def load_BIOM(table, informat='json', v=1):
"""
load a BIOM table from BIOM format. Default format is 'json'.
"""
from biom.table import Table
import json
import sys
informats = ['json','tsv']
if not informat in informats:
print "\nPlease specify a valid BIOM input format. Currently we support: '%s'.\n" %"', '".join(informats)
else:
if v:
print "\nSpecified BIOM input format '%s' - ok!" %(informat)
if informat == 'json':
with open(table) as data_file:
data = json.load(data_file)
t = Table.from_json(data)
elif informat == 'tsv':
tsv = open(in_tsv)
func = lambda x : x
t = Table.from_tsv(tsv, obs_mapping=None, sample_mapping=None, process_func=func)
tsv.close()
return t
def BIOM_tsv_to_R_transpose(in_tsv, out_csv):
"""
Parse a biom table in tsv format and transpose it for input into R
"""
from biom import Table
tsv = open(in_tsv)
#in_tsv = open('COI-trim30min100-merge-c3-id97-OTU-taxonomy.kraken.tsv')
func = lambda x : x
intable = Table.from_tsv(tsv,obs_mapping=None, sample_mapping=None, process_func=func)
outtable = intable.transpose()
out=open("transposed.tsv","w")
out.write(outtable.to_tsv(header_key=None, header_value=None))
out.close()
#refine
intable = open('transposed.tsv','r')
temp = intable.next()
out=''
for line in intable:
if line.startswith('#'):
if line.strip().endswith('taxomomy'):
print "Removing taxonomy"
line = ",".join(line.strip().split("\t")[:-1]).replace('#OTU ID','Sample').replace('\t',',')+'\n'
line = line.replace('#OTU ID','Sample').replace('\t',',')
out+=line
else:
line = line.replace('\t',',')
out+=line
outtable = open(out_csv,'w')
outtable.write(out)
outtable.close()
def convert_table_to_biom(table_f, sample_mapping, obs_mapping,
process_func, **kwargs):
"""Convert a contigency table to a biom table
sample_mapping : dict of {'sample_id':metadata} or None
obs_mapping : dict of {'obs_id':metadata} or None
process_func: a function to transform observation metadata
dtype : type of table data
"""
otu_table = Table.from_tsv(table_f, obs_mapping, sample_mapping,
process_func, **kwargs)
return otu_table.to_json(generatedby())
def convert_table_to_biom(table_f, sample_mapping, obs_mapping, process_func,
**kwargs):
"""Convert a contigency table to a biom table
sample_mapping : dict of {'sample_id':metadata} or None
obs_mapping : dict of {'obs_id':metadata} or None
process_func: a function to transform observation metadata
dtype : type of table data
"""
otu_table = Table.from_tsv(table_f, obs_mapping, sample_mapping,
process_func, **kwargs)
return otu_table.to_json(generatedby())
def parse_biom_table(fp, ids=None, axis='sample', input_is_dense=False):
r"""Parses the biom table stored in the filepath `fp`
Parameters
----------
fp : file like
File alike object storing the BIOM table
ids : iterable
The sample/observation ids of the samples/observations that we need
to retrieve from the biom table
axis : {'sample', 'observation'}, optional
The axis to subset on
input_is_dense : boolean
Indicates if the BIOM table is dense or sparse. Valid only for JSON
tables.
Returns
-------
Table
The BIOM table stored at fp
Raises
------
ValueError
If `samples` and `observations` are provided.
