def _open(x, f, writer=writer, out_file=out_file):
if out_file is not None and x == 'pass':
handle = open(out_file, 'w')
else:
handle = getOutputHandle(aligner_file,
out_label='db-%s' % x,
out_dir=out_args['out_dir'],
out_name=out_args['out_name'],
out_type=out_args['out_type'])
return handle, writer(handle, fields=f)
def mergeDbFiles(db_files,
drop=False,
out_file=None,
out_args=default_out_args):
"""
Updates field and value pairs to a database file
Arguments:
db_files : list of database file names.
drop : if True drop columns not present in all files.
out_file : output file name. Automatically generated from the input file if None.
out_args : common output argument dictionary from parseCommonArgs.
Returns:
str : output file name.
"""
log = OrderedDict()
log['START'] = 'ParseDb'
log['COMMAND'] = 'merge'
log['FILES'] = ','.join([os.path.basename(f) for f in db_files])
log['DROP'] = drop
printLog(log)
# Open input
db_handles = [open(f, 'rt') for f in db_files]
db_iters = [TSVReader(x) for x in db_handles]
result_count = sum([countDbFile(f) for f in db_files])
# Define output fields
field_list = [x.fields for x in db_iters]
if drop:
field_set = set.intersection(*map(set, field_list))
else:
field_set = set.union(*map(set, field_list))
field_order = OrderedDict([(f, None) for f in chain(*field_list)])
out_fields = [f for f in field_order if f in field_set]
# Open output file
if out_file is not None:
pass_handle = open(out_file, 'w')
else:
__, __, out_args['out_type'] = splitName(db_files[0])
pass_handle = getOutputHandle(db_files[0],
out_label='parse-merge',
out_dir=out_args['out_dir'],
out_name=out_args['out_name'],
out_type=out_args['out_type'])
pass_writer = TSVWriter(pass_handle, out_fields)
# Iterate over records
start_time = time()
rec_count = 0
for db in db_iters:
for rec in db:
# Print progress for previous iteration
printProgress(rec_count, result_count, 0.05, start_time=start_time)
rec_count += 1
# Write records
pass_writer.writeDict(rec)
# Print counts
printProgress(rec_count, result_count, 0.05, start_time=start_time)
log = OrderedDict()
log['OUTPUT'] = os.path.basename(pass_handle.name)
log['RECORDS'] = rec_count
log['END'] = 'ParseDb'
printLog(log)
# Close file handles
pass_handle.close()
for x in db_handles:
x.close()
return pass_handle.name
def updateDbFile(db_file,
field,
values,
updates,
out_file=None,
out_args=default_out_args):
"""
Updates field and value pairs to a database file
Arguments:
db_file : the database file name.
field : the field to update.
values : a list of values to specifying which rows to update.
updates : a list of values to update each value with.
out_file : output file name. Automatically generated from the input file if None.
out_args : common output argument dictionary from parseCommonArgs.
Returns:
str : output file name
"""
log = OrderedDict()
log['START'] = 'ParseDb'
log['COMMAND'] = 'update'
log['FILE'] = os.path.basename(db_file)
log['FIELD'] = field
log['VALUES'] = ','.join(values)
log['UPDATES'] = ','.join(updates)
printLog(log)
# Open input
db_handle = open(db_file, 'rt')
db_iter = TSVReader(db_handle)
out_fields = db_iter.fields
__, __, out_args['out_type'] = splitName(db_file)
# Open output
if out_file is not None:
pass_handle = open(out_file, 'w')
else:
pass_handle = getOutputHandle(db_file,
out_label='parse-update',
out_dir=out_args['out_dir'],
out_name=out_args['out_name'],
out_type=out_args['out_type'])
pass_writer = TSVWriter(pass_handle, out_fields)
# Count records
result_count = countDbFile(db_file)
# Iterate over records
start_time = time()
rec_count, pass_count = 0, 0
for rec in db_iter:
# Print progress for previous iteration
printProgress(rec_count, result_count, 0.05, start_time=start_time)
rec_count += 1
# Updated values if found
for x, y in zip(values, updates):
if rec[field] == x:
rec[field] = y
pass_count += 1
# Write records
pass_writer.writeDict(rec)
# Print counts
printProgress(rec_count, result_count, 0.05, start_time=start_time)
log = OrderedDict()
log['OUTPUT'] = os.path.basename(pass_handle.name)
log['RECORDS'] = rec_count
log['UPDATED'] = pass_count
log['END'] = 'ParseDb'
printLog(log)
# Close file handles
pass_handle.close()
db_handle.close()
return pass_handle.name
def sortDbFile(db_file,
field,
numeric=False,
descend=False,
out_file=None,
out_args=default_out_args):
"""
Sorts records by values in an annotation field
Arguments:
db_file : the database filename
field : the field name to sort by
numeric : if True sort field numerically;
if False sort field alphabetically
descend : if True sort in descending order;
if False sort in ascending order
out_file : output file name. Automatically generated from the input file if None.
