œ_#ÁÕ§TE NAŒ“KeÉ:”(åŽÖJÞùY’‚ñùž7; «]Û ý`8g“¯B© jdÖÖ¸ðzœ¸¦4Ç3Kó^(ÍÖ¼ Õ€pvìwšõB4df$Èü^0˜…åÌC$#2FŽÑ§±¦ÛZ/÷š&m£ñzÒÖ ’.Î]!Î;ƒ(Õ–¢d/—#Kª+tZyuÏB>NÛÖ†(¸ŒSà'³„Y˜´-_•¦¼´˜OlNK§¶ÒàŠˆTHµƒeTPå·fïM’…þuÏÍüp6دªE£åü‡ZØ'CKF#â«;‹eyO Qp„†l"ö1èíÙP ÏŒúl! BÝ2ñª•_VÁÉ÷3eu`–F¸ìI--ö<¿žë¯4õ캿¢)34Å{wMÉ2ÆÖFŸ¥`e9Ú¶¸P‡.”FÔï rY ‚²ÈTB,{ÛœéJ}«àQ4¹0Rû4D‚B§S‘ dO•v¾„™Sן¯3FeŸ™«+ÓâwH dÕÛÌì·P4ë&¥#rÜÉ Ù¦ê†ý·xòqk¯2,¹§™E\ék‚×Sá”ÚºÙ⺷ö£6…à ʾ qSá³Å|;àû}4Ÿ($â¹VY~óÍ!èÜÒŒËX½Ù1j‚VíÍŸš³+œ]«½g{_{/vµ½\¢¶vÉWKÿ:ñám½ ¥ S²x‘t ŽšÝÙÿÀÇ^ný PK IW™k‚½÷ á _rels/.relsUT dìd dìd dìd’ÏNÃ0‡ï{ŠÈ÷ÕÝ@¡¥» ¤Ý*`%îÑ&QâÁöö‚J£ì°cœŸ¿|¶²ÙÆA½rL½wVE Šñ¶w†çúay * 9Kƒw¬áÈ ¶ÕbóÄIîI]’Ê—4t"á1™ŽGJ…ìòMããH’±Å@æ…ZÆuYÞ`üÍ€jÂT;«!îì T}|Û7MoøÞ›ýÈNN<|v–í2ÄÜ¥ÏèšbË¢Ázó˜Ë )„"£OÏ7ú{ZYÈ’yÞç#1'tuÉM?6o>Z´_å9›ëKÚ˜}?þ³žÏÌ·N>fµx PK IWª½e ¢ U € word/document.xmlUT dìdPK IWþË3” z €J¢ word/settings.xmlUT dìdPK IWC‡{š' ƒ €¤ docProps/custom.xmlUT dìdPK IW츱=Œ €‡¥ [Content_Types].xmlUT dìdPK IWV%ë±" €U§ docProps/app.xmlUT dìdPK IW€RŒ 3 €¶¨ docProps/core.xmlUT dìdPK IWkòDn ô €ª word/_rels/document.xml.relsUT dìdPK IW;$î €Î« word/fontTable.xmlUT dìdPK IW+åäz] ÷. €ý¬ word/numbering.xmlUT dìdPK IW¤2×r- ¿ €›° word/styles.xmlUT dìdPK IWMFÒ ø €´ word/header1.xmlUT dìdPK IWF— T e €· word/media/image1.jpegUT dìdPK IW!Yéáå €°Ë word/media/image2.pngUT dìdPK IW°Àºë ú €ÙÌ word/media/image3.pngUT dìdPK IW$“†ª L €Î word/footer1.xmlUT dìdPK IWzaGôM €ñÑ word/footer2.xmlUT dìdPK IW–µâº P €}Õ word/theme/theme1.xmlUT dìdPK IW™k‚½÷ á €{Û _rels/.relsUT PK ! bîh^ [Content_Types].xml ¢( ¬”ËNÃ0E÷HüCä-Jܲ@5í‚Ç*Q>Àēƪc[žiiÿž‰ûB¡j7±ÏÜ{2ñÍh²nm¶‚ˆÆ»R‹ÈÀU^7/ÅÇì%¿’rZYï @1__f› ˜q·ÃR4DáAJ¬h>€ãÚÇV߯¹ªZ¨9ÈÛÁàNVÞ8Ê©ÓãÑÔji){^óã-I‹"{Üv^¥P!XS)bR¹rú—K¾s(¸3Õ`cÞ0†½ÝÎß»¾7M4²©ŠôªZÆk+¿|\|z¿(Ž‹ôPúº6h_-[ž@!‚ÒØ Pk‹´2nÏ}Ä?£LËð Ýû%áÄßdºždN"m,à¥ÇžDO97*‚~§Èɸ8ÀOíc|n¦Ñ äEøÿöéºóÀBÉÀ!$}‡íàÈé;{ìÐå[ƒîñ–é2þ ÿÿ PK ! µU0#ô L _rels/.rels ¢( ¬’MOÃ0†ïHü‡È÷ÕÝBKwAH»!T~€Iܵ£$Ý¿'TƒG½~üÊÛÝ<êÈ!öâ4¬‹;#¶w†—úqu *&r–Fq¬áÄvÕõÕö™GJy(v½*«¸¨¡KÉß#FÓñD±Ï.W ¥†=™ZÆMYÞbø®ÕBS톰·7 ê“Ï›×–¦é ?ˆ9LìÒ™ÈsbgÙ®|Èl!õùUSh9i°bžr:"y_dlÀóD›¿ý|-NœÈR"4ø2ÏGÇ% õZ´4ñËyÄ7 ëÈðÉ‚‹¨Þ ÿÿ PK ! Q48wÛ — xl/workbook.xml¤UÙnâ0}iþ!cñ‡ *–¢AšVU×$dC¬&vÆv UÕŸë@XÊK§/¹p|Žï¹N÷b“¥Ö •Š ÞC¸î"‹òHÄŒ¯zèá~b·‘¥4á1I§=ôJºèÿüÑ] ù¼âÙ ®z(Ñ:GE ͈ª‹œrˆ,…̈†©\9*—”Ä*¡Tg©ã¹nàd„q´Eåg0ÄrÉ":Q‘Q®· ’¦D}•°\UhYô¸ŒÈç"·#‘å ±`)Ó¯%(²²(œ®¸d‘‚ì nZ w v¡ñª• t¶TÆ")”Xê:@;[Ògú±ë`|²›ó=ø’ïHúÂL÷¬dðEVÁ+8€a÷Ûh¬Uz%„Íû"ZsÏÍCýî’¥ôqk]‹äù5ÉL¦Rd¥Dé˘i÷P ¦bM/|dÉ",…¨çãFNoçiûéë>aêiçsó#ðÄ ÕTr¢éHp ÜIú®ÝJìQ"ÀÜÖ-ý[0I¡¦ÀZ Z…d¡nˆN¬B¦=4 g %PDF-1.4 %âãÏÓ 3 0 obj << /Linearized 1 /L 422775 ÿØÿà JFIF ÿÛ C ÿÛ C ÿÀ X" ÿÄ ÿÄ H !1A"Qaq2‘¡#±ÁBRÑ3Cbrá$S‚¢²ð4ñ%6DTc’ÂsÿÄ ÿÄ = !1AQ"aq‘Á2R¡±BÑð#3br’²4á$‚¢ÂñÿÚ ? áHBßÝ`„! !@B„ „! !@B„ „! !@B„ „! !@B„ „! !@B„ „! !@B„ „! !@B„ „! !@B„ „! !@B„ „! !@B„ „! !@B„ „! !@B„ „! !@B„ „! !@B„ „! !@B„ „! !@B„ „! !@B„ „! !@B„ „! !@B„ „! !@B„ „! !@B„ „! !@B„ „! ! stream
"""Convenient parallelization of higher order functions.
This module provides two helper functions, with appropriate fallbacks on
