Sample trainng data for task 3 of ICDAR

data-task3-icdar.txt

('d5e1edd79d924caeafbd9f08dbfc71f9',
 ('COMMERCIAL INVOICE\n'
  '\n'
  'INVOICE NO :\n'
  '1111111\n'
  'DATE : 20-Sep-18\n'
  '\n'
  'INVOICE OF :\n'
  'BAGS OF POLYETHYLENE\n'
  'TENO/ 8138\n'
  'SHIPPED ON :\n'
  'YORK\n'
  '\n'
  'FROM :\n'
  'Name Lastname\n'
  'TO : NEW on Account and entire Risk of :\n'
  'LOL :\n'
  '\n'
  'CENTRAL POLY ML LOL CORP\n'
  'ETD : 24-Sep-18\n'
  '2400 Some Place , PO Box xxxx\n'
  'ETA : 17-Oct-18\n'
  '\n'
  'LLLNDX, NJ 07036, USA.\n'
  '45 Days from the date of B/L\n'
  ':\n'
  'DRAWN UNDER / THROUGH ORDER NO :\n'
  '32430, 32507, 32674, 32562,32585\n'
  '24-Sep-18\n'
  '\n'
  'HBL NO :\n'
  '\n'
  'ABCMD229\n'
  '\n'
  'DATE : RATE\n'
  '\n'
  'USD QTY\n'
  '\n'
  'USD RATE\n'
  'COUNTRY OF ORIGIN\n'
  '\n'
  'ITEM NO\n'
  '\n'
  'DESCRIPTION\n'
  '\n'
  'QTY\n'
  '\n'
  'AMT CIF USD\n'
  '\n'
  'USD/ CASE\n'
  '\n'
  "40'\n"
  'X 1 OOLU8652131\n'
  '\n'
  '221660N3M\n'
  'BAGS OF POLYETHYLENE\n'
  '806\n'
  '\n'
  '$\n'
  '\n'
  '11.820$\n'
  '\n'
  '9,526.92\n'
  '\n'
  ': SIZE 515 X 320 X 78 MM\n'
  '\n'
  ': PO NO 32430\n'
  '\n'
  '111112434M\n'
  'BAGS OF POLYETHYLENE\n'
  '205\n'
  '\n'
  '$\n'
  '\n'
  '16.745 $\n'
  '\n'
  '3,432.73\n'
  '\n'
  '505 X 265 X 135 MM\n'
  '\n'
  ': SIZE : PO NO 32507\n'
  '\n'
  '231748N3MM\n'
  '\n'
  'BAGS OF POLYETHYLENE\n'
  '1000\n'
  '\n'
  '$\n'
  '\n'
  '11.408\n'
  '\n'
  '$\n'
  '\n'
  '11,408.00\n'
  '\n'
  '400 x 270 x 100 MM\n'
  '\n'
  'SIZE : : PO NO 32674\n'
  '\n'
  '3340KXXH/250\n'
  'BAGS OF POLYETHYLENE\n'
  '200\n'
  '\n'
  '$\n'
  '\n'
  '8.893\n'
  '\n'
  '$\n'
  '\n'
  '1,778.60\n'
  '\n'
  'INDIA\n'
  '\n'
  '380 X 230 X 150 MM\n'
  '\n'
  ': SIZE : PONO 32562\n'
  '\n'
  '232048N3MM\n'
  'BAGS OF POLYETHYLENE\n'
  '$\n'
  '\n'
  '14.408\n'
  '\n'
  '$\n'
  '\n'
  '100.86\n'
  '\n'
  ': 420 x 290 x 110 MM\n'
  '\n'
  'SIZE 32430\n'
  '\n'
  ': PO NO 232048N3MR\n'
  'BAGS OF POLYETHYLENE\n'
  '350\n'
  '\n'
  '$\n'
  '\n'
  '11.144\n'
  '\n'
  '$\n'
  '\n'
  '3,900.40\n'
  '\n'
  '480 X 285 X 105 MM\n'
  '\n'
  ': SIZE : PONO 32585\n'
  '\n'
  'TOTAL\n'
  '\n'
  '2568\n'
  '\n'
  '30,147.50\n'
  '\n'
  'TOTAL USD CIF\n'
  '\n'
  '$\n'
  '\n'
  '30,147.50\n'
  '\n'
  'CIF USD THIRTY THOUSAND ONE HUNDRED FORTY SEVEN DOLLARS & CENTS FIFTY ONLY\n'
  '\n'
  ':\n'
  '\n'
  'AMOUNT IN WORDS Name of Beneficiary :\n'
  '\n'
  'Drip Capital Inc\n'
  'For LUCRO\n'
  '\n'
  'Bank Name :\n'
  'Silicon Valley Bank\n'
  '\n'
  'This Invoice has been sold to LOLXXXX\n'
  'For M/s. LLLLLXXX\n'
  '\n'
  'Capital Inc; 555 Bryant Street,#356\n'
  '\n'
  '3003 Tasman Drive, Santa Clara, CA 95054, USA.\n'
  ':\n'
  'Palo Alto, CA 94301, USA\n'
  '\n'
  "Bank Address Bank's Routing Number :\n"
  '\n'
  '(ABA #) 112XXX982X\n'
  'Swift Code :\n'
  'SVBKUS6S\n'
  'Authouse 3302198017\n'
  '\n'
  ':\n'
  'Account Number PLEASE PAY "DRIP CAPITAL"\n'
  '\n'
  'nator\n'
  '\n'
  'ON OUR ACCOUNT\n'
  '\n'
  'Authorised Signatory\n'
  '\n'
  'E. & O.E.\n'
  '\n'
  ': NUMBER 2459BCKJDFLKJA33\n'
  '\n'
  '0312007949\n'
  '\n'
  ': IEC CODE GST Goregaon Road '
  'Aashir HEAD OFFICE (W),\n'
  '\n'
  ': INDIA. 4XXXX, Mumbai - Tel 2XXXXX / +91-222222222 : Website : '
  'www.example.com\n'
  '\n'
  '- Gujarat\n'
  '\n'
  'Branch Gujarat :',
  array([10, 10, 10, ..., 10, 10, 10])))

data.py

def data_to_text2(df):
    """Convert google's dataframe to plain text conserving blocks and lines."""
    text = ""
    labels = []
    for row in df.itertuples():
        text += row.Text
        labels.extend([row.label_encoded for i in range(len(row.Text))])
        if row.space_type == 0:
            continue
        elif row.space_type in [1, 2]:
            text += " "
            labels.append(row.label_encoded)
        elif row.space_type in [3, 4]:
            text += "\n"
            labels.append(row.label_encoded)
        else:
            text += "\n\n"
            labels.append(row.label_encoded)
            labels.append(row.label_encoded)
    text = text.strip()
    return text, np.array(labels[:len(text)]).flatten()