out_args : common output argument dictionary from parseCommonArgs
Returns:
str : output file name
"""
log = OrderedDict()
log['START'] = 'ParseDb'
log['COMMAND'] = 'sort'
log['FILE'] = os.path.basename(db_file)
log['FIELD'] = field
log['NUMERIC'] = numeric
printLog(log)
# Open input
db_handle = open(db_file, 'rt')
db_iter = TSVReader(db_handle)
out_fields = db_iter.fields
__, __, out_args['out_type'] = splitName(db_file)
# Open output
if out_file is not None:
pass_handle = open(out_file, 'w')
else:
pass_handle = getOutputHandle(db_file,
out_label='parse-sort',
out_dir=out_args['out_dir'],
out_name=out_args['out_name'],
out_type=out_args['out_type'])
pass_writer = TSVWriter(pass_handle, out_fields)
# Store all records in a dictionary
start_time = time()
printMessage("Indexing: Running", start_time=start_time)
db_dict = {i: r for i, r in enumerate(db_iter)}
result_count = len(db_dict)
# Sort db_dict by field values
tag_dict = {k: v[field] for k, v in db_dict.items()}
if numeric: tag_dict = {k: float(v or 0) for k, v in tag_dict.items()}
sorted_keys = sorted(tag_dict, key=tag_dict.get, reverse=descend)
printMessage("Indexing: Done", start_time=start_time, end=True)
# Iterate over records
start_time = time()
rec_count = 0
for key in sorted_keys:
# Print progress for previous iteration
printProgress(rec_count, result_count, 0.05, start_time=start_time)
rec_count += 1
# Write records
pass_writer.writeDict(db_dict[key])
# Print counts
printProgress(rec_count, result_count, 0.05, start_time=start_time)
log = OrderedDict()
log['OUTPUT'] = os.path.basename(pass_handle.name)
log['RECORDS'] = rec_count
log['END'] = 'ParseDb'
printLog(log)
# Close file handles
pass_handle.close()
db_handle.close()
return pass_handle.name
def selectDbFile(db_file,
fields,
values,
logic='any',
regex=False,
out_file=None,
out_args=default_out_args):
"""
Selects records from a database file
Arguments:
db_file : the database file name
fields : a list of fields to check for selection criteria
values : a list of values defining selection targets
logic : one of 'any' or 'all' defining whether one or all fields must have a match.
regex : if False do exact full string matches; if True allow partial regex matches.
out_file : output file name. Automatically generated from the input file if None.
out_args : common output argument dictionary from parseCommonArgs
Returns:
str : output file name.
"""
# Define string match function
if regex:
def _match_func(x, patterns):
return any([re.search(p, x) for p in patterns])
else:
def _match_func(x, patterns):
return x in patterns
# Define logic function
if logic == 'any':
_logic_func = any
elif logic == 'all':
_logic_func = all
# Print console log
log = OrderedDict()
log['START'] = 'ParseDb'
log['COMMAND'] = 'select'
log['FILE'] = os.path.basename(db_file)
log['FIELDS'] = ','.join(fields)
log['VALUES'] = ','.join(values)
log['REGEX'] = regex
printLog(log)
# Open input
db_handle = open(db_file, 'rt')
db_iter = TSVReader(db_handle)
out_fields = db_iter.fields
__, __, out_args['out_type'] = splitName(db_file)
# Open output
if out_file is not None:
pass_handle = open(out_file, 'w')
else:
pass_handle = getOutputHandle(db_file,
out_label='parse-select',
out_dir=out_args['out_dir'],
out_name=out_args['out_name'],
out_type=out_args['out_type'])
pass_writer = TSVWriter(pass_handle, out_fields)
# Count records
result_count = countDbFile(db_file)
# Iterate over records
start_time = time()
rec_count, pass_count, fail_count = 0, 0, 0
for rec in db_iter:
# Print progress for previous iteration
printProgress(rec_count, result_count, 0.05, start_time=start_time)
rec_count += 1
# Check for selection values in all fields
select = _logic_func(
[_match_func(rec.get(f, False), values) for f in fields])
# Write sequences
if select:
pass_count += 1
pass_writer.writeDict(rec)
else:
fail_count += 1
# Print counts
printProgress(rec_count, result_count, 0.05, start_time=start_time)
log = OrderedDict()
log['OUTPUT'] = os.path.basename(pass_handle.name)
log['RECORDS'] = rec_count
log['SELECTED'] = pass_count
log['DISCARDED'] = fail_count
log['END'] = 'ParseDb'
printLog(log)
# Close file handles
pass_handle.close()
db_handle.close()
return pass_handle.name
def renameDbFile(db_file,
fields,
names,
out_file=None,
out_args=default_out_args):
"""
Renames fields in a database file
Arguments:
db_file : the database file name.
fields : a list of fields to rename.
values : a list of new names for fields.
out_file : output file name. Automatically generated from the input file if None.
out_args : common output argument dictionary from parseCommonArgs.
Returns:
str : output file name.
"""
log = OrderedDict()
log['START'] = 'ParseDb'
log['COMMAND'] = 'rename'
log['FILE'] = os.path.basename(db_file)
log['FIELDS'] = ','.join(fields)
log['NAMES'] = ','.join(names)
printLog(log)
# Open file handles
db_handle = open(db_file, 'rt')
db_iter = TSVReader(db_handle)
__, __, out_args['out_type'] = splitName(db_file)
# Get header and rename fields
out_fields = list(db_iter.fields)
for f, n in zip(fields, names):
i = out_fields.index(f)
out_fields[i] = n
# Open writer
if out_file is not None:
pass_handle = open(out_file, 'w')
else:
pass_handle = getOutputHandle(db_file,
out_label='parse-rename',
out_dir=out_args['out_dir'],
out_name=out_args['out_name'],
out_type=out_args['out_type'])
pass_writer = TSVWriter(pass_handle, out_fields)
# Count records
result_count = countDbFile(db_file)
# Iterate over records
start_time = time()
rec_count = 0
for rec in db_iter:
# Print progress for previous iteration
printProgress(rec_count, result_count, 0.05, start_time=start_time)
