def read_spider_alignment(filename, header=None, **extra):
''' Read a SPIDER alignment data from a file
Args:
filename : str
Input filename containing alignment data
header : str
User-specified header for the alignment file
extra : dict
Unused extra keyword arguments
Returns:
align : list
List of named tuples
'''
align_header = [
"epsi,theta,phi,ref_num,id,psi,tx,ty,nproj,ang_diff,cc_rot,spsi,sx,sy,mirror,micrograph,stack_id,defocus",
"epsi,theta,phi,ref_num,id,psi,tx,ty,nproj,ang_diff,cc_rot,spsi,sx,sy,mirror,micrograph,stack_id",
"epsi,theta,phi,ref_num,id,psi,tx,ty,nproj,ang_diff,cc_rot,spsi,sx,sy,mirror"
]
if 'numeric' in extra: del extra['numeric']
align = None
for h in align_header:
try:
align = format.read(filename, numeric=True, header=h, **extra)
except: pass
else:
if len(align) == 0 or align[0]._fields[0]=='epsi':
break
if align is None:
align = format.read(filename, numeric=True, header=header, **extra)
return align
def read_spider_alignment(filename, header=None, **extra):
''' Read a SPIDER alignment data from a file
Args:
filename : str
Input filename containing alignment data
header : str
User-specified header for the alignment file
extra : dict
Unused extra keyword arguments
Returns:
align : list
List of named tuples
'''
align_header = [
"epsi,theta,phi,ref_num,id,psi,tx,ty,nproj,ang_diff,cc_rot,spsi,sx,sy,mirror,micrograph,stack_id,defocus",
"epsi,theta,phi,ref_num,id,psi,tx,ty,nproj,ang_diff,cc_rot,spsi,sx,sy,mirror,micrograph,stack_id",
"epsi,theta,phi,ref_num,id,psi,tx,ty,nproj,ang_diff,cc_rot,spsi,sx,sy,mirror"
]
if 'numeric' in extra: del extra['numeric']
align = None
for h in align_header:
try:
align = format.read(filename, numeric=True, header=h, **extra)
except:
pass
else:
if len(align) == 0 or align[0]._fields[0] == 'epsi':
break
if align is None:
align = format.read(filename, numeric=True, header=header, **extra)
return align
def decode(dict_file, in_file, out_file): types = idl.parse_es6_idl() ty_script = types.interfaces['Script'] string_dict = strings.read_dict(dict_file, with_signature=True) proggy = format.read(types, string_dict, ty_script, in_file) json.dump(proggy, out_file)
def read_alignment(filename, image_file, use_3d=False, align_cols=7, force_list=False, ctf_params=False, scale_spi=False, apix=0, class_index=0, **extra):
''' Read an alignment file and organize images
This comprehensive function reads alignment parameters from either a
Relion Star file or a SPIDER alignment file. It can optionally read
and return CTF information taken from a Relion star file or the combination
of the pySPIDER alignment file and a SPIDER params file.
.. note::
Supports both SPIDER and Relion alignment files
Args:
filename : str
Filename for alignment file
image_file : str
Filename for image template - not required for relion
use_3d : bool
If True, then translations are RT, otherwise they are TR
align_cols : int
Number of columns in alignment array (must be greater than 7)
force_list : bool
Return filename index tuple list rather than converting
to SPIDER ID label array
ctf_params : bool
Read and return CTF params
scale_spi : bool
Scale translations by the pixel size
class_index : int
Index of class to select
extra : dict
Unused keyword arguments
Returns:
files : list or tuple
List of filename/id tuples or a tuple (filename, label array)
out : array
2d array where rows are images and columns are alignment
parameters: psi,theta,phi,inplane,tx,ty,defocus
param : dict, optional
CTF params
'''
if align_cols < 7: raise ValueError, "align_cols must be greater than 7"
if not isinstance(image_file, str) and len(image_file) == 1: image_file=image_file[0]
id_len = extra['id_len'] if 'id_len' in extra else 0
if 'format' in extra: del extra['format']
if 'numeric' not in extra: extra['numeric']=True
supports_spider_id="" if not force_list else None
if isinstance(filename, list) and len(filename) > 0 and hasattr(filename[0], 'rlnImageName') or (not isinstance(filename, list) and is_relion_star(filename)):
if isinstance(filename, list):
align=filename
else:
align = format.read(filename, **extra)
param = numpy.zeros((len(align), align_cols))
files = []
fileset=set()
if use_3d:
_logger.info("Standard Relion alignment file - leave 3D")
for i in xrange(len(align)):
if hasattr(align[i], 'rlnClassNumber') and class_index > 0:
if align[i].rlnClassNumber != class_index: continue
filename, index = relion_utility.relion_file(align[i].rlnImageName)
if image_file != "":
filename = spider_utility.spider_filename(image_file, filename)
files.append((filename, index))
fileset.add(filename)
if supports_spider_id is not None and spider_utility.is_spider_filename(files[-1][0]):
if supports_spider_id == "":
supports_spider_id = spider_utility.spider_filepath(files[-1][0])
elif supports_spider_id != spider_utility.spider_filepath(files[-1][0]):
supports_spider_id=None
j = len(files)-1
param[j, 0] = align[i].rlnAnglePsi
param[j, 1] = align[i].rlnAngleTilt
param[j, 2] = align[i].rlnAngleRot
#param[j, 3] = align[i].
