Mercurial > repos > estrain > data_manager_fastani
changeset 20:1af0cdf9b1fa draft
Uploaded
| author | estrain |
|---|---|
| date | Wed, 20 Jul 2022 21:54:50 +0000 |
| parents | ceda2bd3e41d |
| children | 7fb415930708 |
| files | data_manager_fastani/data_manager/data_manager_fastani.py data_manager_fastani/data_manager/data_manager_fastani.xml data_manager_fastani/tool-data/accessions.csv |
| diffstat | 3 files changed, 58 insertions(+), 16 deletions(-) [+] |
line wrap: on
line diff
--- a/data_manager_fastani/data_manager/data_manager_fastani.py Mon May 30 00:56:12 2022 +0000 +++ b/data_manager_fastani/data_manager/data_manager_fastani.py Wed Jul 20 21:54:50 2022 +0000 @@ -1,4 +1,4 @@ -#!/usr/bin/env python +#/usr/bin/env python # Errol Strain, estrain@gmail.com # Database downloads for FastANI @@ -9,6 +9,7 @@ import re import argparse import requests +import csv def download_D1(output_directory): @@ -37,20 +38,24 @@ baseurl="https://ftp.ncbi.nih.gov/genomes/refseq/bacteria/" - urldict={'Aeromonas_caviae_strain_WP8_S18_ESBL_04':'Aeromonas_caviae/representative/GCF_014169735.1_ASM1416973v1/GCF_014169735.1_ASM1416973v1_genomic.fna.gz', - 'Aeromonas_veronii_strain_FDAARGOS_632':'Aeromonas_veronii/representative/GCF_008693705.1_ASM869370v1/GCF_008693705.1_ASM869370v1_genomic.fna.gz', - 'Aeromonas_sobria_strain_CECT_4245':'Aeromonas_sobria/representative/GCF_000820145.1_PRJEB7040/GCF_000820145.1_PRJEB7040_genomic.fna.gz', - 'Edwardsiella_piscicida_strain_18EpOKYJ':'Edwardsiella_piscicida/representative/GCF_021733145.1_ASM2173314v1/GCF_021733145.1_ASM2173314v1_genomic.fna.gz', - 'Vibrio_alginolyticus_FDAARGOS_97':'Vibrio_alginolyticus/representative/GCF_001471275.2_ASM147127v2/GCF_001471275.2_ASM147127v2_cds_from_genomic.fna.gz', - 'Vibrio_harveyi_ATCC_33843':'Vibrio_harveyi/representative/GCF_000770115.1_ASM77011v2/GCF_000770115.1_ASM77011v2_genomic.fna.gz', - 'Vibrio_rotiferianus_strain_B64D1':'Vibrio_rotiferianus/representative/GCF_002214395.1_ASM221439v1/GCF_002214395.1_ASM221439v1_genomic.fna.gz', - 'Staphylococcus_warneri_strain_22_1':'Staphylococcus_warneri/representative/GCF_003571725.1_ASM357172v1/GCF_003571725.1_ASM357172v1_genomic.fna.gz'} - +def download_VL(output_directory,accfile): + + cwd = os.getcwd() + os.chdir(output_directory) + + baseurl="https://ftp.ncbi.nih.gov/genomes/refseq/bacteria/" + + urldict = {} + with open(accfile,mode='r') as inp: + reader = csv.reader(inp) + urldict = {rows[0]:rows[1] for rows in reader} + #FastANI uses filenames in output. Creating user friendly names - #for fish pathogens - for key in urldict: - url=baseurl+urldict[key] - filename = key + '.fna.gz' + #for fish pathogens + for key in urldict: + url=baseurl+key+"/representative/"+urldict[key]+"/"+urldict[key]+"_genomic.fna.gz" + filename = key + '.fna.gz' + print(url+"\n") with open(filename, "wb") as f: r = requests.get(url) f.write(r.content) @@ -84,6 +89,7 @@ parser = argparse.ArgumentParser(description='Download FastANI Databases') parser.add_argument('--type', type=str, required=True, nargs=1, help='Database Type') parser.add_argument('--name', type=str, required=True, nargs=1, help='Unique