source: main/trunk/model-sites-dev/heritage-nz/collect/pdf-reports/prepare/xls-to-csv.py@ 33191

import xlrd
import csv
import os

import xlsutil

letter_to_folder = {
  'a': "A - B",
  'b': "A - B",
  'c': "C - D",
  'd': "C - D",
  'e': "E - F",
  'f': "E - F",
  'g': "G - H",
  'h': "G - H",
  'i': "I - J",
  'j': "I - J",
  'k': "K - L",
  'l': "K - L",
  'm': "M - N",
  'n': "M - N",
  'o': "O - P",
  'p': "O - P",
  'q': "Q - R",
  'r': "Q - R",
  's': "S - T",
  't': "S - T",
  'u': "U - V",
  'v': "U - V",
  'w': "W - Z",
  'x': "W - Z",
  'y': "W - Z",
  'z': "W - Z"
}

def id_to_relative_pdf_file(id):
    first_char = id[0:1]
    first_char_lower = first_char.lower()

    file_tail = letter_to_folder[first_char_lower]
    file_root = "Reports " + file_tail

    pdf_file = os.path.join("Digital-Library-PDFs",file_root,id+".pdf")

    return pdf_file


# **** Written, but never used
# **** Col numbers refer to older version of spreadsheet
#
# There is now a mapping of header-name to index position which would be
# a better way to do thing

def fixup_cell(col_num,cell_str):
    # col == 1 => Author
    # col == 2 => Title
    # col == 4 => TLA
    # col == 5 => AuthorityNo
    # col == 6 => SiteNo

    if col_num == 1: #Author
        cell_str = re.sub('\|?(et al)|(Associates)', '', cell_str)

    if col_num == 2: #Title
        # Detect any ALL-CAPS titles, and change to title-case
        onlyletters = re.sub('[^A-Za-z]+', '', cell_str)
        if onlyletters.isupper:
            cell_str = cell_str.title()

    if col_num == 4: #TLA
        if cell_str == "All":     cell_str = ""
        if cell_str == "n/a":     cell_str = ""
        if cell_str == "various": cell_str = "" # not a meaningful TLA


    return cell_str


def excel_to_bound_pdf_csv(excel_filename):
    workbook = xlrd.open_workbook(excel_filename)

    worksheet = workbook.sheet_by_index(xlsutil.worksheet_index_pos)
    ## worksheet = workbook.sheet_by_name(xlsutil.worksheet_name)

    excel_filename_split = os.path.split(excel_filename)
    excel_dirname = excel_filename_split[0]
    excel_file = excel_filename_split[1]
    excel_file_splitext = os.path.splitext(excel_file)
    excel_file_root = excel_file_splitext[0];

    pdfbound_filename = os.path.join(excel_dirname,"PDF-BOUND "+excel_file_root+".csv")
    unbound_filename  = os.path.join(excel_dirname,"UNBOUND "+excel_file_root+".csv")

    csv_pdfbound_ofile = open(pdfbound_filename, 'wb')
    csv_unbound_ofile  = open(unbound_filename, 'wb')

    pdfbound_wr = csv.writer(csv_pdfbound_ofile, quoting=csv.QUOTE_ALL)
    unbound_wr  = csv.writer(csv_unbound_ofile, quoting=csv.QUOTE_ALL)

    # 1. For header-line, build up hashmap of header-names to column number
    # 2. Write out header-line as Greenstone friendly Metadata terms 
    #      => remove sp spaces, change "Report Identifier" to hnz.Identifier
    # 3. Process the rest of the file, checking if accompanying 
    #     PDF file present or not and only printing out the header-names
    #     specified in 'sanitized_headers'

    num_rows = worksheet.nrows

    # 1. Get header-line hashmap of header-names to column numbers
    header_names_mapping = xlsutil.getHeaderMappingToColNum(worksheet)

    if header_names_mapping is None:
        print "Failed to find metadata label 'Report Identifier' in column 0"
        exit()
    else:
        # 2. Process Header into Greenstone friendly metadata terms     

        entry_utf8_row = []
        row_i = 0;

        for header_name in xlsutil.sanitized_headers:

            if header_name in header_names_mapping:
                header_col_j = header_names_mapping[header_name]

                header_cell_value = worksheet.cell_value(row_i,header_col_j)
                
                if header_cell_value == "Report Identifier":
                    header_cell_value = "hnz.Identifier"

                if header_cell_value == "Title":
                    header_cell_value = "dc.Title"

                if header_cell_value == "RelevantTLA's":
                    header_cell_value = "TLA"

                header_cell_value_utf8 = unicode(header_cell_value).encode("utf-8")
                metadata_name_utf8 = header_cell_value_utf8.replace(" ", "")

                entry_utf8_row.append(metadata_name_utf8)

            else:
                print("Warning: Failed to column mapping in spreadsheet for header name \""+header_name+"\" => skipping")

        unbound_wr.writerow(entry_utf8_row)

        # Add in 'Filename' as first column in pdfbound CSV file
        entry_utf8_row.insert(0, "Filename")
        pdfbound_wr.writerow(entry_utf8_row)
                      
        # 3. Process the rest of the file (metadata values) ...

        row_i = row_i + 1
        while row_i<num_rows:
            num_cols = worksheet.row_len(row_i)
            formatted_utf8_row = []

            found_pdf = False

#            for col_j in range(num_cols):
            for header_name in xlsutil.sanitized_headers:

                if header_name in header_names_mapping:
                    col_j = header_names_mapping[header_name]

                    cell_value = worksheet.cell_value(row_i,col_j)

                    formatted_cell_value = xlsutil.format_if_int(cell_value)

                    if header_name == "Report Identifier":
                        # Check to see if companion PDF file present
                        # pdf_filename = os.path.join(excel_dirname,"pdfs",formatted_cell_value+".pdf")
                        id = formatted_cell_value.replace(" ","")
                        
                        pdf_file = id_to_relative_pdf_file(id)
                        pdf_file_root, pdf_ext = os.path.splitext(pdf_file)
                        PDF_file = pdf_file_root + ".PDF"
                    
                        pdf_filename = os.path.join(excel_dirname,pdf_file)
                        PDF_filename = os.path.join(excel_dirname,PDF_file)

                        if os.path.exists(pdf_filename):
                            found_pdf = True
                            formatted_utf8_row.insert(0, pdf_file)
                        elif os.path.exists(PDF_filename):
                            found_pdf = True
                            formatted_utf8_row.insert(0, PDF_file)
                        else:
                            print "Unbound id: '" + id + "'"

                    if isinstance(formatted_cell_value, basestring):
                        # Remove any trailing whitespace.  
                        # Newline at end particular harmful for a entry in the CSV file
                        formatted_cell_value = formatted_cell_value.rstrip();

                    formatted_cell_value_utf8 = unicode(formatted_cell_value).encode("utf-8")

                    ## Perform any cell transformations to make DL used spreadsheet
                    ## cleaner to build
                    # formatted_cell_value_utf8 = fixup_cell(col_j,formatted_cell_value_utf8)

                    formatted_utf8_row.append(formatted_cell_value_utf8)
                else:
                    print "Warning: No column number mapping for header name \""+header_name+"\" => skipping"

            if found_pdf:
                pdfbound_wr.writerow(formatted_utf8_row)
            else:
                unbound_wr.writerow(formatted_utf8_row)

            row_i = row_i + 1


    csv_pdfbound_ofile.close()
    csv_unbound_ofile.close()

if __name__ == "__main__":
    excel_to_bound_pdf_csv(os.sys.argv[1])