ValueError: could not convert string to float: '29,6455'

[Marcopterix]

Hello,
When I try to run vSNP3 with the test data I get this error with the new version

vsnp3_step1.py -r1 _R1.fastq.gz -r2 _R2.fastq.gz -t Mycobacterium_AF2122

vsnp3_step1.py SET ARGUMENTS:
Namespace(SAMPLE_NAME=None, FASTQ_R1='13-1941-6_S4_L001_R1_600000.fastq.gz', FASTQ_R2='13-1941-6_S4_L001_R2_600000.fastq.gz', FASTAtoFASTQ=None, FASTA=None, gbk=None, reference_type='Mycobacterium_AF2122', nanopore=False, output_dir=None, assemble_unmap=False, spoligo=False, debug=False)

Traceback (most recent call last):
File "/home/bioinfocenasa/Programas_bioinfo/vSNP3/bin/vsnp3_step1.py", line 301, in
vsnp.run()
File "/home/bioinfocenasa/Programas_bioinfo/vSNP3/bin/vsnp3_step1.py", line 128, in run
fastq_stats.latex(self.latex_report.tex)
File "/home/bioinfocenasa/Programas_bioinfo/vSNP3/bin/vsnp3_fastq_stats_seqkit.py", line 119, in latex
print(f'Mean Read Score & {float(self.R1.read_quality_average):0.1f} & {float(self.R2.read_quality_average):0.1f} \\', file=tex)
ValueError: could not convert string to float: '29,6455'

[stuber]

It looks like you're encountering a locale-specific decimal separator issue in vSNP3. It appears it is failing because in some countries, the decimal separator is a comma (,) instead of a period (.), causing Python to fail when trying to convert the string "29,6455" to a float.

Here's a possible fix:

1. Temporarily change the locale settings - Users can set their locale to use period as the decimal separator before running the script:

   export LC_NUMERIC=en_US.UTF-8

   Then run the vSNP3 command.

2. Create a wrapper script that sets the locale before running vSNP3:

   #!/bin/bash
   export LC_NUMERIC=en_US.UTF-8
   /path/to/vsnp3_step1.py "$@"

3. Modify the vSNP3 code if they have access to it - In the file /home/bioinfocenasa/Programas_bioinfo/vSNP3/bin/vsnp3_fastq_stats_seqkit.py around line 119, they could add a conversion function:

   def convert_to_float(value):
       if isinstance(value, str):
           return float(value.replace(',', '.'))
       return float(value)
       
   # Then change the line 119 to:
   print(f'Mean Read Score & {convert_to_float(self.R1.read_quality_average):0.1f} & {convert_to_float(self.R2.read_quality_average):0.1f} \\', file=tex)

The first two options are probably the easiest to implement without modifying the source code.

Hope this helps,

[stuber]

Starting in version 3.29, vSNP3 will force decimal point as separator to ensure consistent parsing of numeric values across different international locale settings. The locale function will handle conversion from comma-based decimal separators to period-based ones.