>ieUddlmZddlZddlmZddlZddlmZddl m Z ddl m Z ddlmZddlmZdd lmZdd lmZdd lmZmZmZmZmZmZmZmZdd lm Z dd l!m"Z"ddl#m$Z$m%Z%m&Z&m'Z'ddl(m)Z)ddl*m+Z+ddl,m-Z-ddl.m/Z/m0Z0m1Z1m2Z2m3Z3e4de5fiZ6dZ7dZ8e$e-dddddZ9dZ:dZ;dZdS)!) annotationsN)Integral)is_datetime64_any_dtype)Rolling) normalize_arg)tokenize)BlockwiseDepDict)methods)check_axis_keyword_deprecation)Scalar_Frame_get_divisions_map_partitions_get_meta_map_partitions_maybe_from_pandasapply_and_enforce new_dd_objectpartitionwise_graph) from_pandas)_maybe_align_partitions)insert_meta_param_descriptionis_dask_collectionis_dataframe_likeis_series_like)unpack_collections)HighLevelGraph) no_default)Mapply derived_fromfuncname has_keywordCombinedOutputcd}|5t|tr |jd|krt||5t|tr |jd|krt|d|||fD}t j|}t ||t|nd|t|ndfS)NzqPartition size is less than overlapping window size. Try using ``df.repartition`` to increase the partition size.rcg|]}||SN).0ps 6lib/python3.11/site-packages/dask/dataframe/rolling.py z#_combined_parts..8s N N N1 Q ) isinstancershapeNotImplementedErrorr concatr"len) prev_part current_part next_partbeforeaftermsgpartscombineds r)_combined_partsr9)s * FH!=!= ?1  ' '%c** *E8!.Es' ? ? ?" 2~ > > ?2 ? ? ?r+rcLg|]!}t|tr|dn|"Srr<r'args r)r*z!overlap_chunk..Hs/ O O O3jn55 >CFF3 O O Or+)r,datetime timedeltar-iloc) funcr4r5argskwargsdfsr8prev_part_lengthnext_part_lengthout expansions r) overlap_chunkrNDs ? ? ? ? ?C36q60H 0 O O$ O O OD $   C&(,--"!I~aAIaLHN1$55  ))x  %+,,!    8FE6M ""r+T)metaenforce_metadatatransform_divisionsalign_dataframesc t|st|rt|st|dn|}|f|z}t t rt jt t rt jt tj st tj r.t|j j j stdnEt tr!dkrt trdkst!d| dd} | dd} t%|sJ| t'g|Ri| } n%dt)|z} t'|g|Ri| } | d | } |rEt+|} t-|}n%#t $r} t!| d | d} ~ wwxYwd |D}t/|||| | t1d |DrH| dft2|t4d |Df| fi}t7j| ||}t;|| Sg}g}t=|||||| fd}|D]}t |t>r6||}| || |MtC|}tE|\}}|r+| ||#|| |i}d}| $D]D\}}tC|}tE|\}}|#||||<|rd}EtK|drKdtMddD}|'dtQ||fd}|r#tStT| |g|R|tVd|}n#|r| n|}tStV| |g|Ri|d|i}t7j| ||}tY|| S)a] Apply a function to each partition, sharing rows with adjacent partitions. Parameters ---------- func : function The function applied to each partition. If this function accepts the special ``partition_info`` keyword argument, it will recieve information on the partition's relative location within the dataframe. df: dd.DataFrame, dd.Series args, kwargs : Positional and keyword arguments to pass to the function. Positional arguments are computed on a per-partition basis, while keyword arguments are shared across all partitions. The partition itself will be the first positional argument, with all other arguments passed *after*. Arguments can be ``Scalar``, ``Delayed``, or regular Python objects. DataFrame-like args (both dask and pandas) will be repartitioned to align (if necessary) before applying the function; see ``align_dataframes`` to control this behavior. enforce_metadata : bool, default