Mercurial > repos > estrain > data_manager_amrfinderplus

--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/data_manager_amrfinderplus/README	Fri Mar 18 11:53:03 2022 +0000
@@ -0,0 +1,1 @@
+AMRFinderPlus Data Manager
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/data_manager_amrfinderplus/data_manager/data_manager_amrfinderplus.py	Fri Mar 18 11:53:03 2022 +0000
@@ -0,0 +1,101 @@
+#!/usr/bin/env python
+# Errol Strain, estrain@gmail.com
+# Database downloads for NCBI AMRFinderPlus
+
+import sys
+import os
+import tempfile
+import shutil
+import json
+import re
+import argparse
+from ftplib import FTP
+
+
+def download_from_ncbi():
+    NCBI_FTP_SERVER = 'ftp.ncbi.nlm.nih.gov'
+    FILENAME = 'version.txt'
+    NCBI_DOWNLOAD_PATH = '/pathogen/Antimicrobial_resistance/AMRFinderPlus/database/latest/'
+
+    email = 'anonymous@example.com'
+
+    ftp = FTP( NCBI_FTP_SERVER )
+    ftp.login( 'anonymous', email)
+    ftp.cwd(NCBI_DOWNLOAD_PATH)
+
+    #exclude the allele counts folder
+    files = ftp.nlst()
+    files = filter(lambda x: re.search(r'^((?!allele|(?:invokername=allele)).)*$', x), files)
+
+    for f in files:
+      ftp.retrbinary("RETR " + f, open(f, 'wb').write)
+
+    files = ftp.nlst()
+    files = filter(lambda x: re.search(r'^((?!allele|(?:invokername=allele)).)*$', x), files)
+    pointmuts = filter(lambda x: re.search(r'^((?!tab|(?:invokername=tab)).)*$', x), files)
+    pointmuts = filter(lambda x: re.search(r'AMR_DNA-', x), pointmuts)
+
+
+    # Make blast databases
+    blastcmd = "makeblastdb -in AMRProt -dbtype prot -logfile /dev/null"
+    os.system(blastcmd)
+    blastcmd = "makeblastdb -in AMR_CDS -dbtype nucl -logfile /dev/null"
+    os.system(blastcmd)
+
+    for f in pointmuts:
+      blastcmd = "makeblastdb -in " + f +" -dbtype nucl -logfile /dev/null"
+      os.system(blastcmd)
+
+    # Make HMM indexes
+    hmmcmd="hmmpress -f AMR.LIB > /dev/null 2> /dev/null"
+    os.system(hmmcmd)
+
+    # Read in version
+    with open("version.txt") as f:
+      version = f.readline().rstrip()
+
+    ftp.quit()
+
+    return version
+
+def print_json (version,argskey,argspath,argsname,argsvalue,argsout):
+
+    valueentry="AMRFinderPlus_"+version
+
+    data_table_entry = {
+      'data_tables' : {
+        'amrfinder_databases': [
+          {
+            "dbkey":argskey,
+            "value":argsvalue,
+            "name":argsname,
+            "path":argspath,
+          }
+        ]
+      }
+    }
+
+    with open(argsout, 'w') as fh:
+      json.dump(data_table_entry,fh, sort_keys=True)
+
+def main():
+
+    parser = argparse.ArgumentParser(description='Download NCBI amrFinderPlus Databases')
+    parser.add_argument('--name', type=str, required=True, nargs=1, help='Database name')
+    parser.add_argument('--path', type=str, required=True, nargs=1, help='Absolute path to new database folder')
+    parser.add_argument('--value', type=str, required=True, nargs=1, help='Database value')
+    parser.add_argument('--key', type=str, required=True, nargs=1, help='Database value')
+    parser.add_argument('--out', type=str, required=True, nargs=1, help='output file')
+
+    args = parser.parse_args()
+
+    os.mkdir("output")
+    os.chdir("output")
+    #Fetch the files and build blast databases
+    version=download_from_ncbi()
+    os.chdir("..")
+
+    shutil.copytree("output",args.path[0])
+    print_json(version,args.key[0],args.path[0],args.name[0],args.value[0],args.out[0])
+
+if __name__ == "__main__": main()
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/data_manager_amrfinderplus/data_manager/data_manager_amrfinderplus.xml	Fri Mar 18 11:53:03 2022 +0000
@@ -0,0 +1,29 @@
+<tool id="amrfinderplus_database_builder" name="AMRFinderPlus" tool_type="manage_data" version="0.0.1" profile="16.01">
+    <description> Database builder</description>
+    <requirements>
+        <requirement type="package">blast</requirement>
+        <requirement type="package">hmmer</requirement>
+    </requirements>
+    <command detect_errors="exit_code"><![CDATA[
+        python $__tool_directory__/data_manager_amrfinderplus_database_builder.py --path $amrdbpath --name $amrdbname --value $amrdbvalue --key $amrdbkey --out ${out_file};
+    ]]></command>
+    <inputs>
+      <param name="amrdbkey" type="text" label="Unique Database Key"/>
+      <param name="amrdbpath" type="text" label="Absolute path of new amrFinderPlus database folder"/>
+      <param name="amrdbname" type="text" label="Name of amrFinderPlus database"/>
+      <param name="amrdbvalue" type="text" label="Value of amrFinderPlus database"/>
+    </inputs>
+    <outputs>
+        <data name="out_file" format="data_manager_json" dbkey="amrdbkey"/>
+    </outputs>
+    <tests>
+        <test>
+        </test>
+    </tests>
+    <help>
+    </help>
+    <citations>
+     Refer to https://ftp.ncbi.nlm.nih.gov/pathogen/Antimicrobial_resistance/AMRFinderPlus/database/latest/ for latest database version number.
+
+    </citations>
+</tool>
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/data_manager_amrfinderplus/data_manager_conf.xml	Fri Mar 18 11:53:03 2022 +0000
@@ -0,0 +1,13 @@
+<?xml version="1.0"?>
+<data_managers>
+    <data_manager tool_file="data_manager/data_manager_amrfinderplus.xml" id="data_manager_amrfinderplus">
+      <data_table name="amrfinder_databases">
+        <output>
+          <column name="value" />
+          <column name="name" />
+          <column name="path" />
+          <column name="dbkey" />
+        </output>
+      </data_table>
+    </data_manager>
+</data_managers>
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/data_manager_amrfinderplus/tool_data_table_conf_sample.xml.sample	Fri Mar 18 11:53:03 2022 +0000
@@ -0,0 +1,12 @@
+<tables>
+    <!-- Locations of all fasta files required to build Diamond databases -->
+    <table name="diamond_database" comment_char="#">
+        <columns>value, name, db_path</columns>
+        <file path="tool-data/diamond_database.loc" />
+    </table>
+    <!-- Locations of taxonomy data downloaded from NCBI -->
+    <table name="ncbi_accession2taxid" comment_char="#">
+        <columns>value, name, path</columns>
+        <file path="tool-data/ncbi_accession2taxid.loc" />
+    </table>
+</tables>
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/data_manager_amrfinderplus/tool_data_table_conf_sample.xml.test	Fri Mar 18 11:53:03 2022 +0000
@@ -0,0 +1,7 @@
+<tables>
+    <!-- Locations of fasta files to build amrfinderplus databases -->
+    <table name="amrfinderplus_databases" comment_char="#">
+        <columns>value, name, db_path</columns>
+        <file path="${__HERE__}/test-data/amrfinderplus_database.loc" />
+    </table>
+</tables>