Fix for 0 division error in fastp module.

[bjohnnyd]

Fixes issue where a division by 0 occurs if there are 0 reads in input fastq.

Additional fix on top of #1451 which updates CHANGELOG.md in 797674f.

fixes #1466
fixes #1444

• This comment contains a description of changes (with reason)
• CHANGELOG.md has been updated

[bjohnnyd]

@ewels one thing to add is that this additional fix causes the warning to be reported twice. Would it be better to add the s_name to the warnings so that it is clear it refers to the same input sample/fastq? Essentially it would be: FastQ input file for sample <s_name> has zero raw reads.

[ewels]

On my phone so a bit tricky to review properly, but looks good - thanks! 👍🏻

Please just add to the log message which sample had zero reads. If the sample makes it through to the report (isn't skipped) and shows zero reads, then no log message is needed.

[ewels]

Could you please log the sample name here?

[ewels]

In fact I'd be inclined to just merge this with the catch / log message above.. What do you think?

[bjohnnyd]

I think it would be best to either report once per sample or report the sample in the message. What is your opinion of adding a pass instead of logging sample?

Using the data from #1499 and MultiQC build from 2dd31c1 results in the following:

  /// MultiQC 🔍 | v1.12.dev0

|           multiqc | Search path : /home/johnny/data/testing/multiqc_fastp_bug/PCRblank.fastp.json
|         searching | ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 100% 1/1
|             fastp | FastQ input file has zero raw reads
|             fastp | FastQ input file has zero raw reads
|             fastp | Found 1 reports
|           multiqc | Compressing plot data
|           multiqc | Report      : report/multiqc_report.html
|           multiqc | Data        : report/multiqc_data
|           multiqc | MultiQC complete

which I think is confusing given that it is one sample.

[bjohnnyd]

the sample does make it into the report.

[ewels]

Sorry for the delay. Yup, I think that we're saying the same thing here - as it is the error message is a bit confusing (repeated, sample is still there).

So I was suggesting that we log the same name (no longer repeated) or just merge with the statement above (debug so not in the main console output unless verbose, also specifies what is going wrong and the sample name). I think I'll go for the latter if that's ok.

[ewels]

Tweaked in fe471c6

[ewels]

Does this also fix #1499 ?