Mercurial > repos > galaxytrakr > aws_sra

comparison aws_sra.xml @ 7:3529ce6bcde7 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/CFSAN-Biostatistics/galaxytrakr-tools commit 74cea66080315c44e81547de5fe02f046bcaa1f0
author galaxytrakr
date Mon, 23 Mar 2026 18:32:46 +0000
parents 8c60cd4c0ca7
children ef51d44623c2
comparison
equal deleted inserted replaced
6:8c60cd4c0ca7 7:3529ce6bcde7
1 <tool id="aws_sra" name="NCBI SRA AWS Fetch" version="0.3.0+gt_0.6" profile="23.0"> 1 <tool id="aws_sra" name="NCBI SRA AWS Fetch" version="0.3.0+gt_0.7" profile="23.0">
2 <description>Fetch SRA data files from NCBI's public AWS S3 buckets</description> 2 <description>Fetch SRA data files from NCBI's public AWS S3 buckets</description>
3 3
4 <requirements> 4 <requirements>
5 <requirement type="package" version="2.34.8">awscli</requirement> 5 <requirement type="package" version="2.34.8">awscli</requirement>
6 <requirement type="package" version="3.2.1">sra-tools</requirement> 6 <requirement type="package" version="3.2.1">sra-tools</requirement>
47 47
48 ## ── FASTQ DUMP mode (sra-pub-run-odp only) ─────────────────────────────── 48 ## ── FASTQ DUMP mode (sra-pub-run-odp only) ───────────────────────────────
49 #elif $action.mode == 'fastq_dump' 49 #elif $action.mode == 'fastq_dump'
50 #set $acc = $action.accession.strip() 50 #set $acc = $action.accession.strip()
51 51
52 ## 1. Download the .sra file from S3 52 ## 1. Download the file from S3
53 mkdir -p sra_cache && 53 mkdir -p sra_cache &&
54 aws s3 cp 54 aws s3 cp \
55 --no-sign-request 55 --no-sign-request \
56 '${s3_base}/sra/${acc}/${acc}.sra' 56 '${s3_base}/sra/${acc}/${acc}' \
57 ./sra_cache/${acc}.sra && 57 ./sra_cache/${acc} &&
58 58
59 ## 2. Convert with fasterq-dump --split-files. 59 ## 2. Convert with fasterq-dump (CORRECTED: uses extensionless file)
60 ## Paired runs → <acc>_1.fastq + <acc>_2.fastq
61 ## Single runs → <acc>.fastq (no _1/_2 suffix)
62 ## We always use --split-files; single-end runs simply produce one file.
63 mkdir -p fastq_out && 60 mkdir -p fastq_out &&
64 fasterq-dump 61 fasterq-dump \
65 ./sra_cache/${acc}.sra 62 ./sra_cache/${acc} \
66 --outdir ./fastq_out 63 --outdir ./fastq_out \
67 --temp . 64 --temp . \
68 --threads \${GALAXY_SLOTS:-4} 65 --threads \${GALAXY_SLOTS:-4} \
69 --split-files 66 --split-files
70 && 67 &&
71 68
72 ## 3. Compress with pigz (fasterq-dump does not gzip natively) 69 ## 3. Compress with pigz
73 pigz -p \${GALAXY_SLOTS:-4} ./fastq_out/*.fastq && 70 pigz -p \${GALAXY_SLOTS:-4} ./fastq_out/*.fastq &&
74 71
75 ## 4. Stage outputs 72 ## 4. Stage outputs
76 #if $action.layout == 'paired' 73 #if $action.layout == 'paired'
77 cp ./fastq_out/${acc}_1.fastq.gz '$output_r1' && 74 cp ./fastq_out/${acc}_1.fastq.gz '$output_r1' &&
145 <!-- ── FASTQ DUMP ── --> 142 <!-- ── FASTQ DUMP ── -->
146 <when value="fastq_dump"> 143 <when value="fastq_dump">
147 <param name="accession" type="text" label="SRA Accession" 144 <param name="accession" type="text" label="SRA Accession"
148 help="SRA run accession to fetch and convert, e.g. SRR000001. Must be present in sra-pub-run-odp."> 145 help="SRA run accession to fetch and convert, e.g. SRR000001. Must be present in sra-pub-run-odp.">
149 <validator type="empty_field" message="An SRA accession is required."/> 146 <validator type="empty_field" message="An SRA accession is required."/>
150 <validator type="regex" message="Must be a valid SRA run accession (SRR, ERR, or DRR followed by digits).">^\[SED\]RR\[0-9\]+$</validator> 147 <validator type="regex" message="Must be a valid SRA run accession (SRR, ERR, or DRR followed by digits).">^[SED]RR[0-9]+$</validator>
151 </param> 148 </param>
152 <param name="layout" type="select" label="Read layout" 149 <param name="layout" type="select" label="Read layout"
153 help="Paired-end produces two datasets labelled accession_1 and accession_2. Single-end produces one dataset. Check the SRA record to confirm layout before running."> 150 help="Paired-end produces two datasets labelled accession_1 and accession_2. Single-end produces one dataset. Check the SRA record to confirm layout before running.">
154 <option value="paired" selected="true">Paired-end (R1 + R2)</option> 151 <option value="paired" selected="true">Paired-end (R1 + R2)</option>
155 <option value="single">Single-end</option> 152 <option value="single">Single-end</option>