Mercurial > repos > morinlab > fetch_ensembl_cache_data_manager
changeset 0:c6b7640b24e9 draft
Uploaded
| author | morinlab |
|---|---|
| date | Sun, 04 Dec 2016 18:10:19 -0500 |
| parents | |
| children | c3150a2a03f7 |
| files | fetch_ensembl_cache_data_manager/data_manager/data_manager_fetch_ensembl_cache.py fetch_ensembl_cache_data_manager/data_manager/data_manager_fetch_ensembl_cache.xml fetch_ensembl_cache_data_manager/data_manager/ensembl_vep_macros.xml fetch_ensembl_cache_data_manager/data_manager_conf.xml fetch_ensembl_cache_data_manager/tool-data/ensembl_cache.loc fetch_ensembl_cache_data_manager/tool_data_table_conf.xml.sample fetch_ensembl_cache_data_manager/tool_dependencies.xml |
| diffstat | 7 files changed, 207 insertions(+), 0 deletions(-) [+] |
line wrap: on
line diff
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/fetch_ensembl_cache_data_manager/data_manager/data_manager_fetch_ensembl_cache.py Sun Dec 04 18:10:19 2016 -0500 @@ -0,0 +1,68 @@ +import sys +import os +import tempfile +import optparse +import subprocess + +from json import loads, dumps + +DEFAULT_DATA_TABLE_NAME = "ensembl_cache" + +def fetch_ensembl_cache(data_manager_dict, ensembl_vep_path, species, assembly, ensembl_version, params, target_directory, data_table_name): + + perl_executable = '/'.join([ensembl_vep_path, "INSTALL.pl"]) + + commands = [ "perl" , perl_executable, "--AUTO", "cf", "--SPECIES", species, "--ASSEMBLY", assembly, "--CACHEDIR", target_directory ] + + print commands + + proc = subprocess.Popen( args=commands, shell=False, cwd=target_directory) + return_code = proc.wait() + if return_code: + print >> sys.stderr, "Error fetching ensembl cache" + sys.exit( return_code ) + + data_table_entry = dict(value='.'.join([species,assembly,ensembl_version]), name='.'.join([species,assembly,ensembl_version]), species=species, assembly=assembly, version=ensembl_version, path=target_directory ) + + _add_data_table_entry( data_manager_dict, data_table_name, data_table_entry ) + +def _add_data_table_entry( data_manager_dict, data_table_name, data_table_entry ): + data_manager_dict['data_tables'] = data_manager_dict.get( 'data_tables', {} ) + data_manager_dict['data_tables'][ data_table_name ] = data_manager_dict['data_tables'].get( data_table_name, [] ) + data_manager_dict['data_tables'][ data_table_name ].append( data_table_entry ) + return data_manager_dict + +def main(): + + print >> sys.stderr, "Starting" + + parser = optparse.OptionParser() + parser.add_option( '-i', '--organism_string', dest='input', action='store', type="string") + parser.add_option( '-e', '--ensembl_vep_path', dest='ensembl_vep_path', action='store', type='string') + (options, args) = parser.parse_args() + + filename = args[0] + + print >> sys.stderr, "What" + print >> sys.stderr, "When" + + ensembl_vep_path = options.ensembl_vep_path + species = options.input.split("-")[0] + assembly = options.input.split("-")[1] + ensembl_version = 84 + + print >> sys.stderr, "hello" + + params = loads( open( filename ).read() ) + target_directory = params[ 'output_data' ][0]['extra_files_path'] + os.mkdir( target_directory ) + data_manager_dict = {} + + #Fetch the Ensembl Cache + fetch_ensembl_cache( data_manager_dict, ensembl_vep_path, species, assembly, ensembl_version, params, target_directory, data_table_name=DEFAULT_DATA_TABLE_NAME ) + + + #save info to json file + open( filename, 'wb' ).write( dumps( data_manager_dict ) ) + +if __name__ == "__main__": main()
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/fetch_ensembl_cache_data_manager/data_manager/data_manager_fetch_ensembl_cache.xml Sun Dec 04 18:10:19 2016 -0500 @@ -0,0 +1,17 @@ +<tool id="fetch_ensembl_cache_data_manager" name="Fetch Ensembl Cache" tool_type="manage_data" version="0.0.2"> + <description>fetcher</description> + <requirements> + <requirement type="package" version="84">ensembl_vep</requirement> + </requirements> + <macros> + <import>ensembl_vep_macros.xml</import> + </macros> + <command interpreter="python"> + $__tool_directory__/data_manager_fetch_ensembl_cache.py "${out_file}" --organism_string $species_select.value --ensembl_vep_path "\$ENSEMBL_VEP_84_INSTALL_DIR" </command> + <inputs> + <expand macro="genome_list"/> + </inputs> + <outputs> + <data name="out_file" format="data_manager_json"/> + </outputs> +</tool>