rec_count += 1
# TODO: repeating renaming is unnecessary.
# Rename fields
for f, n in zip(fields, names):
rec[n] = rec.pop(f)
# Write
pass_writer.writeDict(rec)
# Print counts
printProgress(rec_count, result_count, 0.05, start_time=start_time)
log = OrderedDict()
log['OUTPUT'] = os.path.basename(pass_handle.name)
log['RECORDS'] = rec_count
log['END'] = 'ParseDb'
printLog(log)
# Close file handles
pass_handle.close()
db_handle.close()
return pass_handle.name
def splitDbFile(db_file, field, num_split=None, out_args=default_out_args):
"""
Divides a tab-delimited database file into segments by description tags
Arguments:
db_file : filename of the tab-delimited database file to split
field : the field name by which to split db_file
num_split : the numerical threshold by which to group sequences;
if None treat field as textual
out_args : common output argument dictionary from parseCommonArgs
Returns:
list : a list of output file names.
"""
log = OrderedDict()
log['START'] = 'ParseDb'
log['COMMAND'] = 'split'
log['FILE'] = os.path.basename(db_file)
log['FIELD'] = field
log['NUM_SPLIT'] = num_split
printLog(log)
# Open input
db_handle = open(db_file, 'rt')
db_iter = TSVReader(db_handle)
out_fields = db_iter.fields
__, __, out_args['out_type'] = splitName(db_file)
# Determine total numbers of records
rec_count = countDbFile(db_file)
start_time = time()
count = 0
# Sort records into files based on textual field
if num_split is None:
# Create set of unique field tags
with open(db_file, 'rt') as tmp_handle:
tmp_iter = TSVReader(tmp_handle)
tag_list = list(set([row[field] for row in tmp_iter]))
# Forbidden characters in filename and replacements
no_good = {
'\/': 'f',
'\\': 'b',
'?': 'q',
'\%': 'p',
'*': 's',
':': 'c',
'\|': 'pi',
'\"': 'dq',
'\'': 'sq',
'<': 'gt',
'>': 'lt',
' ': '_'
}
# Replace forbidden characters in tag_list
tag_dict = {}
for tag in tag_list:
for c, r in no_good.items():
tag_dict[tag] = (tag_dict.get(tag, tag).replace(c,r) \
if c in tag else tag_dict.get(tag, tag))
# Create output handles
handles_dict = {
tag: getOutputHandle(db_file,
out_label='%s-%s' % (field, label),
out_name=out_args['out_name'],
out_dir=out_args['out_dir'],
out_type=out_args['out_type'])
for tag, label in tag_dict.items()
}
# Create Db writer instances
writers_dict = {
tag: TSVWriter(handles_dict[tag], fields=out_fields)
for tag in tag_dict
}
# Iterate over records
for row in db_iter:
printProgress(count, rec_count, 0.05, start_time=start_time)
count += 1
# Write row to appropriate file
tag = row[field]
writers_dict[tag].writeDict(row)
# Sort records into files based on numeric num_split
else:
num_split = float(num_split)
# Create output handles
handles_dict = {
'under':
getOutputHandle(db_file,
out_label='under-%.1f' % num_split,
out_name=out_args['out_name'],
out_dir=out_args['out_dir'],
out_type=out_args['out_type']),
'atleast':
getOutputHandle(db_file,
out_label='atleast-%.1f' % num_split,
out_name=out_args['out_name'],
out_dir=out_args['out_dir'],
out_type=out_args['out_type'])
}
# Create Db writer instances
writers_dict = {
'under': TSVWriter(handles_dict['under'], fields=out_fields),
'atleast': TSVWriter(handles_dict['atleast'], fields=out_fields)
}
# Iterate over records
for row in db_iter:
printProgress(count, rec_count, 0.05, start_time=start_time)
count += 1
tag = row[field]
tag = 'under' if float(tag) < num_split else 'atleast'
writers_dict[tag].writeDict(row)
# Write log
printProgress(count, rec_count, 0.05, start_time=start_time)
log = OrderedDict()
for i, k in enumerate(handles_dict):
log['OUTPUT%i' % (i + 1)] = os.path.basename(handles_dict[k].name)
log['RECORDS'] = rec_count
log['PARTS'] = len(handles_dict)
log['END'] = 'ParseDb'
printLog(log)
# Close output file handles
db_handle.close()
for t in handles_dict:
handles_dict[t].close()
return [handles_dict[t].name for t in handles_dict]
def dropDbFile(db_file, fields, out_file=None, out_args=default_out_args):
"""
Deletes entire fields from a database file
Arguments:
db_file : the database file name.
fields : a list of fields to drop.
out_file : output file name. Automatically generated from the input file if None.
out_args : common output argument dictionary from parseCommonArgs
Returns:
str : output file name.