Database Folder Name') + parser.add_argument('--acc', type=str, required=True, nargs=1, help='CSV Accession file with NCBI ftp folder IDs, see tool-data') parser.add_argument('--out', type=str, required=True, nargs=1, help='output file') args = parser.parse_args() @@ -101,7 +107,7 @@ version="FastANI D1" elif(args.type[0]=="VL"): download_D1(output_directory) - download_VL(output_directory) + download_VL(output_directory,args.acc[0]) version="FastANI D1 + VetLIRN" print_json(version,output_directory,args.name[0],args.out[0])
--- a/data_manager_fastani/data_manager/data_manager_fastani.xml Mon May 30 00:56:12 2022 +0000 +++ b/data_manager_fastani/data_manager/data_manager_fastani.xml Wed Jul 20 21:54:50 2022 +0000 @@ -2,7 +2,7 @@ <requirements> </requirements> <command detect_errors="exit_code"><![CDATA[ - python $__tool_directory__/data_manager_fastani.py --type $input_source_selector --name $name --out ${output_file}; + python $__tool_directory__/data_manager_fastani.py --type $input_source_selector --name $name --acc ../accessions.csv --out ${output_file}; ]]></command> <inputs> <param name="input_source_selector" type="select" label="Database Selection">
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/data_manager_fastani/tool-data/accessions.csv Wed Jul 20 21:54:50 2022 +0000 @@ -0,0 +1,36 @@ +Aeromonas_allosaccharophila,GCF_016026615.1_ASM1602661v1 +Aeromonas_aquatica,GCF_000764655.1_ASM76465v1 +Aeromonas_australiensis,GCF_000819725.1_PRJEB7021 +Aeromonas_bestiarum,GCF_002906925.1_ASM290692v1 +Aeromonas_bivalvium,GCF_003265465.1_ASM326546v1 +Aeromonas_cavernicola,GCF_002795305.1_ASM279530v1 +Aeromonas_caviae,GCF_014169735.1_ASM1416973v1 +Aeromonas_dhakensis,GCF_020405345.1_ASM2040534v1 +Aeromonas_diversa,GCF_000819805.1_PRJEB7026 +Aeromonas_encheleia,GCF_900637545.1_51438_G01 +Aeromonas_enteropelogenes,GCF_020341435.1_ASM2034143v1 +Aeromonas_eucrenophila,GCF_000819865.1_PRJEB7029 +Aeromonas_finlandensis,GCF_000764645.1_ASM76464v1 +Aeromonas_fluvialis,GCF_000819885.1_PRJEB7030 +Aeromonas_hydrophila,GCF_017310215.1_ASM1731021v1 +Aeromonas_jandaei,GCF_016127195.1_ASM1612719v1 +Aeromonas_lacus,GCF_000764665.1_ASM76466v1 +Aeromonas_lusitana,GCF_002812985.1_MDC2473 +Aeromonas_media,GCF_020423125.1_ASM2042312v1 +Aeromonas_molluscorum,GCF_000388115.1_Amol1.0 +Aeromonas_piscicola,GCF_000820005.1_PRJEB7033 +Aeromonas_popoffii,GCF_000820025.1_PRJEB7034 +Aeromonas_rivipollensis,GCF_010974825.1_ASM1097482v1 +Aeromonas_rivuli,GCF_020149575.1_ASM2014957v1 +Aeromonas_sanarellii,GCF_000820085.1_PRJEB7037 +Aeromonas_schubertii,GCF_001481395.1_ASM148139v1 +Aeromonas_simiae,GCF_014892695.1_ASM1489269v1 +Aeromonas_sobria,GCF_000820145.1_PRJEB7040 +Aeromonas_taiwanensis,GCF_000820165.1_PRJEB7041 +Aeromonas_tecta,GCF_000820185.1_PRJEB7042 +Aeromonas_veronii,GCF_008693705.1_ASM869370v1 +Edwardsiella_piscicida,GCF_021733145.1_ASM2173314v1 +Staphylococcus_warneri,GCF_003571725.1_ASM357172v1 +Vibrio_alginolyticus,GCF_001471275.2_ASM147127v2 +Vibrio_harveyi,GCF_000770115.1_ASM77011v2 +Vibrio_rotiferianus,GCF_002214395.1_ASM221439v1