param[j, 4] = align[i].rlnOriginX
param[j, 5] = align[i].rlnOriginY
if hasattr(align[i], 'rlnDefocusV'):
param[j, 6] = (align[i].rlnDefocusU+align[i].rlnDefocusV)/2
else:
param[j, 6] = align[i].rlnDefocusU
else:
_logger.info("Standard Relion alignment file - convert to 2D")
for i in xrange(len(align)):
if hasattr(align[i], 'rlnClassNumber') and class_index > 0:
if align[i].rlnClassNumber != class_index: continue
filename, index = relion_utility.relion_file(align[i].rlnImageName)
if image_file != "":
filename = spider_utility.spider_filename(image_file, filename)
files.append((filename, index))
fileset.add(filename)
if supports_spider_id is not None and spider_utility.is_spider_filename(files[-1][0]):
if supports_spider_id == "":
supports_spider_id = spider_utility.spider_filepath(files[-1][0])
elif supports_spider_id != spider_utility.spider_filepath(files[-1][0]):
supports_spider_id=None
j = len(files)-1
param[j, 1] = align[i].rlnAngleTilt
param[j, 2] = align[i].rlnAngleRot
rot, tx, ty = spider_transforms.align_param_3D_to_2D(align[i].rlnAnglePsi, align[i].rlnOriginX, align[i].rlnOriginY)
param[j, 3] = rot
param[j, 4] = tx
param[j, 5] = ty
if hasattr(align[i], 'rlnDefocusV'):
param[j, 6] = (align[i].rlnDefocusU+align[i].rlnDefocusV)/2
else:
param[j, 6] = align[i].rlnDefocusU
param = param[:len(files)]
assert(numpy.alltrue(param[:, 6]>0.0))
if supports_spider_id is not None:
label = numpy.zeros((len(param), 2), dtype=numpy.int)
for i in xrange(len(files)): label[i, :] = (spider_utility.spider_id(files[i][0]), files[i][1])
if len(fileset) == len(numpy.unique(label[:, 0]).squeeze()):
label[:, 1]-=1
files = (files[0][0], label)
if label[:, 1].min() < 0: raise ValueError, "Cannot have a negative index"
else:
_logger.warn("Input filenames appear to be SPIDER but the ID is not unique")
if ctf_params:
ctf_param=dict(cs=align[0].rlnSphericalAberration,
voltage=align[0].rlnVoltage,
ampcont=align[0].rlnAmplitudeContrast) if len(align) > 0 else {}
else:
if ctf_params:
ctf_param=spider_params.read(extra['param_file'])
if apix == 0: apix=ctf_param['apix']
if 'ndarray' not in extra or not extra['ndarray']: extra['ndarray']=True
if isinstance(filename, list):
align=numpy.asarray(filename)
else:
align = read_spider_alignment(filename, **extra)[0]
param = numpy.zeros((len(align), align_cols))
if align.shape[1] > 17:
param[:, 6] = align[:, 17]
if use_3d:
if numpy.sum(align[:, 5]) != 0.0:
_logger.info("Standard SPIDER alignment file - convert to 3D")
if numpy.all(align[:,5]==0):
_logger.info("Detected non-standard SPIDER alignment file with no 2D parameters")
param[:, :3] = align[:, :3]
else:
for i in xrange(len(align)):
rot, tx, ty = spider_transforms.align_param_2D_to_3D(align[i, 5], align[i, 6], align[i, 7])
param[i, 0] = rot
param[i, 1:3] = align[i, 1:3]
param[i, 4:6] = (tx, ty)
else:
_logger.info("Detected non-standard SPIDER alignment file with only angles")
param[:, :3] = align[:, :3]
else:
_logger.info("Standard SPIDER alignment file - leave 2D")
param[:, 1:3] = align[:, 1:3]
if numpy.sum(align[:, 5]) != 0.0:
param[:, 3:6] = align[:, 5:8]
else:
_logger.info("Detected non-standard SPIDER alignment file with only angles")
if numpy.any(align[:,0]!=0):
param[:, 3]=-align[:,0]
if force_list or not spider_utility.is_spider_filename(image_file):
if align.shape[1] <= 15:
files = []
if isinstance(image_file, str) and align[:, 4].max() > ndimage_file.count_images(image_file):
for i in xrange(len(align)):
files.append( (image_file, int(i+1)) )
else:
idx = align[:, 4].astype(numpy.int)
for i in xrange(len(align)):
files.append( (image_file, idx[i]) )
