def _search_string_identifier(self, search_string, chunk_size=8):
with open(self.path, 'rb') as seeker:
data = None
total_chunk_size = chunk_size * 512
spec_start = None
# NOTE: This needs to go intp regex_patterns.py
regex_string = re.compile(
"<\s*spectrum[^>]*index=\"[0-9]+\"\sid=\"({0})\"\sdefaultArrayLength=\"[0-9]+\">"
.format("".join(['.*', search_string, '.*'])).encode())
search_string = search_string.encode()
while True:
file_pointer = seeker.tell()
data = seeker.read(total_chunk_size)
string, seeker = self._read_until_tag_end(seeker,
byte_mode=True)
data += string
spec_start = regex_string.search(data)
if spec_start:
spec_start_offset = file_pointer + spec_start.start()
current_index = spec_start.group(1)
if search_string in current_index:
seeker.seek(spec_start_offset)
start, end = self._read_to_spec_end(seeker)
seeker.seek(start)
spec_string = seeker.read(end - start)
xml_string = XML(spec_string)
return spec.Spectrum(xml_string,
measured_precision=5e-6)
elif len(data) == 0:
raise Exception('cant find specified string')
def __getitem__(self, identifier):
"""
Access the item with id 'identifier' in the file by iterating the xml-tree.
Arguments:
identifier (str): native id of the item to access
Returns:
data (str): text associated with the given identifier
"""
old_pos = self.file_handler.tell()
self.file_handler.seek(0, 0)
mzml_iter = iter(iterparse(self.file_handler, events=['end']))
while True:
event, element = next(mzml_iter)
if event == 'end':
if element.tag.endswith('}spectrum'):
if int(
regex_patterns.SPECTRUM_ID_PATTERN.search(
element.get('id')).group(0)) == identifier:
self.file_handler.seek(old_pos, 0)
return spec.Spectrum(element, measured_precision=5e-6)
elif element.tag.endswith('}chromatogram'):
if element.get('id') == identifier:
self.file_handler.seek(old_pos, 0)
return spec.Chromatogram(element,
measured_precision=5e-6)
def __next__(self):
"""
Iterator for the class :py:class:`Run`.
Iterates all of the spectra in the file.
Returns:
Spectrum (:py:class:`Spectrum`): a spectrum object with interface
to the original spectrum element.
Example:
>>> for spectrum in Reader:
... print(spectrum.mz, end='\\r')
"""
has_ref_group = self.info.get('referenceable_param_group_list', False)
while True:
event, element = next(self.iter, ('END', 'END'))
if event == 'end':
if element.tag.endswith('}spectrum'):
spectrum = spec.Spectrum(element)
if has_ref_group:
spectrum._set_params_from_reference_group(
self.info['referenceable_param_group_list_element']
)
ms_level = spectrum.ms_level
spectrum.measured_precision = self.ms_precisions[ms_level]
spectrum.calling_instance = self
return spectrum
elif event == 'END':
raise StopIteration
def __next__(self):
"""
Iterator for the class :py:class:`Run`.
Iterates all of the spectra in the file.
Returns:
Spectrum (:py:class:`Spectrum`): a spectrum object with interface
to the original spectrum element.
Example:
>>> for spectrum in Reader:
... print(spectrum.mz, end='\\r')
"""
while True:
event, element = next(self.iter, ('END', 'END'))
if event == 'end':
if element.tag.endswith('}spectrum'):
spectrum = spec.Spectrum(element)
ms_level = spectrum.ms_level
spectrum.measured_precision = self.ms_precisions[ms_level]
spectrum.calling_instance = self
return spectrum
elif event == 'END':
raise StopIteration
def _search_linear(self, seeker, index, chunk_size=8):
"""
Fallback to linear search if interpolated search fails.
"""
data = None
i = 0
total_chunk_size = chunk_size * 512
spec_start = None
spec_end = None
i = 0
# print('target', index)
while True:
file_pointer = seeker.tell()
data = seeker.read(total_chunk_size)
string, seeker = self._read_until_tag_end(seeker)
data += string
spec_start = self.spec_open.search(data)
if spec_start:
spec_start_offset = file_pointer + spec_start.start()
seeker.seek(spec_start_offset)
current_index = int(
re.search(b'[0-9]*$', spec_start.group('id')).group())
# print(current_index)
spec_end = self.spec_close.search(data[spec_start.start():])
if spec_end:
spec_end_offset = file_pointer + spec_end.end(
) + spec_start.start()
seeker.seek(spec_end_offset)
while spec_end is None:
file_pointer = seeker.tell()
data = seeker.read(total_chunk_size)
string, seeker = self._read_until_tag_end(seeker)
data += string
spec_end = self.spec_close.search(data)
if spec_end:
spec_end_offset = file_pointer + spec_end.end()
self.offset_dict[current_index] = (spec_start_offset,
spec_end_offset)
seeker.seek(spec_end_offset)
break
if current_index == index:
seeker.seek(spec_start_offset)
spec_string = seeker.read(spec_end_offset -
spec_start_offset)
self.offset_dict[current_index] = (spec_start_offset,
spec_end_offset)
xml_string = XML(spec_string)
seeker.close()
return spec.Spectrum(xml_string, measured_precision=5e-6)
def __getitem__(self, identifier):
"""
Access the item with id 'identifier'.
Either use linear, binary or interpolated search.