"""
log = OrderedDict()
log['START'] = 'ParseDb'
log['COMMAND'] = 'add'
log['FILE'] = os.path.basename(db_file)
log['FIELDS'] = ','.join(fields)
printLog(log)
# Open input
db_handle = open(db_file, 'rt')
db_iter = TSVReader(db_handle)
__, __, out_args['out_type'] = splitName(db_file)
# Exclude dropped field from output
out_fields = [f for f in db_iter.fields if f not in fields]
# Open output
if out_file is not None:
pass_handle = open(out_file, 'w')
else:
pass_handle = getOutputHandle(db_file,
out_label='parse-drop',
out_dir=out_args['out_dir'],
out_name=out_args['out_name'],
out_type=out_args['out_type'])
pass_writer = TSVWriter(pass_handle, out_fields)
# Count records
result_count = countDbFile(db_file)
# Iterate over records
start_time = time()
rec_count = 0
for rec in db_iter:
# Print progress for previous iteration
printProgress(rec_count, result_count, 0.05, start_time=start_time)
rec_count += 1
# Write row
pass_writer.writeDict(rec)
# Print counts
printProgress(rec_count, result_count, 0.05, start_time=start_time)
log = OrderedDict()
log['OUTPUT'] = os.path.basename(pass_handle.name)
log['RECORDS'] = rec_count
log['END'] = 'ParseDb'
printLog(log)
# Close file handles
pass_handle.close()
return pass_handle.name
def indexDbFile(db_file,
field=default_index_field,
out_file=None,
out_args=default_out_args):
"""
Adds an index column to a database file
Arguments:
db_file : the database file name.
field : the name of the index field to add.
out_file : output file name. Automatically generated from the input file if None.
out_args : common output argument dictionary from parseCommonArgs.
Returns:
str : output file name.
"""
log = OrderedDict()
log['START'] = 'ParseDb'
log['COMMAND'] = 'index'
log['FILE'] = os.path.basename(db_file)
log['FIELD'] = field
printLog(log)
# Open input
db_handle = open(db_file, 'rt')
db_iter = TSVReader(db_handle)
__, __, out_args['out_type'] = splitName(db_file)
# Append index field
out_fields = list(db_iter.fields)
out_fields.append(field)
# Open output
if out_file is not None:
pass_handle = open(out_file, 'w')
else:
pass_handle = getOutputHandle(db_file,
out_label='parse-index',
out_dir=out_args['out_dir'],
out_name=out_args['out_name'],
out_type=out_args['out_type'])
pass_writer = TSVWriter(pass_handle, out_fields)
# Count records
result_count = countDbFile(db_file)
# Iterate over records
start_time = time()
rec_count = 0
for rec in db_iter:
# Print progress for previous iteration
printProgress(rec_count, result_count, 0.05, start_time=start_time)
rec_count += 1
# Add count and write updated row
rec.update({field: rec_count})
pass_writer.writeDict(rec)
# Print counts
printProgress(rec_count, result_count, 0.05, start_time=start_time)
log = OrderedDict()
log['OUTPUT'] = os.path.basename(pass_handle.name)
log['RECORDS'] = rec_count
log['END'] = 'ParseDb'
printLog(log)
# Close file handles
pass_handle.close()
db_handle.close()
return pass_handle.name
def addDbFile(db_file,
fields,
values,
out_file=None,
out_args=default_out_args):
"""
Adds field and value pairs to a database file
Arguments:
db_file : the database file name.
fields : a list of fields to add.
values : a list of values to assign to all rows of each field.
out_file : output file name. Automatically generated from the input file if None.
out_args : common output argument dictionary from parseCommonArgs.
Returns:
str : output file name.
"""
log = OrderedDict()
log['START'] = 'ParseDb'
log['COMMAND'] = 'add'
log['FILE'] = os.path.basename(db_file)
log['FIELDS'] = ','.join(fields)
log['VALUES'] = ','.join(values)
printLog(log)
# Open inut
db_handle = open(db_file, 'rt')
db_iter = TSVReader(db_handle)
__, __, out_args['out_type'] = splitName(db_file)
# Add fields
out_fields = list(db_iter.fields)
out_fields.extend(fields)
# Open output
if out_file is not None:
pass_handle = open(out_file, 'w')
else:
pass_handle = getOutputHandle(db_file,
out_label='parse-add',
out_dir=out_args['out_dir'],
out_name=out_args['out_name'],
out_type=out_args['out_type'])
pass_writer = TSVWriter(pass_handle, out_fields)
# Count records
result_count = countDbFile(db_file)
# Define fields and values to append
add_dict = {
k: v
for k, v in zip(fields, values) if k not in db_iter.fields
}
# Iterate over records
start_time = time()
rec_count = 0
for rec in db_iter:
# Print progress for previous iteration
printProgress(rec_count, result_count, 0.05, start_time=start_time)
rec_count += 1
# Write updated row
rec.update(add_dict)
pass_writer.writeDict(rec)
# Print counts
printProgress(rec_count, result_count, 0.05, start_time=start_time)
log = OrderedDict()
log['OUTPUT'] = os.path.basename(pass_handle.name)
log['RECORDS'] = rec_count
log['END'] = 'ParseDb'
printLog(log)
# Close file handles
pass_handle.close()
db_handle.close()
return pass_handle.name
def convertToGenbank(db_file, inference=None, db_xref=None, molecule=default_molecule,
product=default_product, features=None, c_field=None, label=None,
count_field=None, index_field=None, allow_stop=False,
asis_id=False, asis_calls=False, allele_delim=default_allele_delim,
build_asn=False, asn_template=None, tbl2asn_exec=default_tbl2asn_exec,
format=default_format, out_file=None,
out_args=default_out_args):
"""
Builds GenBank submission fasta and table files
Arguments:
db_file : the database file name.
