The requests for each main process should send a POST request with the json:

req = {
    "project_id": <id_int>,
    "pipeline_id": <id_int>,
    "process_id": <id_int>,
    "status": <status_str>,
    "log": <log_file_str>,
    "report": <report_file_str>
    "info": <str>,
    "stats": <str>
}

The curl request line:

curl \
    -H "Content-Type: application/json" \ # set data type to json
    -X POST -d '{<json>}' \ "Provide json data"
    http://<url>

• integrity_coverage
• fastqc
• fastqc_report
• trimmomatic
• spades
• process_spades
• assembly_mapping
• process_assembly_mapping
• pilon

JSON fields:

• project_id
• pipeline_id
• process_id
• run_property
• run_property_value
• type

chewBBACA quality assessment should be done not only by counting the LNF but also the:

• PLOT
• NIPH
• ALM
• ASM

Current QC processes were designed for paired end reads. This should be extended to support both single and paired end reads.

In some cases where coverage is very low, some contigs do not yield coverage information using samtools depth. The problem is in this line, where some of the output files are empty. This means that when the coverage sum is calculated with:

parallel -j ${task.cpus} echo -n {.} '"\t"' '&&' cut -f3 {} '|' paste -sd+ '|' bc >> coverages.tsv  ::: *.tab

The lines of the empty samtools output files will be malformed and will yield an error when parsing the coverages.tsv file.

Setup auto-generated API docs for the templates submodule:

• integrity_coverage
• fastqc
• fastqc_report
• trimmomatic
• trimmomatic_report
• spades
• process_spades
• assembly_report
• process_assembly_mapping
• process_abricate

run_info.json:

{"Sample name": 
    {"modules_run_report": {"spades": [null, null, 0, {"sample": "Not run"}, {"sample": "Not run"}]},
                           {<module>: [<status>,<qc>,<time>,{<fail>},{<warning>}]},
     "run_successfully": true,
     "pass_qc": "FAIL"
}

run_stats.json

{"Sample name": {<module>:<stats>}}

A QC warning/failing system could be implemented throughout the pipeline's processes. Whenever a QC step occurs, warnings could be written in a standardized way into a ".warnings" file that is sent through the main Status channels and through POST requests.

The Warnings should be the following:

• FastQC
  • Per base sequence quality != pass
  • Overrepresented sequences != pass
  • Bad per base sequence content == fail
• Process spades
  • Assembled genome lower than 80% or higher than 150%
  • Excessive number of contigs
• Process assembly mapping
  • Low mapped reads
  • Assembled genome lower than 80% or higher than 150%
  • Excessive number of contigs

The Failures should be the following:

• Integrity coverage
  • File corruption
  • Low coverage
• FastQC (2)
  • Per base sequence quality Fail
  • Overrepresented sequences Fail
  • Per sequence GC content Fail
  • Per base N content != pass
  • Sequence length distribution Fail
  • Adapter content != pass