else:
files=[]
label = align[:, 15:17].astype(numpy.int)
for mic, pid in label:
files.append((spider_utility.spider_filename(image_file, int(mic), id_len), int(pid)))
else:
label = numpy.zeros((len(param), 2), dtype=numpy.int)
if align.shape[1] <= 15:
label[:, 0] = spider_utility.spider_id(image_file)
if isinstance(image_file, str) and align[:, 4].max() > ndimage_file.count_images(image_file):
label[:, 1] = numpy.arange(1, len(align)+1)
else:
label[:, 1] = align[:, 4]
else:
label[:, :] = align[:, 15:17].astype(numpy.int)
label[:, 1]-=1
files = (image_file, label)
if label[:, 1].min() < 0: raise ValueError, "Cannot have a negative index"
if ctf_params: return files, param, ctf_param
return files, param
def read_alignment(filename,
image_file,
use_3d=False,
align_cols=7,
force_list=False,
ctf_params=False,
scale_spi=False,
apix=0,
class_index=0,
**extra):
''' Read an alignment file and organize images
This comprehensive function reads alignment parameters from either a
Relion Star file or a SPIDER alignment file. It can optionally read
and return CTF information taken from a Relion star file or the combination
of the pySPIDER alignment file and a SPIDER params file.
.. note::
Supports both SPIDER and Relion alignment files
Args:
filename : str
Filename for alignment file
image_file : str
Filename for image template - not required for relion
use_3d : bool
If True, then translations are RT, otherwise they are TR
align_cols : int
Number of columns in alignment array (must be greater than 7)
force_list : bool
Return filename index tuple list rather than converting
to SPIDER ID label array
ctf_params : bool
Read and return CTF params
scale_spi : bool
Scale translations by the pixel size
class_index : int
Index of class to select
extra : dict
Unused keyword arguments
Returns:
files : list or tuple
List of filename/id tuples or a tuple (filename, label array)
out : array
2d array where rows are images and columns are alignment
parameters: psi,theta,phi,inplane,tx,ty,defocus
param : dict, optional
CTF params
'''
if align_cols < 7: raise ValueError, "align_cols must be greater than 7"
if not isinstance(image_file, str) and len(image_file) == 1:
image_file = image_file[0]
id_len = extra['id_len'] if 'id_len' in extra else 0
if 'format' in extra: del extra['format']
if 'numeric' not in extra: extra['numeric'] = True
supports_spider_id = "" if not force_list else None
if isinstance(filename, list) and len(filename) > 0 and hasattr(
filename[0], 'rlnImageName') or (not isinstance(filename, list)
and is_relion_star(filename)):
if isinstance(filename, list):
align = filename
else:
align = format.read(filename, **extra)
param = numpy.zeros((len(align), align_cols))
files = []
fileset = set()
if use_3d:
_logger.info("Standard Relion alignment file - leave 3D")
for i in xrange(len(align)):
if hasattr(align[i], 'rlnClassNumber') and class_index > 0:
if align[i].rlnClassNumber != class_index: continue
filename, index = relion_utility.relion_file(
align[i].rlnImageName)
if image_file != "":
filename = spider_utility.spider_filename(
image_file, filename)
files.append((filename, index))
fileset.add(filename)
if supports_spider_id is not None and spider_utility.is_spider_filename(
files[-1][0]):
if supports_spider_id == "":
supports_spider_id = spider_utility.spider_filepath(
files[-1][0])
elif supports_spider_id != spider_utility.spider_filepath(
files[-1][0]):
supports_spider_id = None
j = len(files) - 1
param[j, 0] = align[i].rlnAnglePsi
param[j, 1] = align[i].rlnAngleTilt
param[j, 2] = align[i].rlnAngleRot
#param[j, 3] = align[i].