inference : reference alignment tool.
db_xref : reference database link.
molecule : source molecule (eg, "mRNA", "genomic DNA")
product : Product (protein) name.
features : dictionary of sample features (BioSample attributes) to add to the description of each record.
c_field : column containing the C region gene call.
label : a string to use as a label for the ID. if None do not add a field label.
count_field : field name to populate the AIRR_READ_COUNT note.
index_field : field name to populate the AIRR_CELL_INDEX note.
allow_stop : if True retain records with junctions having stop codons.
asis_id : if True use the original sequence ID for the output IDs.
asis_calls : if True do not parse gene calls for IMGT nomenclature.
allele_delim : delimiter separating the gene name from the allele number when asis_calls=True.
build_asn : if True run tbl2asn on the generated .tbl and .fsa files.
asn_template : template file (.sbt) to pass to tbl2asn.
tbl2asn_exec : name of or path to the tbl2asn executable.
format : input and output format.
out_file : output file name without extension. Automatically generated from the input file if None.
out_args : common output argument dictionary from parseCommonArgs.
Returns:
tuple : the output (feature table, fasta) file names.
"""
log = OrderedDict()
log['START'] = 'ConvertDb'
log['COMMAND'] = 'genbank'
log['FILE'] = os.path.basename(db_file)
printLog(log)
# Define format operators
try:
reader, __, schema = getFormatOperators(format)
except ValueError:
printError('Invalid format %s.' % format)
# Open input
db_handle = open(db_file, 'rt')
db_iter = reader(db_handle)
# Check for required columns
try:
required = ['sequence_input',
'v_call', 'd_call', 'j_call',
'v_seq_start', 'd_seq_start', 'j_seq_start']
checkFields(required, db_iter.fields, schema=schema)
except LookupError as e:
printError(e)
# Open output
if out_file is not None:
out_name, __ = os.path.splitext(out_file)
fsa_handle = open('%s.fsa' % out_name, 'w')
tbl_handle = open('%s.tbl' % out_name, 'w')
else:
fsa_handle = getOutputHandle(db_file, out_label='genbank', out_dir=out_args['out_dir'],
out_name=out_args['out_name'], out_type='fsa')
tbl_handle = getOutputHandle(db_file, out_label='genbank', out_dir=out_args['out_dir'],
out_name=out_args['out_name'], out_type='tbl')
# Count records
result_count = countDbFile(db_file)
# Define writer
writer = csv.writer(tbl_handle, delimiter='\t', quoting=csv.QUOTE_NONE)
# Iterate over records
start_time = time()
rec_count, pass_count, fail_count = 0, 0, 0
for rec in db_iter:
# Print progress for previous iteration
printProgress(rec_count, result_count, 0.05, start_time=start_time)
rec_count += 1
# Extract table dictionary
name = None if asis_id else rec_count
seq = makeGenbankSequence(rec, name=name, label=label, count_field=count_field, index_field=index_field,
molecule=molecule, features=features)
tbl = makeGenbankFeatures(rec, start=seq['start'], end=seq['end'], product=product,
db_xref=db_xref, inference=inference, c_field=c_field,
allow_stop=allow_stop, asis_calls=asis_calls, allele_delim=allele_delim)
if tbl is not None:
pass_count +=1
# Write table
writer.writerow(['>Features', seq['record'].id])
for feature, qualifiers in tbl.items():
writer.writerow(feature)
if qualifiers:
for x in qualifiers:
writer.writerow(list(chain(['', '', ''], x)))
# Write sequence
SeqIO.write(seq['record'], fsa_handle, 'fasta')
else:
fail_count += 1
# Final progress bar
printProgress(rec_count, result_count, 0.05, start_time=start_time)
# Run tbl2asn
if build_asn:
start_time = time()
printMessage('Running tbl2asn', start_time=start_time, width=25)
result = runASN(fsa_handle.name, template=asn_template, exec=tbl2asn_exec)
printMessage('Done', start_time=start_time, end=True, width=25)
# Print ending console log
log = OrderedDict()
log['OUTPUT_TBL'] = os.path.basename(tbl_handle.name)
log['OUTPUT_FSA'] = os.path.basename(fsa_handle.name)
log['RECORDS'] = rec_count
log['PASS'] = pass_count
log['FAIL'] = fail_count
log['END'] = 'ConvertDb'
printLog(log)
# Close file handles
tbl_handle.close()
fsa_handle.close()
db_handle.close()
return (tbl_handle.name, fsa_handle.name)
def convertToFasta(db_file, id_field=default_id_field, seq_field=default_seq_field,
meta_fields=None, out_file=None, out_args=default_out_args):
"""
Builds fasta files from database records
Arguments:
db_file : the database file name.
id_field : the field containing identifiers.
seq_field : the field containing sequences.
meta_fields : a list of fields to add to sequence annotations.
out_file : output file name. Automatically generated from the input file if None.
out_args : common output argument dictionary from parseCommonArgs.
Returns:
str : output file name.