True Whether to enforce at runtime that the structure of the DataFrame produced by ``func`` actually matches the structure of ``meta``. This will rename and reorder columns for each partition, and will raise an error if this doesn't work, but it won't raise if dtypes don't match. before : int, timedelta or string timedelta The rows to prepend to partition ``i`` from the end of partition ``i - 1``. after : int, timedelta or string timedelta The rows to append to partition ``i`` from the beginning of partition ``i + 1``. transform_divisions : bool, default True Whether to apply the function onto the divisions and apply those transformed divisions to the output. align_dataframes : bool, default True Whether to repartition DataFrame- or Series-like args (both dask and pandas) so their divisions align before applying the function. This requires all inputs to have known divisions. Single-partition inputs will be split into multiple partitions. If False, all inputs must have either the same number of partitions or a single partition. Single-partition inputs will be broadcast to every partition of multi-partition inputs. $META See Also -------- dd.DataFrame.map_overlap zMMust have a `DatetimeIndex` when using string offset for `before` and `after`rz*before and after must be positive integerstokenN parent_metazoverlap--zx. If you don't want the partitions to be aligned, and are calling `map_overlap` directly, pass `align_dataframes=False`.c<g|]}t|t|Sr&)r,r r=s r)r*zmap_overlap..s' 7 7 7" 2v 6 6 72 7 7 7r+c3@K|]}t|tVdSr%)r,r rAs r) zmap_overlap..s, 3 3s:c6 " " 3 3 3 3 3 3r+c g|] }|jdf Sr@)_namerAs r)r*zmap_overlap..s888C#)Q888r+ dependenciesci}t|\}}||t| \}}||dt|z}t t |||D]\}\}} } ||f} t|| |  f|| <tj |||g} t| |S)Nzoverlap-concat-r]) _get_previous_partitionsupdate_get_nexts_partitionsr enumeratezip __dask_keys__r9rfrom_collectionsr)rBdskprevs_parts_dskprevsnexts_parts_dsknextsname_aiprevcurrentnextkeygraphr5r4 divisionsrOs r)_handle_frame_argumentz+map_overlap.._handle_frame_arguments!9#v!F!F ?###!6sE!B!B ?###"Xc]]2(1 s((**E 2 2) )  M M $A$gt1+C'wfeLCHH/3%PPPUFD)<<zmap_overlap..s8   8DQH55   r+c|i|d|iS)Nrur&)rurGrH orig_funcs r)rFzmap_overlap..funcs#9dLfLL^LLL Lr+)r^_func_metar^)-rrrrr,strpd to_timedeltarCrDrindex_meta_nonempty inferred_type TypeErrorr ValueErrorpopcallablerr rrrallrtuplerrfr rr appendrrextenditemsr!rcinsertr rrrNr) rFr>r4r5rOrPrQrRrGrHnamerVrUerIlayerrrargs2r^rtrBarg2 collectionskwargs3simplekvrurgkwargs4rsr|s ``` @@r) map_overlapr_sD 2   "3B"7"7 ASTVAWAW  B  54'LMM +   vx ( ( K! 5(++ IJJ J ::gt $ $D**]D11K D>> vu>t>>>v>>HTNN*tVUDTDDDVDD  U  D!$'' *400DD   QQQ   8 7 7 7 7C #D#tVT; O OD 3 3d 3 3 333 ) 1I8848889  /e$OOOeT4((( EL--sD$I========&   c6 " " ((--C LL       $ $ $ C  .s33k   LL       , , , , LL    G F 1 !  +A..;K(((  F4)** M  (3B388     Q(88999  M M M M M !          &      #/&&!              &      +D#L Q Q QE dI 6 66sG$$ H.HHci}|j}d}dt||z}|rrt|tr]g}t d|jD]/}||f}t j||f|f||<||0|dnt|tj rtj |j jdd} || krt#|g}t d|jD]0}||f}t$||dz f||f|f||<||1|dn dg|jz}||fS)zE Helper to get the nexts partitions required for the overlap ziPartition size is less than specified window. Try using ``df.repartition`` to increase the partition sizezoverlap-append-rTNrzr)r\rr,rrange npartitionsrheadrrCrDrSeriesrsdiffrEanyr_head_timedelta) r>r5rgdf_nametimedelta_partition_messagename_brkrmrqdeltass r)rbrb<s ChG F "e!4!4 4F (E8,,(q".))  A1+C! e4CH LL     T E8- . . (2<((--//4QrT: FN   ! ! :899 9q".))  A1+C''1q5)9GQ<OCH LL     T' :r+ci}|jdt||z}|rat|trLdg}t |jdz D]/}||f}t j|f|f||<||0nt|tj r^tj |j }|jdd}||kr|d} dg}t |jdz D]}||dz} t#| |z | } |||} } | | kr"| dkr| || z } | dz } | | kr| dk||f}t$fdt | |dzD|dzf|f||<||nXdg}t |jdz D]1}||f}t$|fg|dzf|f||<||2n dg|jz}||fS)zH Helper to get the previous partitions required for the overlap zoverlap-prepend-NrTrzrcg|]}|fSr&r&)r'rrs r)r*z,_get_previous_partitions..s;;;agq\;;;r+)r\rr,rrrrtailrrCrDrrrsrrErmax_tail_timedelta)r>r4rgrlrirmrqdivsrpt_zpt_ilbfirstjrs @r)r`r`as ChG (2v"6"6 6F 5(*VX..5(r~)**  A1+C! f5CH LL      FH. / /.(y&&!!B$' VO " "# "7DFE2>A-.. " "AE{--7AqbjjQUU!F1I-EAAbjjQUUqk#;;;;5AE??;;;a!e$ C  S!!!!) ".FE2>A-.. " "qk#q\Na!e$ C  S!!!! "' :r+cX||j|j|zkS)zReturn rows of ``next_`` whose index is before the last observation in ``current`` + ``after``. Parameters ---------- current : DataFrame next_ : DataFrame after : timedelta Returns ------- overlapped : DataFrame )rr)ronext_r5s r)rrs)  1 1 3 3e ;< ==r+cLtjfd|D}|S)a4Return the concatenated rows of each dataframe in ``prevs`` whose index is after the first observation in ``current`` - ``before``. Parameters ---------- current : DataFrame prevs : list of DataFrame objects before : timedelta Returns ------- overlapped : DataFrame cfg|]-}||jjz k.Sr&)rmin)r'rnr4ros r)r*z#_tail_timedelta..s8MMMtdjGM--//&89 :MMMr+)r r/)riror4selecteds `` r)rrs8~MMMMMuMMMH Or+ceZdZdZddddefdZdZedZe dZ dZ e e d Ze e d Ze e d Ze e d Ze e d Ze e dZe e dZe e ddZe e ddZe e dZe e dZe e dZe e  ddZe e dZeZdZdS)rz%Provides rolling window calculations.NFc||_||_||_||_||_||_|jjdi|t|jtrdnd|_ dS)Nfreqr&) objwindow min_periodscenteraxiswin_typer~rolling_rolling_kwargsr,int _win_type)selfrrrrrrs r)__init__zRolling.__init__sx &     33D0022333 ",DK!=!=I6r+cl|j|j|j|jd}|jt ur |j|d<|S)N)rrrrr)rrrrrr)rrHs r)rzRolling._rolling_kwargssBk+k    9J & &!YF6N r+c~|jdvp4t|jto |jdkp|jjdkS)zm Indicator for whether the object has a single partition (True) or multiple (False). )rTcolumnsrT)rr,rrrrrs r)_has_single_partitionzRolling._has_single_partitionsE I ' )4;11FdkQ6F )x#q( r+ct5|jdi|}dddn #1swxYwYt|||i|S)Nr&)r rgetattr)r>rolling_kwargsrrGrHrs r)pandas_rolling_methodzRolling.pandas_rolling_methods + - - 3 3 