annotate xlsx2tsv.py @ 0:a21347be425d draft default tip

planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
author ufz
date Mon, 16 Dec 2024 20:56:29 +0000
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
1 import argparse
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
2
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
3 import pandas as pd
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
4
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
5
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
6 def convert_xlsx_to_tsv(input_file, sheet_name, output):
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
7 try:
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
8 # Read the specified sheet and convert them to tsv
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
9 df = pd.read_excel(input_file, sheet_name=sheet_name)
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
10 df.to_csv(output, sep='\t', index=False)
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
11 print(f"Extracted sheet '{sheet_name}' from {input_file}")
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
12
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
13 except Exception as e:
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
14 print(f"Failed to convert sheet '{sheet_name}' from {input_file}: {e}")
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
15
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
16
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
17 def main():
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
18 parser = argparse.ArgumentParser(description="Convert specific sheets from a single .xlsx file to .tsv format in the same directory.")
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
19 parser.add_argument("--input-file", type=str, required=True, help="Path to the input .xlsx file.")
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
20 parser.add_argument("--sheet-names", type=str, required=True, help="Comma-separated list of sheet names to convert.")
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
21 parser.add_argument("--output", type=str, default="extracted_sheet.tsv", required=False, help="Suffix for the tsv file")
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
22 args = parser.parse_args()
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
23
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
24 # Convert sheet names from str to list
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
25 sheet_names = args.sheet_names
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
26
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
27 # Call the conversion function with the provided arguments
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
28 convert_xlsx_to_tsv(args.input_file, sheet_names, args.output)
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
29
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
30
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
31 if __name__ == "__main__":
a21347be425d planemo upload for repository https://github.com/Helmholtz-UFZ/galaxy-tools/tree/main/tools/xls2tsv commit 2e75a14496fa80104e76b307289c58b9b7013ae1
ufz
parents:
diff changeset
32 main()