object an Aggregation object is any object that defines the method agg(target, accumulator) For example, here we get a map of even and odd counts:: >>> glom(range(10), Group({lambda x: x % 2: T})) {0: 8, 1: 9} And here we create a `"bucketized" `_ map of even and odd numbers:: >>> glom(range(10), Group({lambda x: x % 2: [T]})) {0: [0, 2, 4, 6, 8], 1: [1, 3, 5, 7, 9]} target is the current target, accumulator is a dict maintained by Group mode unlike Iter(), Group() converts an iterable target into a single result; Iter() converts an iterable target into an iterable result cXlgNspec)selfr&s r__init__Group.__init__I c6[U['SU['0U['[ UR 5[ [4;a[ UR 5"5nOSn[X5H*nX2["X@R U5p5U[LdM(Us $ U$r$) GROUPrrrtyper&dictlistrrr )r'rrrettlasts rglomit Group.glomitLs~d gh  ?tTl *tyy/#CCV+A4[IIu=#d{ , r+cVURRnU<SUR<S3$N(r)rrr&)r'cns r__repr__Group.__repr__^s ^^ $ $tyy))r+r%N) r __module__ __qualname____firstlineno____doc__r(r4r:__static_attributes__r+rr!r!,s8$*r+r!c^^UU4SjnT[n[[USS55(aURTU5$[U5(aU"T5$[ U5nU[ [ 4;a[SU<35e[U5nXFnU[ LaSnUR5HupURU S5[LaM U"U 5n U [LaSnM5U [La [XI'MHX;a0XK'XKT['U"U 5n U [La [XI'MwSnU [LdMXU 'M U(a[$U$U[ La`UHXn [ U 5[ La[SU-5eU"U 5n U [La[s $U [LdMGURU 5 MZ U$[SR!U55e![a U"5=otU'GNQf=f) z? Group mode dispatcher; also sentinel for current mode = group c&>T["TUT5$r$)r)r&rrs rGROUP..gs5;vtU;r+aggNz>Group mode expected dict, list, callable, or aggregator, not: TFz:dicts within lists are not allowed while in Group mode: %rz'{} not a valid spec type for Group mode)rcallablegetattrrFr.r/r0r idKeyErroritemsgetr rappend ValueErrorformat) rr&rrecursetree _spec_type_spec_idaccdonekeyspecvalspeckeyresults ` ` rr-r-csEEjQ RRQ ,)|+8n,sG G$#G$c(\rSrSrSrSrSrSrSrg)FirstzC holds onto the first value >>> glom([1, 2, 3], Group(First())) 1 rAc,X;a [X 'U$[$r$)r r'rrQs rrF First.aggs  DJM r+c4SURR-$Nz%s()rrr's rr:First.__repr__////r+N rr<r=r>r? __slots__rFr:r@rAr+rr[r[s I 0r+r[c(\rSrSrSrSrSrSrSrg)Avgzy takes the numerical average of all values; raises exception on non-numeric value >>> glom([1, 2, 3], Group(Avg())) 2.0 rAcX nUS==U- ss'US==S- ss'USUS- $![a SS/=o2U'N9f=f)Ngrr)rJ)r'rrQavg_accs rrFAvg.aggs_ ,jG  f  a qzGAJ&&  ,$'8 +G4j ,s+AAc4SURR-$rarbrcs rr: Avg.__repr__rer+NrfrAr+rririsI'0r+ric(\rSrSrSrSrSrSrSrg)Maxzz takes the maximum of all values; raises exception on values that are not comparable >>> glom([1, 2, 3], Group(Max())) 3 rAc,X;dXU:aXU'X $r$rAr^s rrFMax.agg  vT 2Jzr+c4SURR-$rarbrcs rr: Max.__repr__rer+NrfrAr+rrqrqI 0r+rqc(\rSrSrSrSrSrSrSrg)Minzz takes the minimum of all values; raises exception on values that are not comparable >>> glom([1, 2, 3], Group(Min())) 1 rAc,X;dXU:aXU'X $r$rAr^s rrFMin.aggrur+c4SURR-$rarbrcs rr: Min.__repr__rer+NrfrAr+rrzrzrxr+rzc.\rSrSrSrSrSrSrSrSr g)Sampleatakes a random sample of the values >>> glom([1, 2, 3], Group(Sample(2))) # doctest: +SKIP [1, 3] >>> glom(range(5000), Group(Sample(2))) # doctest: +SKIP [272, 2901] The advantage of this over :func:`random.sample` is that this can take an arbitrarily-sized, potentially-very-long streaming input and returns a fixed-size output. Note that this does not stream results out, so your streaming input must have finite length. sizecXlgr$r)r'rs rr(Sample.__init__r*r+cX;aS//X 'X up4[U5UR:aURU5 O*[R"SU5nXPR:aXU'X S==S- ss'U$)Nrr)lenrrMrandomrandint)r'rrQnum_seensampleposs rrF Sample.aggsp  RDJ: v; " MM& !..H-CYY$s  1   r+cRURR<SUR<S3$r7)rrrrcs rr:Sample.__repr__s>>22DII>>r+N) rr<r=r>r?rgr(rFr:r@rAr+rrrs I ?r+rc4\rSrSrSrSr\4SjrSrSr Sr g)Limitia Limits the number of values passed to sub-accumulator >>> glom([1, 2, 3], Group(Limit(2))) [1, 2] To override the default untransformed list output, set the subspec kwarg: >>> glom(range(10), Group(Limit(3, subspec={(lambda x: x % 2): [T]}))) {0: [0, 2], 1: [1]} You can even nest Limits in other ``Group`` specs: >>> glom(range(10), Group(Limit(5, {(lambda x: x % 2): Limit(2)}))) {0: [0, 2], 1: [1, 3]} )nsubspecc<U[La[/nXlX lgr$)rr rr)r'rrs rr(Limit.__init__*s h cG r+cU[[La [S5eU[nX;aS0/X0'X0SU['X0S==S- ss'X0SUR:a[ $U[ "XRU5$)Nz Limit() only valid in Group moderr)rr-r rrr rr)r'rrrQs rr4 Limit.glomit0s ;e #<= =X  RDJ*Q-h 1  :a=466 !KT{6<<77r+cnURR<SUR<SUR<S3$)Nr8z, r)rrrrrcs rr:Limit.__repr__<s!#~~66 MMr+N) rr<r=r>r?rgrr(r4r:r@rAr+rrrs!"!I"* 8Nr+r)r? __future__rrboltons.typeutilsrcorerrrr r r r r rrrrobjectr!r-r[rirqrzrrrAr+rrs +TTT  $  "4*F4*n9Sx0F0&0&000&0&0&0&"?V"?L'NF'Nr+