Notes
-----
Subsetting from the BIOM table is only supported in one axis
Examples
--------
Parse a hdf5 biom table
>>> from h5py import File # doctest: +SKIP
>>> from biom.parse import parse_biom_table
>>> f = File('rich_sparse_otu_table_hdf5.biom') # doctest: +SKIP
>>> t = parse_biom_table(f) # doctest: +SKIP
Parse a hdf5 biom table subsetting observations
>>> from h5py import File # doctest: +SKIP
>>> from biom.parse import parse_biom_table
>>> f = File('rich_sparse_otu_table_hdf5.biom') # doctest: +SKIP
>>> t = parse_biom_table(f, ids=["GG_OTU_1"],
... axis='observation') # doctest: +SKIP
"""
if axis not in ['observation', 'sample']:
UnknownAxisError(axis)
try:
return Table.from_hdf5(fp, ids=ids, axis=axis)
except ValueError:
pass
except RuntimeError:
pass
if hasattr(fp, 'read'):
old_pos = fp.tell()
# Read in characters until first non-whitespace
# If it is a {, then this is (most likely) JSON
c = fp.read(1)
while c.isspace():
c = fp.read(1)
if c == '{':
fp.seek(old_pos)
t = Table.from_json(json.load(fp, object_pairs_hook=OrderedDict),
input_is_dense=input_is_dense)
else:
fp.seek(old_pos)
t = Table.from_tsv(fp, None, None, lambda x: x)
elif isinstance(fp, list):
try:
t = Table.from_json(json.loads(''.join(fp),
object_pairs_hook=OrderedDict),
input_is_dense=input_is_dense)
except ValueError:
t = Table.from_tsv(fp, None, None, lambda x: x)
else:
t = Table.from_json(json.loads(fp, object_pairs_hook=OrderedDict),
input_is_dense=input_is_dense)
def subset_ids(data, id_, md):
return id_ in ids
def gt_zero(vals, id_, md):
return np.any(vals)
if ids is not None:
t.filter(subset_ids, axis=axis)
axis = 'observation' if axis == 'sample' else 'sample'
t.filter(gt_zero, axis=axis)
return t
def parse_biom_table(fp, ids=None, axis='sample', input_is_dense=False):
r"""Parses the biom table stored in the filepath `fp`
Parameters
----------
fp : file like
File alike object storing the BIOM table
ids : iterable
The sample/observation ids of the samples/observations that we need
to retrieve from the biom table
axis : {'sample', 'observation'}, optional
The axis to subset on
input_is_dense : boolean
Indicates if the BIOM table is dense or sparse. Valid only for JSON
tables.
Returns
-------
Table
The BIOM table stored at fp
Raises
------
ValueError
If `samples` and `observations` are provided.
Notes
-----
Subsetting from the BIOM table is only supported in one axis
Examples
--------
Parse a hdf5 biom table
>>> from h5py import File # doctest: +SKIP
>>> from biom.parse import parse_biom_table
>>> f = File('rich_sparse_otu_table_hdf5.biom') # doctest: +SKIP
>>> t = parse_biom_table(f) # doctest: +SKIP
Parse a hdf5 biom table subsetting observations
>>> from h5py import File # doctest: +SKIP
>>> from biom.parse import parse_biom_table
>>> f = File('rich_sparse_otu_table_hdf5.biom') # doctest: +SKIP
>>> t = parse_biom_table(f, ids=["GG_OTU_1"],
... axis='observation') # doctest: +SKIP
"""
if axis not in ['observation', 'sample']:
UnknownAxisError(axis)
try:
return Table.from_hdf5(fp, ids=ids, axis=axis)
except:
pass
if hasattr(fp, 'read'):
old_pos = fp.tell()
try:
t = Table.from_json(json.load(fp), input_is_dense=input_is_dense)
except ValueError:
fp.seek(old_pos)
t = Table.from_tsv(fp, None, None, lambda x: x)
elif isinstance(fp, list):
try:
t = Table.from_json(json.loads(''.join(fp)),
input_is_dense=input_is_dense)
except ValueError:
t = Table.from_tsv(fp, None, None, lambda x: x)
else:
t = Table.from_json(json.loads(fp), input_is_dense=input_is_dense)
if ids is not None:
f = lambda data, id_, md: id_ in ids
t.filter(f, axis=axis)
axis = 'observation' if axis == 'sample' else 'sample'
f = lambda vals, id_, md: np.any(vals)
t.filter(f, axis=axis)
return t
def parse_biom_table(fp, ids=None, axis='sample', input_is_dense=False):
r"""Parses the biom table stored in the filepath `fp`
Parameters
----------
fp : file like
File alike object storing the BIOM table
ids : iterable
The sample/observation ids of the samples/observations that we need
to retrieve from the biom table
axis : {'sample', 'observation'}, optional
The axis to subset on
input_is_dense : boolean
Indicates if the BIOM table is dense or sparse. Valid only for JSON
tables.
Returns
-------
Table
The BIOM table stored at fp
Raises
------
ValueError
If `samples` and `observations` are provided.
