Python科学计算(第2版)

Python科学计算(第2版)张若愚著北京内容简介本书详细介绍Python科学计算中最常用的扩展库NumPy、SciPy、matplotlib、Pandas、SymPy、TTK、Mayavi、OpenCV、Cython，涉及数值计算、界面制作、三维可视化、图像处理、提高运算效率等多方面的内容。所附光盘中包含所有章节的Notebook以及便携式运行环境WinPython，以方便读者运行书中所有实例。本书封面贴有清华大学出版社防伪标签，无标签者不得销售。版权所有，侵权必究。侵权举报电话：010-6278298913701121933图书在版编目(CIP)数据Python科学计算/张若愚著.—2版.—北京：清华大学出版社，2016ISBN978-7-302-42658-5Ⅰ.①P…Ⅱ.①张…Ⅲ.①软件工具—程序设计Ⅳ.①TP311.56中国版本图书馆CIP数据核字(2016)第014000号责任编辑：李维杰装帧设计：牛静敏责任校对：成凤进责任印制：出版发行：清华大学出版社网址：，地址：北京清华大学学研大厦A座邮编：100084社总机：010-62770175邮购：010-62786544投稿与读者服务：010-62776969，[e-mail]质量反馈：010-62772015，[e-mail]印装者：经销：全国新华书店开本：185mm×260mm彩插：6印张：45.75字数：1083千字版次：2012年1月第1版2016年4月第2版印次：2016年4月第1次印刷印数：1~3500定价：118.00元—————————————————————————————————————————————产品编号：PrefacePythonisrightfullyviewedasageneralpurposelanguage,wellsuitedforwebdevelopment,systemadministration,andgeneralpurposebusinessapplications.It’shasearnedthisreputationwellbypoweringwebsitessuchasYouTube,installationtoolsintegraltoRedHat’soperatingsystem,andlargecorporateITsystemsfromcloudclustermanagementtoinvestmentbanking.Pythonhasalsoestablisheditselffirmlyintheworldofscientificcomputingcoveringawiderangeofapplicationsfromseismicprocessingforoilexplorationtoquantumphysics.Thisbreadthofapplicabilityissignificantbecausetheseseeminglydisparateusesoftenoverlapinimportantways.Applicationsthatcaneasilyconnecttodatabasespublishinformationtotheweb,andefficientlycarryoutcomplexcalculationsarenowcriticalinmanyindustries.Python’sprimarystrengthisthatitallowsdeveloperstobuildsuchtoolsquickly.Python’sscientificcomputingrootsactuallygoquitedeep.GuidovanRossumcreatedthelanguagewhileatCWI,theCenterforMathematicsandComputerScience,intheNetherlands.Asinterestdevelopedoutsidethecenter,othersbegantocontribute.ThefirstseveralPythonworkshops,startingin1994,wereheldanoceanawayatscientificinstitutionssuchasNIST(NationalInstituteofInstrumentsandTechnology),theUSGeologicalSociety,andLLNL(LawrenceLivermoreNationalLaboratories),allsciencecentricinstitutions.Atthetime,Python1.0hadrecentlybeenreleasedandtheattendeeswerejustbeginningtohammeroutthedesignofitsmathematicaltools.Adecadeandahalflater,itisgratifyingtoseehowfarwehavecomebothintheamazingcapabilitiesofthetoolsetandthediversityofthecommunity.Itissomehowfittingthatthefirstcomprehensivebook(thatIknowof)coveringtheprimaryscientificcomputingtoolsforPythoniscomposedandpublished,anotheroceanaway,inChinese.Lookingforwardadecadeandahalf,Icanhardlywaittoseewhatwewillallbuildtogether.Guido,himself,wasnotascientistorengineer.HesatsquarelyinthecomputersciencebranchofCWIandcreatedPythontoeasethepainofbuildingsystemadministrationtoolsfortheAmoebaoperatingsystem.Atthetime,thetoolswerebeingwritteninC.Pythonwastobethetoolthat“bridgedthegapbetweenshellscriptingandC.”OperatingsystemtoolsarenoteveninthesameneighborhoodasmatrixinversionsorfastFouriertransforms,but,asthelanguageemerged,scientistsaroundtheworldweresomeofitsearliestadopters.GuidohadsucceededincreatinganelegantlyexpressivelanguagethatcouplednicelywiththeirexistingCandFortrancode.And,inGuido,theyhadalanguagedesignerwillingtolistenandaddcriticalfeatures,suchascomplexnumbers,specificallyforthescientificcommunity.WiththecreationofNumeric,theprecursortoNumPy,PythongainedafastandpowerfulnumbercrunchingtoolthatsolidifiedPython’sroleasaleadingcomputationallanguageinthecomingdecades.IIPrefacePython科学计算(第2版)Forsome,theterm“scientificprogramming”conjuresupvisionsofintricatealgorithmsdescribedfrom“NumericalRecipesinC”orforgedinlatenightprogrammingsessionsbygraduatestudents.ButtherealityisthedomainencompassesamuchwiderrangeofprogrammingtasksfromlowlevelalgorithmstoGUIdevelopmentwithadvancedgraphics.Thislattertopicistoooftenunderestimatedintermsofimportanceandeffort.Fortunately,RuoyuZhanghasdoneustheserviceofcoveringallfacetsofthescientificprogramminginthisbook.BeginningwiththefoundationalNumpylibrarythealgorithmictoolboxesinSciPyheprovidesthefundamentaltoolsforanyscientificapplication.Hethenaptlycoversthe2Dplottingand3Dvisualizationlibrariesprovidedbymatplotlib,chaco,andmayavi.ApplicationandGUIdevelopmentwithTraitsandTraitsUI,andcouplingtolegacyClibrariesthroughCython,Weave,ctypes,andSWIGarewellcoveredaswell.ThesecoretoolsareroundedoutbycoverageofsymbolicmathematicswithSymPyandvariousotherusefultopics.It’strulygratifyingtoseeallofthesetopicsaggregatedintoasinglevolume.Itprovidesaone-stopshopthatcanleadyoufromthebeginningstepstoapolishedandfullfeaturedapplicationforanalysisandsimulation.EricJones2011/12/8第1版序Python理所当然地被视为一门通用的程序设计语言，非常适合于网站开发、系统管理以及通用的业务应用程序。它为诸如YouTube这样的网站系统、RedHat操作系统中不可或缺的安装工具以及从云管理到投资银行等大型企业的IT系统提供技术支持，从而赢得了如此高的声誉。Python还在科学计算领域建立了牢固的基础，覆盖了从石油勘探的地震数据处理到量子物理等范围广泛的应用场景。Python这种广泛的适用性在于，这些看似不同的应用领域通常在某些重要的方面是重叠的。易于与数据库连接、在网络上发布信息并高效地进行复杂计算的应用程序对于许多行业是至关重要的，而Python最主要的长处就在于它能让开发者迅速地创建这样的工具。实际上，Python与科学计算的关系源远流长。吉多·范罗苏姆创建这门语言，还是在他在荷兰阿姆斯特丹的国家数学和计算机科学研究学会(CWI)的时候。当时只是作为“课余”的开发，但是很快其他人也开始为之做出贡献。从1994年开始的头几次Python研讨会，都是在大洋彼岸的科研机构举行的。例如国家标准技术研究所(NIST)、美国地质