"""
log = OrderedDict()
log['START'] = 'ConvertDb'
log['COMMAND'] = 'fasta'
log['FILE'] = os.path.basename(db_file)
log['ID_FIELD'] = id_field
log['SEQ_FIELD'] = seq_field
if meta_fields is not None: log['META_FIELDS'] = ','.join(meta_fields)
printLog(log)
# Open input
out_type = 'fasta'
db_handle = open(db_file, 'rt')
db_iter = TSVReader(db_handle)
result_count = countDbFile(db_file)
# Open output
if out_file is not None:
pass_handle = open(out_file, 'w')
else:
pass_handle = getOutputHandle(db_file, out_label='sequences', out_dir=out_args['out_dir'],
out_name=out_args['out_name'], out_type=out_type)
# Iterate over records
start_time = time()
rec_count, pass_count, fail_count = 0, 0, 0
for rec in db_iter:
# Print progress for previous iteration
printProgress(rec_count, result_count, 0.05, start_time=start_time)
rec_count += 1
# Get SeqRecord
seq = buildSeqRecord(rec, id_field, seq_field, meta_fields)
# Write sequences
if seq is not None:
pass_count += 1
SeqIO.write(seq, pass_handle, out_type)
else:
fail_count += 1
# Print counts
printProgress(rec_count, result_count, 0.05, start_time=start_time)
log = OrderedDict()
log['OUTPUT'] = os.path.basename(pass_handle.name)
log['RECORDS'] = rec_count
log['PASS'] = pass_count
log['FAIL'] = fail_count
log['END'] = 'ConvertDb'
printLog(log)
# Close file handles
pass_handle.close()
db_handle.close()
return pass_handle.name
def convertToBaseline(db_file, id_field=default_id_field, seq_field=default_seq_field,
germ_field=default_germ_field, cluster_field=None,
meta_fields=None, out_file=None, out_args=default_out_args):
"""
Builds fasta files from database records
Arguments:
db_file : the database file name.
id_field : the field containing identifiers.
seq_field : the field containing sample sequences.
germ_field : the field containing germline sequences.
cluster_field : the field containing clonal groupings;
if None write the germline for each record.
meta_fields : a list of fields to add to sequence annotations.
out_file : output file name. Automatically generated from the input file if None.
out_args : common output argument dictionary from parseCommonArgs.
Returns:
str : output file name
"""
log = OrderedDict()
log['START'] = 'ConvertDb'
log['COMMAND'] = 'fasta'
log['FILE'] = os.path.basename(db_file)
log['ID_FIELD'] = id_field
log['SEQ_FIELD'] = seq_field
log['GERM_FIELD'] = germ_field
log['CLUSTER_FIELD'] = cluster_field
if meta_fields is not None: log['META_FIELDS'] = ','.join(meta_fields)
printLog(log)
# Open input
db_handle = open(db_file, 'rt')
db_iter = TSVReader(db_handle)
result_count = countDbFile(db_file)
# Open output
if out_file is not None:
pass_handle = open(out_file, 'w')
else:
pass_handle = getOutputHandle(db_file, out_label='sequences', out_dir=out_args['out_dir'],
out_name=out_args['out_name'], out_type='clip')
# Iterate over records
start_time = time()
rec_count, germ_count, pass_count, fail_count = 0, 0, 0, 0
cluster_last = None
for rec in db_iter:
# Print progress for previous iteration
printProgress(rec_count, result_count, 0.05, start_time=start_time)
rec_count += 1
# Update cluster ID
cluster = rec.get(cluster_field, None)
# Get germline SeqRecord when needed
if cluster_field is None:
germ = buildSeqRecord(rec, id_field, germ_field, meta_fields)
germ.id = '>' + germ.id
elif cluster != cluster_last:
germ = buildSeqRecord(rec, cluster_field, germ_field)
germ.id = '>' + germ.id
else:
germ = None
# Get read SeqRecord
seq = buildSeqRecord(rec, id_field, seq_field, meta_fields)
# Write germline
if germ is not None:
germ_count += 1
SeqIO.write(germ, pass_handle, 'fasta')
# Write sequences
if seq is not None:
pass_count += 1
SeqIO.write(seq, pass_handle, 'fasta')
else:
fail_count += 1
# Set last cluster ID
cluster_last = cluster
# Print counts
printProgress(rec_count, result_count, 0.05, start_time=start_time)
log = OrderedDict()
log['OUTPUT'] = os.path.basename(pass_handle.name)
log['RECORDS'] = rec_count
log['GERMLINES'] = germ_count
log['PASS'] = pass_count
log['FAIL'] = fail_count
log['END'] = 'ConvertDb'
printLog(log)
# Close file handles
pass_handle.close()
db_handle.close()
return pass_handle.name
def convertToChangeo(db_file, out_file=None, out_args=default_out_args):
"""
Converts an AIRR formatted file into an Change-O formatted file
Arguments:
db_file: the database file name.
out_file : output file name. Automatically generated from the input file if None.
out_args : common output argument dictionary from parseCommonArgs.
Returns:
str : output file name.