Notes
-----
Subsetting from the BIOM table is only supported in one axis
Examples
--------
Parse a hdf5 biom table
>>> from h5py import File # doctest: +SKIP
>>> from biom.parse import parse_biom_table
>>> f = File('rich_sparse_otu_table_hdf5.biom') # doctest: +SKIP
>>> t = parse_biom_table(f) # doctest: +SKIP
Parse a hdf5 biom table subsetting observations
>>> from h5py import File # doctest: +SKIP
>>> from biom.parse import parse_biom_table
>>> f = File('rich_sparse_otu_table_hdf5.biom') # doctest: +SKIP
>>> t = parse_biom_table(f, ids=["GG_OTU_1"],
... axis='observation') # doctest: +SKIP
"""
if axis not in ['observation', 'sample']:
UnknownAxisError(axis)
try:
return Table.from_hdf5(fp, ids=ids, axis=axis)
except:
pass
if hasattr(fp, 'read'):
old_pos = fp.tell()
try:
t = Table.from_json(json.load(fp), input_is_dense=input_is_dense)
except ValueError:
fp.seek(old_pos)
t = Table.from_tsv(fp, None, None, lambda x: x)
elif isinstance(fp, list):
try:
t = Table.from_json(json.loads(''.join(fp)),
input_is_dense=input_is_dense)
except ValueError:
t = Table.from_tsv(fp, None, None, lambda x: x)
else:
t = Table.from_json(json.loads(fp), input_is_dense=input_is_dense)
if ids is not None:
f = lambda data, id_, md: id_ in ids
t.filter(f, axis=axis)
axis = 'observation' if axis == 'sample' else 'sample'
f = lambda vals, id_, md: np.any(vals)
t.filter(f, axis=axis)
return t
def parse_biom_table(file_obj, ids=None, axis='sample', input_is_dense=False):
r"""Parses the biom table stored in `file_obj`
Parameters
----------
file_obj : file-like object, or list
file-like object storing the BIOM table (tab-delimited or JSON), or
a list of lines of the BIOM table in tab-delimited or JSON format
ids : iterable
The sample/observation ids of the samples/observations that we need
to retrieve from the biom table
axis : {'sample', 'observation'}, optional
The axis to subset on
input_is_dense : boolean
Indicates if the BIOM table is dense or sparse. Valid only for JSON
tables.
Returns
-------
Table
The BIOM table stored at file_obj
Raises
------
ValueError
If `samples` and `observations` are provided.
Notes
-----
Subsetting from the BIOM table is only supported in one axis
Examples
--------
Parse a hdf5 biom table
>>> from h5py import File # doctest: +SKIP
>>> from biom.parse import parse_biom_table
>>> f = File('rich_sparse_otu_table_hdf5.biom') # doctest: +SKIP
>>> t = parse_biom_table(f) # doctest: +SKIP
Parse a hdf5 biom table subsetting observations
>>> from h5py import File # doctest: +SKIP
>>> from biom.parse import parse_biom_table
>>> f = File('rich_sparse_otu_table_hdf5.biom') # doctest: +SKIP
>>> t = parse_biom_table(f, ids=["GG_OTU_1"],
... axis='observation') # doctest: +SKIP
"""
if axis not in ['observation', 'sample']:
UnknownAxisError(axis)
try:
return Table.from_hdf5(file_obj, ids=ids, axis=axis)
except ValueError:
pass
except RuntimeError:
pass
if hasattr(file_obj, 'read'):
old_pos = file_obj.tell()
# Read in characters until first non-whitespace
# If it is a {, then this is (most likely) JSON
c = file_obj.read(1)
while c.isspace():
c = file_obj.read(1)
if c == '{':
file_obj.seek(old_pos)
t = Table.from_json(json.load(file_obj,
object_pairs_hook=OrderedDict),
input_is_dense=input_is_dense)
else:
file_obj.seek(old_pos)
t = Table.from_tsv(file_obj, None, None, lambda x: x)
elif isinstance(file_obj, list):
try:
t = Table.from_json(json.loads(''.join(file_obj),
object_pairs_hook=OrderedDict),
input_is_dense=input_is_dense)
except ValueError:
t = Table.from_tsv(file_obj, None, None, lambda x: x)
else:
t = Table.from_json(json.loads(file_obj,
object_pairs_hook=OrderedDict),
input_is_dense=input_is_dense)
def subset_ids(data, id_, md):
return id_ in ids
def gt_zero(vals, id_, md):
return np.any(vals)
if ids is not None:
t.filter(subset_ids, axis=axis)
axis = 'observation' if axis == 'sample' else 'sample'
t.filter(gt_zero, axis=axis)
return t