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/fetch_ensembl_cache_data_manager/data_manager/ensembl_vep_macros.xml Sun Dec 04 18:10:19 2016 -0500 @@ -0,0 +1,76 @@ +<macros> + <xml name="genome_list"> + <param name="species_select" type="select" label="Select Species"> + <option value="ailuropoda melanoleuca">ailuropoda melanoleuca</option> + <option value="anas platyrhynchos">anas platyrhynchos</option> + <option value="anolis carolinensis">anolis carolinensis</option> + <option value="astyanax mexicanus">astyanax mexicanus</option> + <option value="bos taurus">bos taurus</option> + <option value="caenorhabditis elegans">caenorhabditis elegans</option> + <option value="callithrix jacchus">callithrix jacchus</option> + <option value="canis familiaris">canis familiaris</option> + <option value="cavia porcellus">cavia porcellus</option> + <option value="chlorocebus sabaeus">chlorocebus sabaeus</option> + <option value="choloepus hoffmanni">choloepus hoffmanni</option> + <option value="ciona intestinalis">ciona intestinalis</option> + <option value="ciona savignyi">ciona savignyi</option> + <option value="danio rerio">danio rerio</option> + <option value="dasypus novemcinctus">dasypus novemcinctus</option> + <option value="dipodomys ordii">dipodomys ordii</option> + <option value="drosophila melanogaster">drosophila melanogaster</option> + <option value="echinops telfairi">echinops telfairi</option> + <option value="equus caballus">equus caballus</option> + <option value="erinaceus europaeus">erinaceus europaeus</option> + <option value="felis catus">felis catus</option> + <option value="ficedula albicollis">ficedula albicollis</option> + <option value="gadus morhua">gadus morhua</option> + <option value="gallus gallus">gallus gallus</option> + <option value="gasterosteus aculeatus">gasterosteus aculeatus</option> + <option value="gorilla gorilla">gorilla gorilla</option> + <option value="homo_sapiens-GRCh37">homo sapiens - GRCh37</option> + <option value="homo_sapiens-GRCh38" selected="true">homo sapiens - GRCh38</option> + <option value="ictidomys tridecemlineatus">ictidomys tridecemlineatus</option> + <option value="latimeria chalumnae">latimeria chalumnae</option> + <option value="lepisosteus oculatus">lepisosteus oculatus</option> + <option value="loxodonta africana">loxodonta africana</option> + <option value="macaca mulatta">macaca mulatta</option> + <option value="macropus eugenii">macropus eugenii</option> + <option value="meleagris gallopavo">meleagris gallopavo</option> + <option value="microcebus murinus">microcebus murinus</option> + <option value="monodelphis domestica">monodelphis domestica</option> + <option value="mus musculus">mus musculus</option> + <option value="mustela putorius furo">mustela putorius furo</option> + <option value="myotis lucifugus">myotis lucifugus</option> + <option value="nomascus leucogenys">nomascus leucogenys</option> + <option value="ochotona princeps">ochotona princeps</option> + <option value="oreochromis niloticus">oreochromis niloticus</option> + <option value="ornithorhynchus anatinus">ornithorhynchus anatinus</option> + <option value="oryctolagus cuniculus">oryctolagus cuniculus</option> + <option value="oryzias latipes">oryzias latipes</option> + <option value="otolemur garnettii">otolemur garnettii</option> + <option value="ovis aries">ovis aries</option> + <option value="pan troglodytes">pan troglodytes</option> + <option value="papio anubis">papio anubis</option> + <option value="pelodiscus sinensis">pelodiscus sinensis</option> + <option value="petromyzon