"""
log = OrderedDict()
log['START'] = 'ConvertDb'
log['COMMAND'] = 'changeo'
log['FILE'] = os.path.basename(db_file)
printLog(log)
# Open input
db_handle = open(db_file, 'rt')
db_iter = AIRRReader(db_handle)
# Set output fields replacing length with end fields
in_fields = [AIRRSchema.toReceptor(f) for f in db_iter.fields]
out_fields = []
for f in in_fields:
out_fields.append(f)
if f in ReceptorData.end_fields and ReceptorData.end_fields[f][0] in in_fields:
out_fields.append(ReceptorData.end_fields[f][1])
out_fields = list(OrderedDict.fromkeys(out_fields))
out_fields = [ChangeoSchema.fromReceptor(f) for f in out_fields]
# Open output writer
if out_file is not None:
pass_handle = open(out_file, 'w')
else:
pass_handle = getOutputHandle(db_file, out_label='changeo', out_dir=out_args['out_dir'],
out_name=out_args['out_name'], out_type=ChangeoSchema.out_type)
pass_writer = ChangeoWriter(pass_handle, fields=out_fields)
# Count records
result_count = countDbFile(db_file)
# Iterate over records
start_time = time()
rec_count = 0
for rec in db_iter:
# Print progress for previous iteration
printProgress(rec_count, result_count, 0.05, start_time=start_time)
rec_count += 1
# Write records
pass_writer.writeReceptor(rec)
# Print counts
printProgress(rec_count, result_count, 0.05, start_time=start_time)
log = OrderedDict()
log['OUTPUT'] = os.path.basename(pass_handle.name)
log['RECORDS'] = rec_count
log['END'] = 'ConvertDb'
printLog(log)
# Close file handles
pass_handle.close()
db_handle.close()
return pass_handle.name
def insertGaps(db_file, references=None, format=default_format,
out_file=None, out_args=default_out_args):
"""
Inserts IMGT numbering into V fields
Arguments:
db_file : the database file name.
references : folder with germline repertoire files. If None, do not updated alignment columns wtih IMGT gaps.
format : input format.
out_file : output file name. Automatically generated from the input file if None.
out_args : common output argument dictionary from parseCommonArgs.
Returns:
str : output file name
"""
log = OrderedDict()
log['START'] = 'ConvertDb'
log['COMMAND'] = 'imgt'
log['FILE'] = os.path.basename(db_file)
printLog(log)
# Define format operators
try:
reader, writer, schema = getFormatOperators(format)
except ValueError:
printError('Invalid format %s.' % format)
# Open input
db_handle = open(db_file, 'rt')
db_iter = reader(db_handle)
# Check for required columns
try:
required = ['sequence_imgt', 'v_germ_start_imgt']
checkFields(required, db_iter.fields, schema=schema)
except LookupError as e:
printError(e)
# Load references
reference_dict = readGermlines(references)
# Check for IMGT-gaps in germlines
if all('...' not in x for x in reference_dict.values()):
printWarning('Germline reference sequences do not appear to contain IMGT-numbering spacers. Results may be incorrect.')
# Open output writer
if out_file is not None:
pass_handle = open(out_file, 'w')
else:
pass_handle = getOutputHandle(db_file, out_label='gap', out_dir=out_args['out_dir'],
out_name=out_args['out_name'], out_type=schema.out_type)
pass_writer = writer(pass_handle, fields=db_iter.fields)
# Count records
result_count = countDbFile(db_file)
# Iterate over records
start_time = time()
rec_count = pass_count = 0
for rec in db_iter:
# Print progress for previous iteration
printProgress(rec_count, result_count, 0.05, start_time=start_time)
rec_count += 1
# Update IMGT fields
imgt_dict = correctIMGTFields(rec, reference_dict)
# Write records
if imgt_dict is not None:
pass_count += 1
rec.setDict(imgt_dict, parse=False)
pass_writer.writeReceptor(rec)
# Print counts
printProgress(rec_count, result_count, 0.05, start_time=start_time)
log = OrderedDict()
log['OUTPUT'] = os.path.basename(pass_handle.name)
log['RECORDS'] = rec_count
log['PASS'] = pass_count
log['FAIL'] = rec_count - pass_count
log['END'] = 'ConvertDb'
printLog(log)
# Close file handles
pass_handle.close()
db_handle.close()
return pass_handle.name
def createGermlines(db_file, references, seq_field=default_seq_field, v_field=default_v_field,
d_field=default_d_field, j_field=default_j_field,
cloned=False, clone_field=default_clone_field, germ_types=default_germ_types,
format=default_format, out_file=None, out_args=default_out_args):
"""
Write germline sequences to tab-delimited database file
Arguments:
db_file : input tab-delimited database file.
references : folders and/or files containing germline repertoire data in FASTA format.
seq_field : field in which to look for sequence.
v_field : field in which to look for V call.
d_field : field in which to look for D call.
j_field : field in which to look for J call.
cloned : if True build germlines by clone, otherwise build individual germlines.
clone_field : field containing clone identifiers; ignored if cloned=False.
germ_types : list of germline sequence types to be output from the set of 'full', 'dmask', 'vonly', 'regions'
format : input and output format.
out_file : output file name. Automatically generated from the input file if None.
out_args : arguments for output preferences.
Returns:
dict: names of the 'pass' and 'fail' output files.
