使用python从格式化的PDF中提取文本

Question

我必须解析格式化的 pdf 以获得一些字段。 PDF 为here。我需要解析的内容显示在thisimgur 中。我使用 PyPDF2 获取文本，但它返回没有任何格式的原始文本。

import PyPDF2
pdfFileObj = open('GPO-PLUMBOOK-2000-4-1.pdf','rb')
pdfReader = PyPDF2.PdfFileReader(pdfFileObj)
pageObj = pdfReader.getPage(0)
print(pageObj.extractText())

我得到的输出如下：

LEGISLATIVE BRANCHLocationPosition TitleName of IncumbentPayPlanType ofAppt.Level,Grade, orPayTenureExpiresARCHITECT OF THE CAPITOLAlan M. HantmanWashington, DCArchitect of the Capitol10 years02/02/07IIIEXPASLocationPosition TitleName of IncumbentPayPlanType ofAppt.Level,Grade, orPayTenureExpiresGENERAL ACCOUNTING OFFICEDavid M. WalkerWashington, DCComptroller General of the United States11/09/1315 years$141,300OTPASVacant  Do...........Deputy Comptroller General of the United States..................OTXSLocationPosition TitleName of IncumbentPayPlanType ofAppt.Level,Grade, orPayTenureExpiresGOVERNMENT PRINTING OFFICEMichael F. DiMarioWashington, DCPublic Printer............IIIEXPASRobert T. Mansker  Do...........Deputy Public Printer............IVEXXSFrancis J. Buckley, Jr.  Do...........Superintendent of Documents..................SLXSRobert G. Andary  Do...........Inspector General..................SLXSMary Beth Lawler  Do...........Staff Assistant............14OTSCLocationPosition TitleName of IncumbentPayPlanType ofAppt.Level,Grade, orPayTenureExpiresLIBRARY OF CONGRESSLIBRARIAN OF CONGRESSJames H. BillingtonWashington, DCLibrarian of Congress............IIIEXPASLIBRARY OF CONGRESS TRUST FUND BOARDJames H. Billington  Do...........Chairman (Ex-Officio)..................WCPASTed Stevens  Do...........Chairman of the Joint Committee of the Library (Ex-Officio)..................WCXSLawrence Summers  Do...........Member (Ex-Officio), Secretary of the Treasury..................WCPASDonald Hammond  Do...........Member (Designee for the Secretary of the Treasurer)..................WCXSCeil Pulitzer  Do...........Member5 years03/23/03......WCPASNajeeb Halaby  Do...........Member5 years08/31/05......WCPASJohn Kluge  Do...........Member5 years03/10/03......WCXSWayne Berman  Do...........Member5 years12/22/01......WCXSEdwin Cox  Do...........Member5 years03/31/04......WCXSJohn Henry  Do...........Member5 years12/22/03......WCXSDonald Jones  Do...........Member5 years10/08/02......WCXSJulie Finley  Do...........Member5 years06/29/01......WCXSBernard Rappaport  Do...........Member5 years12/22/01......WCXS(1)

我需要分离数据，例如Location列下的数据等等。

Answer 1

看看tabula 库（这里是github）。这会返回一个 pandas 数据框。

df = tabula.read_pdf("/home/michael/Downloads/GPO-PLUMBOOK-2000-4-1.pdf", pages=1)
df.dropna(inplace=True)
print(df[:2])

如果您需要阅读其他表格，或者想节省时间，您还可以调整应该使用 pdf 的哪个部分。这样，您只需阅读所有 pdf 表格，然后将我的数据切片为您想要的输出。