def hash_me(file_dict):
uris = {}
uris[file_dict['basename']] = {}
uris[file_dict['basename']]['uriIsolate'] = gu(
':spfy' + str(file_dict['count']))
uris[file_dict['basename']]['uriGenome'] = gu(
':' + generate_hash(file_dict['withpath']))
return uris
def spfyids_single(args_dict):
from settings import database
# this is temporary, TODO: include a spqarql query to the db
uriIsolate = gu(':spfy' + str(database['count']))
uriGenome = gu(':' + generate_hash(args_dict['i']))
args_dict['uriIsolate'] = uriIsolate
args_dict['uriGenome'] = uriGenome
return args_dict
def parse_serotype(graph, serotyper_dict, uriIsolate):
if 'O type' in serotyper_dict:
graph.add(
(uriIsolate, gu('ge:0001076'), Literal(serotyper_dict['O type'])))
if 'H type' in serotyper_dict:
graph.add(
(uriIsolate, gu('ge:0001077'), Literal(serotyper_dict['H type'])))
if 'K type' in serotyper_dict:
graph.add(
(uriIsolate, gu('ge:0001684'), Literal(serotyper_dict['K type'])))
return graph
def generate_turtle_skeleton(graph, fasta_file, uriIsolate, uriGenome):
'''
Handles the main generation of a turtle object.
NAMING CONVENTIONS:
uriIsolate: this is the top-most entry, a uniq. id per file is allocated by checking our DB for the greatest most entry (not in this file)
ex. :spfy234
uriAssembly: aka. the genome ID, this is a sha1 hash of the file contents
ex. :4eb02f5676bc808f86c0f014bbce15775adf06ba
uriContig: indiv contig ids; from SeqIO.record.id - this should be uniq to a contig (at least within a given file)
ex. :4eb02f5676bc808f86c0f014bbce15775adf06ba/contigs/FLOF01006689.1
note: the record.id is what RGI uses as a prefix for ORF_ID (ORF_ID has additional _314 or w/e #s)
Args:
graph(rdflib.Graph): the graph instance that is 1:1 with a .fasta file
fasta_file(str): path to the .fasta file (this should already incl the directory)
spfyID(hash): currently a hash value generated from the name of the fasta file
Returns:
graph: the graph with all the triples generated from the .fasta file
'''
# ex. :spfy234
graph.add((uriIsolate, gu('rdf:type'), gu('ncbi:562')))
graph.add((uriIsolate, gu('ge:0001567'), Literal("bacterium")))
graph.add((uriIsolate, gu('dc:description'),
Literal(uri_to_basename(uriIsolate))))
# ex. :4eb02f5676bc808f86c0f014bbce15775adf06ba
# associatting isolate URI with assembly URI
graph.add((uriIsolate, gu('g:Genome'), uriGenome))
# this is used as the human readable display of Genome
graph.add((uriGenome, gu('dc:description'), Literal(basename(fasta_file))))
# uri for bag of contigs
# ex. :4eb02f5676bc808f86c0f014bbce15775adf06ba/contigs/
uriContigs = gu(uriGenome, "/contigs")
graph.add((uriGenome, gu('so:0001462'), uriContigs))
for record in SeqIO.parse(open(fasta_file), "fasta"):
# ex. :4eb02f5676bc808f86c0f014bbce15775adf06ba/contigs/FLOF01006689.1
uriContig = gu(':' + record.id)
# linking the spec contig and the bag of contigs
graph.add((uriContigs, gu('g:Contig'), uriContig))
graph.add((uriContig, gu('g:DNASequence'), Literal(record.seq)))
graph.add(
(uriContig, gu('g:Description'), Literal(record.description)))
graph.add(
(uriContig, gu('dc:description'), Literal(record.description)))
return graph
def parse_gene_dict(graph, gene_dict, uriGenome):
'''
My intention is to eventually use ECTyper for all of the calls it was meant for.
Just need to update ECTyper dict format to ref. AMR / VF by contig. as opposed to genome directly.