Arguments:
identifier (str): native id of the item to access
Returns:
data (str): text associated with the given identifier
"""
#############################################################################
# DOES NOT HOLD IF NUMBERS DONT START WITH ONE AND/OR DONT INCREASE BY ONE #
# TODO FIXME #
#############################################################################
self.file_handler.seek(0)
spectrum = None
if str(identifier).upper() == 'TIC':
# print(str(identifier).upper())
found = False
mzmliter = iter(iterparse(self.file_handler, events=['end']))
while found is False:
event, element = next(mzmliter, ('STOP', 'STOP'))
if event == 'end':
if element.tag.endswith('}chromatogram'):
if element.get('id') == 'TIC':
found = True
spectrum = spec.Chromatogram(
element, measured_precision=5e-6)
elif event == 'STOP':
raise StopIteration
elif identifier in self.offset_dict:
start = self.offset_dict[identifier]
with open(self.path, 'rb') as seeker:
seeker.seek(start[0])
start, end = self._read_to_spec_end(seeker)
self.file_handler.seek(start, 0)
data = self.file_handler.read(end - start)
spectrum = spec.Spectrum(XML(data), measured_precision=5e-6)
elif type(identifier) == str:
return self._search_string_identifier(identifier)
else:
spectrum = self._interpol_search(identifier)
return spectrum
def __getitem__(self, key):
"""
Execute a SQL request, process the data and return a spectrum object.
Args:
key (str or int): unique identifier for the given spectrum in the
database
"""
self.cursor.execute("SELECT * FROM spectra WHERE id=?", key)
ID, element = self.cursor.fetchone()
element = et.XML(element)
if "spectrum" in element.tag:
spectrum = spec.Spectrum(element)
elif "chromatogram" in element.tag:
spectrum = spec.Chromatogram(element)
return spectrum
def __getitem__(self, identifier):
"""
Access the item with id 'identifier' in the file.
Arguments:
identifier (str): native id of the item to access
Returns:
data (str): text associated with the given identifier
"""
#TODO more elegant way to add NameSpace (.register_namespace maybe??)
ns_prefix = '<mzML xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://psi.hupo.org/ms/mzml http://psidev.info/files/ms/mzML/xsd/mzML1.1.0.xsd" id="test_Creinhardtii_QE_pH8" version="1.1.0" xmlns="http://psi.hupo.org/ms/mzml">'
ns_suffix = '</mzML>'
data = self.Reader.read_block(identifier)
element = XML(ns_prefix + data.decode('utf-8') + ns_suffix)
if 'chromatogram' in element[0].tag:
return spec.Chromatogram(list(element)[0], measured_precision=5e-6)
else:
return spec.Spectrum(list(element)[0], measured_precision=5e-6)
def _interpol_search(self,
target_index,
chunk_size=8,
fallback_cutoff=100):
"""
Use linear interpolation search to find spectra faster.
Arguments:
target_index (str or int) : native id of the item to access
Keyword Arguments:
chunk_size (int) : size of the chunk to read in one go in kb
"""
# print('target ', target_index)
seeker = open(self.path, 'rb')
seeker.seek(0, 2)
chunk_size = chunk_size * 512
lower_bound = 0
upper_bound = seeker.tell()
mid = int(upper_bound / 2)
seeker.seek(mid, 0)
current_position = seeker.tell()
used_indices = set()
spectrum_found = False
spectrum = None
while spectrum_found is False:
jumper_scaling = 1
file_pointer = seeker.tell()
data = seeker.read(chunk_size)
spec_start = self.spec_open.search(data)
if spec_start is not None:
spec_start_offset = file_pointer + spec_start.start()
seeker.seek(spec_start_offset)
current_index = int(
re.search(b'[0-9]*$', spec_start.group('id')).group())
self.offset_dict[current_index] = (spec_start_offset, )
if current_index in used_indices:
# seeker.close()
if current_index > target_index:
jumper_scaling -= 0.1
else:
jumper_scaling += 0.1
used_indices.add(current_index)
dist = current_index - target_index
if dist < -1 and dist > -(fallback_cutoff):
spectrum = self._search_linear(seeker, target_index)
seeker.close()
spectrum_found = True
break
elif dist > 0 and dist < fallback_cutoff:
while current_index > target_index:
offset = int(current_position - chunk_size)
seeker.seek(offset if offset > 0 else 0)
lower_bound = current_position
current_position = seeker.tell()
data = seeker.read(chunk_size)
if self.spec_open.search(data):
spec_start = self.spec_open.search(data)
current_index = int(
re.search(b'[0-9]*$',
spec_start.group('id')).group())
seeker.seek(current_position)
spectrum = self._search_linear(seeker, target_index)
seeker.close()
spectrum_found = True
break
if int(current_index) == target_index:
seeker.seek(spec_start_offset)
start, end = self._read_to_spec_end(seeker)
seeker.seek(start)
self.offset_dict[current_index] = (start, end)
xml_string = seeker.read(end - start)
seeker.close()
spectrum = spec.Spectrum(XML(xml_string),
measured_precision=5e-6)
spectrum_found = True
break
elif int(current_index) > target_index:
scaling = target_index / current_index
seeker.seek(
int(current_position * scaling * jumper_scaling))
upper_bound = current_position
current_position = seeker.tell()
elif int(current_index) < target_index:
scaling = target_index / current_index
seeker.seek(
int(current_position * scaling * jumper_scaling))
lower_bound = current_position
current_position = seeker.tell()
elif len(data) == 0:
sorted_keys = sorted(self.offset_dict.keys())
pos = bisect.bisect_left(
sorted_keys, target_index) - 2 # dat magic number :)
try:
key = sorted_keys[pos]
spec_start_offset = self.offset_dict[key][0]
except:
key = sorted_keys[pos]
spec_start_offset = self.offset_dict[key][0]
seeker = open(self.path, 'rb')
seeker.seek(spec_start_offset)
spectrum = self._search_linear(seeker, target_index)
seeker.close()
spectrum_found = True
break
return spectrum