marinus">petromyzon marinus</option> + <option value="poecilia formosa">poecilia formosa</option> + <option value="pongo abelii">pongo abelii</option> + <option value="procavia capensis">procavia capensis</option> + <option value="pteropus vampyrus">pteropus vampyrus</option> + <option value="rattus norvegicus">rattus norvegicus</option> + <option value="saccharomyces cerevisiae">saccharomyces cerevisiae</option> + <option value="sarcophilus harrisii">sarcophilus harrisii</option> + <option value="sorex araneus">sorex araneus</option> + <option value="sus scrofa">sus scrofa</option> + <option value="taeniopygia guttata">taeniopygia guttata</option> + <option value="takifugu rubripes">takifugu rubripes</option> + <option value="tarsius syrichta">tarsius syrichta</option> + <option value="tetraodon nigroviridis">tetraodon nigroviridis</option> + <option value="tupaia belangeri">tupaia belangeri</option> + <option value="tursiops truncatus">tursiops truncatus</option> + <option value="vicugna pacos">vicugna pacos</option> + <option value="xenopus tropicalis">xenopus tropicalis</option> + <option value="xiphophorus maculatus">xiphophorus maculatus</option> + </param> + </xml> +</macros>
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/fetch_ensembl_cache_data_manager/data_manager_conf.xml Sun Dec 04 18:10:19 2016 -0500 @@ -0,0 +1,24 @@ +<?xml version="1.0"?> +<data_managers> + + <data_manager tool_file="data_manager/data_manager_fetch_ensembl_cache.xml" id="fetch_ensembl_cache" version="0.0.1"> + <data_table name="ensembl_cache"> + <output> + <column name="value" /> + <column name="name"/> + <column name="species" /> + <column name="assembly" /> + <column name="version" /> + <column name="path" output_ref="out_file" > + <move type="directory" relativize_symlinks="True"> + <!-- <source>${path}</source>--> <!-- out_file.extra_files_path is used as base by default --> <!-- if no source, eg for type=directory, then refers to base --> + <target base="${GALAXY_DATA_MANAGER_DATA_PATH}">${species}/${assembly}/${version}</target> + </move> + <value_translation>${GALAXY_DATA_MANAGER_DATA_PATH}/${species}/${assembly}/${version}/${path}</value_translation> + <value_translation type="function">abspath</value_translation> + </column> + </output> + </data_table> + </data_manager> + +</data_managers>
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/fetch_ensembl_cache_data_manager/tool-data/ensembl_cache.loc Sun Dec 04 18:10:19 2016 -0500 @@ -0,0 +1,1 @@ +#<value> <name> <assembly> <species> <ensembl_version> <file_path>
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/fetch_ensembl_cache_data_manager/tool_data_table_conf.xml.sample Sun Dec 04 18:10:19 2016 -0500 @@ -0,0 +1,6 @@ +<tables> + <table name="ensembl_cache" comment_char="#"> + <columns>value, name, species, assembly, version, path</columns> + <file path="tool-data/ensembl_cache.loc" /> + </table> +</tables>
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/fetch_ensembl_cache_data_manager/tool_dependencies.xml Sun Dec 04 18:10:19 2016 -0500 @@ -0,0 +1,15 @@ +<?xml version="1.0"?> +<tool_dependency> + <package name="perl" version="5.18.1"> + <repository changeset_revision="fa8f8d81b13e" name="package_perl_5_18" owner="morinlab" toolshed="https://testtoolshed.g2.bx.psu.edu" /> + </package> + <package name="mysql" version="6.1.0"> + <repository changeset_revision="766f27cb3cd7" name="package_mysql_6_1_0" owner="morinlab" toolshed="https://testtoolshed.g2.bx.psu.edu" /> + </package> + <package name="ensembl_vep_perl_env" version="84"> + <repository changeset_revision="9976000ce924" name="package_ensembl_vep_perl_environment_84" owner="morinlab" toolshed="https://testtoolshed.g2.bx.psu.edu" /> + </package> + <package name="ensembl_vep" version="84"> + <repository changeset_revision="c035dd918ed4" name="package_ensembl_vep_84" owner="morinlab" toolshed="https://testtoolshed.g2.bx.psu.edu" /> + </package> +</tool_dependency>
