Mercurial > repos > dfornika > ncbi_acc_download
changeset 30:e098e839ff97 draft default tip
"planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/ncbi_acc_download commit 1531ae67584525f6eeec71a72d7861f4ad48f96a-dirty"
| author | dfornika |
|---|---|
| date | Mon, 02 Dec 2019 23:04:09 +0000 |
| parents | 94aa25f863d6 |
| children | |
| files | ncbi_acc_download.xml test-data/accessions |
| diffstat | 1 files changed, 85 insertions(+), 7 deletions(-) [+] |
line wrap: on
line diff
--- a/ncbi_acc_download.xml Fri Nov 29 23:31:08 2019 +0000 +++ b/ncbi_acc_download.xml Mon Dec 02 23:04:09 2019 +0000 @@ -8,22 +8,36 @@ </requirements> <command detect_errors="exit_code"><![CDATA[ #if $query_source.select == "accession_file": - grep -v "^[ \t]*$" $query_source.accession_file > accessions && + { grep -v "^[ \t]*$" $query_source.accession_file > accessions || + { echo "No accession numbers in input. Aborting." 1>&2; exit 1; } } && #else if $query_source.select == "accession_list": echo '$query_source.accession_list' | sed -r 's/(\,|__cn__)/\n/g' | grep -v "^[ \t]*$" > accessions && #end if mkdir outdir && cd outdir && + ignore_errors=$ignore_failed && while read accession; do - echo "Downloading accession number: " \$accession && + echo "Downloading accession number: " \$accession " ..." >> ../error.log && ncbi-acc-download --molecule '${molecule.select}' --format '${molecule.format}' #if $molecule.format != 'featuretable' and $molecule.format != 'gff3': --extended-validation all #end if - \${accession} || echo \$accession >> ../failed.txt && sleep 2; - done < ../accessions 2> >(tee -a ../error.log >&2) + \${accession}; + failure=\$?; + if [ \$failure -ne 0 ]; then + echo " failed." >> ../error.log; + if [ \$ignore_errors -ne 0 ]; then + echo \$accession >> ../failed.txt; + else + exit 1; + fi; + else + echo " done." >> ../error.log; + fi; + sleep 2; + done < ../accessions 2> >(tee -a ../error.log >&2); ]]></command> <inputs> <conditional name="query_source"> @@ -35,7 +49,9 @@ <param label="Accession File" name="accession_file" type="data" format="txt,tabular"/> </when> <when value="accession_list"> - <param label="ID List" name="accession_list" type="text" area="true" help="Newline/Comma separated list of IDs"/> + <param label="ID List" name="accession_list" type="text" area="true" help="Newline/Comma separated list of IDs"> + <validator type="expression" message="ID list cannot be empty">value.strip()</validator> + </param> </when> </conditional> <conditional name="molecule"> @@ -57,6 +73,11 @@ </param> </when> </conditional> + <param name="ignore_failed" type="select" display="radio" + label="How to handle download failures"> + <option value="0">Abort with error on first failure</option> + <option value="1">Add accession to failed list and continue</option> + </param> </inputs> <outputs> <collection name="output" type="list" label="${tool.name} on ${on_string}: Downloaded Files"> @@ -65,8 +86,10 @@ <discover_datasets pattern="(?P<name>.+)\.gff$" directory="outdir" format="gff"/> <discover_datasets pattern="(?P<name>.+)\.ft$" directory="outdir" format="txt"/> </collection> - <data name="error_log" from_work_dir="error.log" label="${tool.name} on ${on_string}: Error Log" format="txt"/> - <data name="failed_accessions" from_work_dir="failed.txt" label="${tool.name} on ${on_string}: Failed accessions" format="txt"/> + <data name="error_log" from_work_dir="error.log" label="${tool.name} on ${on_string}: Log" format="txt"/> + <data name="failed_accessions" from_work_dir="failed.txt" label="${tool.name} on ${on_string}: Failed accessions" format="txt"> + <filter>str(ignore_failed)=='1'</filter> + </data> </outputs> <tests> <test> @@ -186,6 +209,61 @@ </conditional> <conditional name="query_source"> <param name="select" value="accession_list" /> + <param name="accession_list" value="CP011064,CP0XXXXX,CP021680"/> + </conditional> + <param name="ignore_failed" value="1" /> + <output_collection name="output" type="list"> + <element name="CP011064" ftype="fasta"> + <assert_contents> + <has_line line=">CP011064.1 Escherichia coli str. Sanji plasmid pSJ_94, complete sequence" /> + </assert_contents> + </element> + <element name="CP021680" ftype="fasta"> + <assert_contents> + <has_line line=">CP021680.1 Escherichia coli strain AR_0162 plasmid tig00002623, complete sequence" /> + </assert_contents> + </element> + </output_collection> + <output name="failed_accessions"> + <assert_contents> + <has_line line="CP0XXXXX" /> + </assert_contents> + </output> + </test> + <test> + <conditional name="molecule"> + <param name="select" value="nucleotide"/> + <param name="format" value="fasta"/> + </conditional> + <conditional name="query_source"> + <param name="select" value="accession_list" /> + <param name="accession_list" value="CP0XXXXX"/> + </conditional> + <param name="ignore_failed" value="1" /> + <output name="failed_accessions"> + <assert_contents> + <has_line line="CP0XXXXX" /> + </assert_contents> + </output> + </test> + <test expect_failure="true"> + <conditional name="molecule"> + <param name="select" value="nucleotide"/> + <param name="format" value="fasta"/> + </conditional> + <conditional name="query_source"> + <param name="select" value="accession_list" /> + <param name="accession_list" value="CP011064,CP0XXXXX,CP021680"/> + </conditional> + <param name="ignore_failed" value="0" /> + </test> + <test> + <conditional name="molecule"> + <param name="select" value="nucleotide"/> + <param name="format" value="fasta"/> + </conditional> + <conditional name="query_source"> + <param name="select" value="accession_list" /> <param name="accession_list" value="CP011064 CP021680"/> </conditional> <output_collection name="output" type="list">