Mercurial > repos > bgruening > data_manager_diamond_database_builder
changeset 13:98de3779045a draft
Uploaded
| author | bgruening |
|---|---|
| date | Sun, 08 Feb 2015 07:46:31 -0500 |
| parents | dbd491c28590 |
| children | 34ce6933d52a |
| files | data_manager/data_manager_diamond_database_builder.py data_manager/data_manager_diamond_database_builder.xml |
| diffstat | 2 files changed, 11 insertions(+), 12 deletions(-) [+] |
line wrap: on
line diff
--- a/data_manager/data_manager_diamond_database_builder.py Sun Feb 08 07:28:34 2015 -0500 +++ b/data_manager/data_manager_diamond_database_builder.py Sun Feb 08 07:46:31 2015 -0500 @@ -5,7 +5,6 @@ import shutil import optparse import urllib2 -#import uuid import subprocess from ftplib import FTP import tarfile @@ -66,19 +65,19 @@ path_contents = _get_files_in_ftp_path( ftp, NCBI_DOWNLOAD_PATH ) - ucsc_file_name = None + ncbi_file_name = None get_stream_reader = None ext = None for ext, get_stream_reader in COMPRESSED_EXTENSIONS: if "%s%s" % ( ncbi_identifier, ext ) in path_contents: - ucsc_file_name = "%s%s%s" % ( NCBI_DOWNLOAD_PATH, ncbi_identifier, ext ) + ncbi_file_name = "%s%s%s" % ( NCBI_DOWNLOAD_PATH, ncbi_identifier, ext ) break - if not ucsc_file_name: - raise Exception( 'Unable to determine filename for UCSC Genome for %s: %s' % ( ncbi_identifier, path_contents ) ) + if not ncbi_file_name: + raise Exception( 'Unable to determine filename for NCBI database for %s: %s' % ( ncbi_identifier, path_contents ) ) - tmp_dir = tempfile.mkdtemp( prefix='tmp-data-manager-ucsc-' ) - ucsc_fasta_filename = os.path.join( tmp_dir, "%s%s" % ( ncbi_identifier, ext ) ) + tmp_dir = tempfile.mkdtemp( prefix='tmp-data-manager-ncbi-' ) + ncbi_fasta_filename = os.path.join( tmp_dir, "%s%s" % ( ncbi_identifier, ext ) ) fasta_base_filename = "%s.fa" % database_id fasta_filename = os.path.join( target_directory, fasta_base_filename ) @@ -87,9 +86,9 @@ tmp_extract_dir = os.path.join ( tmp_dir, 'extracted_fasta' ) os.mkdir( tmp_extract_dir ) - tmp_fasta = open( ucsc_fasta_filename, 'wb+' ) + tmp_fasta = open( ncbi_fasta_filename, 'wb+' ) - ftp.retrbinary( 'RETR %s' % ucsc_file_name, tmp_fasta.write ) + ftp.retrbinary( 'RETR %s' % ncbi_file_name, tmp_fasta.write ) tmp_fasta.flush() tmp_fasta.seek( 0 )
--- a/data_manager/data_manager_diamond_database_builder.xml Sun Feb 08 07:28:34 2015 -0500 +++ b/data_manager/data_manager_diamond_database_builder.xml Sun Feb 08 07:46:31 2015 -0500 @@ -1,5 +1,5 @@ -<tool id="diamond_database_builder" name="Diamond Database" tool_type="manage_data" version="0.0.1"> - <description>builder</description> +<tool id="diamond_database_builder" name="Diamond" tool_type="manage_data" version="0.0.1"> + <description> Database builder</description> <requirements> <requirement type="package" version="0.6.13">diamond</requirement> </requirements> @@ -9,7 +9,7 @@ <inputs> <param type="text" name="database_name" value="" label="Database name or description" help="This will be shown to the user in the databse selectbox."/> - <param type="text" name="database_id" value="" label="Database ID" help="This needs to be unique." /> + <param type="text" name="database_id" value="" label="Database ID" help="This needs to be unique. No whitespaces allowed. Use underscores." /> <conditional name="reference_source"> <param name="reference_source_selector" type="select" label="Choose the source for the reference genome"> <option value="ncbi">NCBI</option>