Python 2 and on systems lacking support for synchronization mechanisms:
- map_multiprocess
- map_multithread
These helpers work like Python 3's map, with two differences:
- They don't guarantee the order of processing of
the elements of the iterable.
- The underlying process/thread pools chop the iterable into
a number of chunks, so that for very long iterables using
a large value for chunksize can make the job complete much faster
than using the default value of 1.
"""
__all__ = ["map_multiprocess", "map_multithread"]
from contextlib import contextmanager
from multiprocessing import Pool as ProcessPool
from multiprocessing import pool
from multiprocessing.dummy import Pool as ThreadPool
from typing import Callable, Iterable, Iterator, TypeVar, Union
from pip._vendor.requests.adapters import DEFAULT_POOLSIZE
Pool = Union[pool.Pool, pool.ThreadPool]
S = TypeVar("S")
T = TypeVar("T")
# On platforms without sem_open, multiprocessing[.dummy] Pool
# cannot be created.
try:
import multiprocessing.synchronize # noqa
except ImportError:
LACK_SEM_OPEN = True
else:
LACK_SEM_OPEN = False
# Incredibly large timeout to work around bpo-8296 on Python 2.
TIMEOUT = 2000000
@contextmanager
def closing(pool: Pool) -> Iterator[Pool]:
"""Return a context manager making sure the pool closes properly."""
try:
yield pool
finally:
# For Pool.imap*, close and join are needed
# for the returned iterator to begin yielding.
pool.close()
pool.join()
pool.terminate()
def _map_fallback(
func: Callable[[S], T], iterable: Iterable[S], chunksize: int = 1
) -> Iterator[T]:
"""Make an iterator applying func to each element in iterable.
This function is the sequential fallback either on Python 2
where Pool.imap* doesn't react to KeyboardInterrupt
or when sem_open is unavailable.
"""
return map(func, iterable)
def _map_multiprocess(
func: Callable[[S], T], iterable: Iterable[S], chunksize: int = 1
) -> Iterator[T]:
"""Chop iterable into chunks and submit them to a process pool.
For very long iterables using a large value for chunksize can make
the job complete much faster than using the default value of 1.
Return an unordered iterator of the results.
"""
with closing(ProcessPool()) as pool:
return pool.imap_unordered(func, iterable, chunksize)
def _map_multithread(
func: Callable[[S], T], iterable: Iterable[S], chunksize: int = 1
) -> Iterator[T]:
"""Chop iterable into chunks and submit them to a thread pool.
For very long iterables using a large value for chunksize can make
the job complete much faster than using the default value of 1.
Return an unordered iterator of the results.
"""
with closing(ThreadPool(DEFAULT_POOLSIZE)) as pool:
return pool.imap_unordered(func, iterable, chunksize)
if LACK_SEM_OPEN:
map_multiprocess = map_multithread = _map_fallback
else:
map_multiprocess = _map_multiprocess
map_multithread = _map_multithread