"""
# Print parameter info
log = OrderedDict()
log['START'] = 'CreateGermlines'
log['FILE'] = os.path.basename(db_file)
log['GERM_TYPES'] = ','.join(germ_types)
log['SEQ_FIELD'] = seq_field
log['V_FIELD'] = v_field
log['D_FIELD'] = d_field
log['J_FIELD'] = j_field
log['CLONED'] = cloned
if cloned: log['CLONE_FIELD'] = clone_field
printLog(log)
# Define format operators
try:
reader, writer, schema = getFormatOperators(format)
except ValueError:
printError('Invalid format %s' % format)
out_args['out_type'] = schema.out_type
# TODO: this won't work for AIRR necessarily
# Define output germline fields
germline_fields = OrderedDict()
seq_type = seq_field.split('_')[-1]
if 'full' in germ_types: germline_fields['full'] = 'germline_' + seq_type
if 'dmask' in germ_types: germline_fields['dmask'] = 'germline_' + seq_type + '_d_mask'
if 'vonly' in germ_types: germline_fields['vonly'] = 'germline_' + seq_type + '_v_region'
if 'regions' in germ_types: germline_fields['regions'] = 'germline_regions'
if cloned:
germline_fields['v'] = 'germline_v_call'
germline_fields['d'] = 'germline_d_call'
germline_fields['j'] = 'germline_j_call'
out_fields = getDbFields(db_file,
add=[schema.fromReceptor(f) for f in germline_fields.values()],
reader=reader)
# Get repertoire and open Db reader
reference_dict = readGermlines(references)
db_handle = open(db_file, 'rt')
db_iter = reader(db_handle)
# Check for required columns
try:
required = ['v_germ_start_imgt', 'd_germ_start', 'j_germ_start',
'np1_length', 'np2_length']
checkFields(required, db_iter.fields, schema=schema)
except LookupError as e:
printError(e)
# Check for IMGT-gaps in germlines
if all('...' not in x for x in reference_dict.values()):
printWarning('Germline reference sequences do not appear to contain IMGT-numbering spacers. Results may be incorrect.')
# Count input
total_count = countDbFile(db_file)
# Check for existence of fields
for f in [v_field, d_field, j_field, seq_field]:
if f not in db_iter.fields:
printError('%s field does not exist in input database file.' % f)
# Translate to Receptor attribute names
v_field = schema.toReceptor(v_field)
d_field = schema.toReceptor(d_field)
j_field = schema.toReceptor(j_field)
seq_field = schema.toReceptor(seq_field)
clone_field = schema.toReceptor(clone_field)
# Define Receptor iterator
if cloned:
start_time = time()
printMessage('Sorting by clone', start_time=start_time, width=20)
sorted_records = sorted(db_iter, key=lambda x: x.getField(clone_field))
printMessage('Done', start_time=start_time, end=True, width=20)
receptor_iter = groupby(sorted_records, lambda x: x.getField(clone_field))
else:
receptor_iter = ((x.sequence_id, [x]) for x in db_iter)
# Define log handle
if out_args['log_file'] is None:
log_handle = None
else:
log_handle = open(out_args['log_file'], 'w')
# Initialize handles, writers and counters
pass_handle, pass_writer = None, None
fail_handle, fail_writer = None, None
rec_count, pass_count, fail_count = 0, 0, 0
start_time = time()
# Iterate over rows
for key, records in receptor_iter:
# Print progress
printProgress(rec_count, total_count, 0.05, start_time=start_time)
# Define iteration variables
records = list(records)
rec_log = OrderedDict([('ID', key)])
rec_count += len(records)
# Build germline for records
if len(records) == 1:
germ_log, germlines, genes = buildGermline(records[0], reference_dict, seq_field=seq_field, v_field=v_field,
d_field=d_field, j_field=j_field)
else:
germ_log, germlines, genes = buildClonalGermline(records, reference_dict, seq_field=seq_field, v_field=v_field,
d_field=d_field, j_field=j_field)
rec_log.update(germ_log)
# Write row to pass or fail file
if germlines is not None:
pass_count += len(records)
# Add germlines to Receptor record
annotations = {}
if 'full' in germ_types: annotations[germline_fields['full']] = germlines['full']
if 'dmask' in germ_types: annotations[germline_fields['dmask']] = germlines['dmask']
if 'vonly' in germ_types: annotations[germline_fields['vonly']] = germlines['vonly']
if 'regions' in germ_types: annotations[germline_fields['regions']] = germlines['regions']
if cloned:
annotations[germline_fields['v']] = genes['v']
annotations[germline_fields['d']] = genes['d']
annotations[germline_fields['j']] = genes['j']
# Write records
try:
for r in records:
r.setDict(annotations)
pass_writer.writeReceptor(r)
except AttributeError:
# Create output file handle and writer
if out_file is not None:
pass_handle = open(out_file, 'w')
else:
pass_handle = getOutputHandle(db_file,
out_label='germ-pass',
out_dir=out_args['out_dir'],
out_name=out_args['out_name'],
out_type=out_args['out_type'])
pass_writer = writer(pass_handle, fields=out_fields)
for r in records:
r.setDict(annotations)
pass_writer.writeReceptor(r)
else:
fail_count += len(records)
if out_args['failed']:
try:
fail_writer.writeReceptor(records)
except AttributeError:
fail_handle = getOutputHandle(db_file,
out_label='germ-fail',
out_dir=out_args['out_dir'],
out_name=out_args['out_name'],
out_type=out_args['out_type'])
fail_writer = writer(fail_handle, fields=out_fields)
fail_writer.writeReceptor(records)
# Write log
printLog(rec_log, handle=log_handle)
# Print log
printProgress(rec_count, total_count, 0.05, start_time=start_time)
log = OrderedDict()
log['OUTPUT'] = os.path.basename(pass_handle.name) if pass_handle is not None else None
log['RECORDS'] = rec_count
log['PASS'] = pass_count
log['FAIL'] = fail_count
log['END'] = 'CreateGermlines'
printLog(log)
# Close file handles
db_handle.close()
output = {'pass': None, 'fail': None}
if pass_handle is not None:
output['pass'] = pass_handle.name
pass_handle.close()
if fail_handle is not None:
output['fail'] = fail_handle.name
fail_handle.close()
if log_handle is not None:
log_handle.close()
return output