123456789101112131415161718192021222324 |
- '''Created Dec 28, 2021 Levi'''
- import camelot
- import pandas as pd
- tables = camelot.read_pdf("C:/Users/Levi/Downloads/ikea/ROINV22000000202597.pdf", flavor='stream')
- df = tables[0].df
- # concat = lambda s1, s2: '{} {}'.format(s1, s2)
- # r3 = df.iloc[3].combine(df.iloc[4], func=concat)
- # df.iloc[3] = r3
- df = df[5:]
- # print(df)
- columns = ['Cod', 'Denumire', 'Cantitate', 'Pret']
- ndf = pd.DataFrame(df[[0, 1, 2, 3]], columns=[0, 1, 2, 3])
- ndf.columns = columns
- ndf['Denumire'] = ndf['Denumire'].str.upper()
- print(ndf)
- # doc = fitz.Document("C:/Users/Levi/Downloads/ikea/ROINV22000000202597.pdf")
- # page = doc.load_page(1)
- # pix = page.get_pixmap(matrix=mat) # render page to an image
- # print(page)
- # print(pix)
- # text = page.get_text('text')
- # print(pix.save('out.png'))
|