param[j, 4] = align[i].rlnOriginX
param[j, 5] = align[i].rlnOriginY
if hasattr(align[i], 'rlnDefocusV'):
param[j, 6] = (align[i].rlnDefocusU +
align[i].rlnDefocusV) / 2
else:
param[j, 6] = align[i].rlnDefocusU
else:
_logger.info("Standard Relion alignment file - convert to 2D")
for i in xrange(len(align)):
if hasattr(align[i], 'rlnClassNumber') and class_index > 0:
if align[i].rlnClassNumber != class_index: continue
filename, index = relion_utility.relion_file(
align[i].rlnImageName)
if image_file != "":
filename = spider_utility.spider_filename(
image_file, filename)
files.append((filename, index))
fileset.add(filename)
if supports_spider_id is not None and spider_utility.is_spider_filename(
files[-1][0]):
if supports_spider_id == "":
supports_spider_id = spider_utility.spider_filepath(
files[-1][0])
elif supports_spider_id != spider_utility.spider_filepath(
files[-1][0]):
supports_spider_id = None
j = len(files) - 1
param[j, 1] = align[i].rlnAngleTilt
param[j, 2] = align[i].rlnAngleRot
rot, tx, ty = spider_transforms.align_param_3D_to_2D(
align[i].rlnAnglePsi, align[i].rlnOriginX,
align[i].rlnOriginY)
param[j, 3] = rot
param[j, 4] = tx
param[j, 5] = ty
if hasattr(align[i], 'rlnDefocusV'):
param[j, 6] = (align[i].rlnDefocusU +
align[i].rlnDefocusV) / 2
else:
param[j, 6] = align[i].rlnDefocusU
param = param[:len(files)]
assert (numpy.alltrue(param[:, 6] > 0.0))
if supports_spider_id is not None:
label = numpy.zeros((len(param), 2), dtype=numpy.int)
for i in xrange(len(files)):
label[i, :] = (spider_utility.spider_id(files[i][0]),
files[i][1])
if len(fileset) == len(numpy.unique(label[:, 0]).squeeze()):
label[:, 1] -= 1
files = (files[0][0], label)
if label[:, 1].min() < 0:
raise ValueError, "Cannot have a negative index"
else:
_logger.warn(
"Input filenames appear to be SPIDER but the ID is not unique"
)
if ctf_params:
ctf_param = dict(cs=align[0].rlnSphericalAberration,
voltage=align[0].rlnVoltage,
ampcont=align[0].rlnAmplitudeContrast
) if len(align) > 0 else {}
else:
if ctf_params:
ctf_param = spider_params.read(extra['param_file'])
if apix == 0: apix = ctf_param['apix']
if 'ndarray' not in extra or not extra['ndarray']:
extra['ndarray'] = True
if isinstance(filename, list):
align = numpy.asarray(filename)
else:
align = read_spider_alignment(filename, **extra)[0]
param = numpy.zeros((len(align), align_cols))
if align.shape[1] > 17:
param[:, 6] = align[:, 17]
if use_3d:
if numpy.sum(align[:, 5]) != 0.0:
_logger.info("Standard SPIDER alignment file - convert to 3D")
if numpy.all(align[:, 5] == 0):
_logger.info(
"Detected non-standard SPIDER alignment file with no 2D parameters"
)
param[:, :3] = align[:, :3]
else:
for i in xrange(len(align)):
rot, tx, ty = spider_transforms.align_param_2D_to_3D(
align[i, 5], align[i, 6], align[i, 7])
param[i, 0] = rot
param[i, 1:3] = align[i, 1:3]
param[i, 4:6] = (tx, ty)
else:
_logger.info(
"Detected non-standard SPIDER alignment file with only angles"
)
param[:, :3] = align[:, :3]
else:
_logger.info("Standard SPIDER alignment file - leave 2D")
param[:, 1:3] = align[:, 1:3]
if numpy.sum(align[:, 5]) != 0.0:
param[:, 3:6] = align[:, 5:8]
else:
_logger.info(
"Detected non-standard SPIDER alignment file with only angles"
)
if numpy.any(align[:, 0] != 0):
param[:, 3] = -align[:, 0]
if force_list or not spider_utility.is_spider_filename(image_file):
if align.shape[1] <= 15:
files = []
if isinstance(image_file, str) and align[:, 4].max(
) > ndimage_file.count_images(image_file):
for i in xrange(len(align)):
files.append((image_file, int(i + 1)))
else:
idx = align[:, 4].astype(numpy.int)
for i in xrange(len(align)):
files.append((image_file, idx[i]))
else:
files = []
label = align[:, 15:17].astype(numpy.int)
for mic, pid in label:
files.append((spider_utility.spider_filename(
image_file, int(mic), id_len), int(pid)))
else:
label = numpy.zeros((len(param), 2), dtype=numpy.int)
if align.shape[1] <= 15:
label[:, 0] = spider_utility.spider_id(image_file)
if isinstance(image_file, str) and align[:, 4].max(
) > ndimage_file.count_images(image_file):
label[:, 1] = numpy.arange(1, len(align) + 1)
else:
label[:, 1] = align[:, 4]
else:
label[:, :] = align[:, 15:17].astype(numpy.int)
label[:, 1] -= 1
files = (image_file, label)
if label[:, 1].min() < 0:
raise ValueError, "Cannot have a negative index"
if ctf_params: return files, param, ctf_param
return files, param