These are the common gene related triples to both AMR / VF.
Note: we are working from uriGenome and assume that the calling functions (
generate_amr() and generate_vf() are doing the transformations to the
gene_dict.keys so that they are contig ids (as they differ in return value
between VF & AMR from ECTyper)
)
TODO: offshore rgi calls to ectyper and make it return a dict in the format we need
-currently, we'll handle ORF_ID to contig id transform in generate_amr()
Args:
graph(rdflib.Graph): the running graph with all our triples
gene_dict({{}}): a dictionary of genes with a assoc info
ex. {'Some_Contig_ID':[{'START','STOP','ORIENTATION','GENE_NAME'}]}
uriGenome(rdflib.URIRef): the base uri of the genome
ex. :4eb02f5676bc808f86c0f014bbce15775adf06ba
TODO: merge common components with generate_amr()
'''
for contig_id in gene_dict.keys():
for gene_record in gene_dict[contig_id]:
# recreating the contig uri
uriContig = gu(':' + contig_id) # now at contig uri
# after this point we switch perspective to the gene and build down to
# relink the gene with the contig
bnode_occurrence = BNode()
bnode_start = BNode()
bnode_end = BNode()
# some gene names, esp those which are effectively a description,
# have spaces
gene_name = gene_record['GENE_NAME'].replace(' ', '_')
graph.add(
(gu(':' + gene_name), gu('faldo:Region'), bnode_occurrence))
graph.add((bnode_occurrence, gu('faldo:Begin'), bnode_start))
graph.add((bnode_occurrence, gu('faldo:End'), bnode_end))
# this is a special case for amr results
if 'CUT_OFF' in gene_dict.keys():
graph.add((bnode_start, gu('dc:Description'),
Literal(amr_results['CUT_OFF'][i])))
graph.add((bnode_end, gu('dc:Description'),
Literal(amr_results['CUT_OFF'][i])))
graph.add((bnode_start, gu('rdf:type'), gu('faldo:Position')))
graph.add((bnode_start, gu('rdf:type'), gu('faldo:ExactPosition')))
graph.add((bnode_end, gu('rdf:type'), gu('faldo:Position')))
graph.add((bnode_end, gu('rdf:type'), gu('faldo:ExactPosition')))
if gene_record['ORIENTATION'] is '+':
graph.add((bnode_start, gu('rdf:type'),
gu('faldo:ForwardStrandPosition')))
graph.add((bnode_end, gu('rdf:type'),
gu('faldo:ForwardStrandPosition')))
else:
graph.add((bnode_start, gu('rdf:type'),
gu('faldo:ReverseStrandPosition')))
graph.add((bnode_end, gu('rdf:type'),
gu('faldo:ReverseStrandPosition')))
graph.add((bnode_start, gu('faldo:Position'),
Literal(gene_record['START'])))
graph.add((bnode_start, gu('faldo:Reference'), uriContig))
graph.add((bnode_end, gu('faldo:Position'),
Literal(gene_record['STOP'])))
graph.add((bnode_end, gu('faldo:Reference'), uriContig))
####
return graph
# mainly used for when a func needs a lot of the args
args_dict = vars(args)
# starting#logging
'''
logging.basicConfig(
filename='outputs/' + __name__ +
args_dict['i'].split('/')[-1] + '.log',
level=logging.INFO
)
'''
# check if a genome uri isn't set yet
if args_dict['uriIsolate'] is None:
# this is temporary, TODO: include a spqarql query to the db
uriIsolate = gu(':spfy' + str(hash(args_dict['i'].split('/')[-1])))
else:
uriIsolate = gu(':spfy' + args_dict['uriIsolate'])
# if the fasta_file hash was not precomputed (batch scripts should
# precompute it), we compute that now
if args_dict['uriGenome'] is None:
uriGenome = gu(':' + generate_hash(args_dict['i']))
else:
uriGenome = gu(':' + args_dict['uriGenome'])
args_dict['uriIsolate'] = uriIsolate
args_dict['uriGenome'] = uriGenome
savvy(args_dict)