bj22>22G 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3%ww%%t6v666s )--c|}|j|jj||g|Ri|}|jr |jj|j||g|R||d|S|jr|jdz}|j|z dz }n3|jdkrtj |j}d}n |jdz }d}t|j|j||||g|R||d|S)N)rUrOrTrr) rrrrrmap_partitionsrrrr Timedeltar)r method_namerGrHrrOr4r5s r) _call_methodzRolling._call_methods^--//)t) H #^[ CG   KQ    % *48**  "   ; [A%FK&(1,EE ^v % %\$+..FEE[1_FE  & H               r+c,|dS)Ncountrrs r)rz Rolling.count's  )))r+c,|dS)Ncovrrs r)rz Rolling.cov+  '''r+c,|dS)Nsumrrs r)rz Rolling.sum/rr+c,|dS)Nmeanrrs r)rz Rolling.mean3  (((r+c,|dS)Nmedianrrs r)rzRolling.median7s  ***r+c,|dS)Nrrrs r)rz Rolling.min;rr+c,|dS)Nrrrs r)rz Rolling.max?rr+rTc0|ddS)NstdrTddofrrrs r)rz Rolling.stdC  Q ///r+c0|ddS)NvarrTrrrs r)rz Rolling.varGrr+c,|dS)Nskewrrs r)rz Rolling.skewKrr+c,|dS)Nkurtrrs r)rz Rolling.kurtOrr+c.|d|S)Nquantiler)rrs r)rzRolling.quantileSs  X666r+cythonc J|pi}|pd}|d||||||S)Nr&r)rawengine engine_kwargsrGrHr)rrFrrrrGrHs r)rz Rolling.applyWsH2zr   '!   r+c&|jd|g|Ri|S)Naggr)rrFrGrHs r) aggregatezRolling.aggregatems' t >t>>>v>>>r+c d}|}|j|d<|j|d<dddt ||DS)Nc,|\}}dddddd}||S)NrrTrrrrrrr&)itemrr_orders r)orderzRolling.__repr__..orderts2DAq  F!9 r+rrz Rolling [{}],c3.K|]\}}||d|VdS)N=r&)r'rrs r)rZz#Rolling.__repr__..s>Aq= q  ===r+)rq)rrrformatjoinsortedr)rrrs r)__repr__zRolling.__repr__ss   --//#';x %)^z"$$ HH">#7#7#9#9uEEE     r+)rT)FrNNN)__name__ __module__ __qualname____doc__rrrpropertyr staticmethodrrr pd_Rollingrrrrrrrrrrrrrrrr r&r+r)rrs//  JJJJ0      X  77\7 & & & P\****\*(((\*(((\*)))\*+++\*(((\*(((\*0000\*0000\*)))\*)))\*777\*      *\*??? C     r+rc\eZdZ d fd ZfdZeddddZfdZxZS) RollingGroupbyNFrc|j|_|j|_|j}|jt |jt r |jg}nt |j}t |jt r||jn| |j||}t ||||||dS)Nr) _groupby_kwargs_slice_groupby_slicerr,rlistbyrrsuperr) rgroupbyrrrrrr sliced_plus __class__s r)rzRollingGroupby.__init__s '6%nk   *$-s33 8#23 "4#677 '*c** /""7:....""7:...k"C  #      r+ct}|dddvr|d|S)Nr)rr)rrgetr)rrHrs r)rzRollingGroupby._rolling_kwargssH((** ::fd # #| 3 3 JJv    r+groupby_kwargs groupby_slicec|jdi|}|r||}|jdi|}t|||i|dS)Nrz)levelr&)rrr sort_index) r>rrr r!rGrHrrs r)rz$RollingGroupby.pandas_rolling_methodst"*..~..  -m,G!'/33N33%ww%%t6v66AAAKKKr+cXtj|g|R|j|jd|S)Nr)rrrr)rrrGrHrs r)rzRollingGroupby._call_methodsP#uww#     /-       r+)NNFNr) r r r rrrrr __classcell__)rs@r)rrs       B  L L L L\ L         r+r)? __future__rrCnumbersrpandasrpandas.api.typesrpandas.core.windowrrdask.array.corer dask.baserdask.blockwiser dask.dataframer dask.dataframe._compatr dask.dataframe.corer r rrrrrrdask.dataframe.iordask.dataframe.multirdask.dataframe.utilsrrrr dask.delayedrdask.highlevelgraphr dask.typingr dask.utilsrrrr r!typerr"r9rNrrbr`rrrr&r+r)r:s""""""444444444444))))))++++++""""""AAAAAA                    *)))))888888 ,+++++......""""""DDDDDDDDDDDDDD&"556###6 Y7Y7Y7Y7Y7x"""J>>>B>>>"(A A A A A A A A H@ @ @ @ @ W@ @ @ @ @ r+