"""Driver for MFP-3D files.
-This driver reads Igor binary waves.
+This driver reads IGOR binary waves.
AUTHORS:
Matlab version: Richard Naud August 2008 (http://lcn.epfl.ch/~naud/)
Hooke submission: Rolf Schmidt, Alberto Gomez-Casado 2009
"""
-# DEFINITION:
-# Reads Igor's (Wavemetric) binary wave format, .ibw, files.
-#
-# ALGORITHM:
-# Parsing proper to version 2, 3, or version 5 (see Technical notes TN003.ifn:
-# http://mirror.optus.net.au/pub/wavemetrics/IgorPro/Technical_Notes/) and data
-# type 2 or 4 (non complex, single or double precision vector, real values).
-#
-# VERSION: 0.1
-#
-# COMMENTS:
-# Only tested for version 2 Igor files for now, testing for 3 and 5 remains to be done.
-# More header data could be passed back if wished. For significance of ignored bytes see
-# the technical notes linked above.
+import copy
+import pprint
import numpy
-import os.path
-import struct
-
-from .. import curve as lhc
+from .. import curve as curve
+from .. import experiment as experiment
+from ..config import Setting
+from . import Driver as Driver
+from .igorbinarywave import loadibw
-__version__='0.0.0.20100310'
+__version__='0.0.0.20100604'
-class DataChunk(list):
- #Dummy class to provide ext and ret methods to the data list.
-
- def ext(self):
- halflen=(len(self)/2)
- return self[0:halflen]
-
- def ret(self):
- halflen=(len(self)/2)
- return self[halflen:]
-class mfp3dDriver(lhc.Driver):
+class MFP3DDriver (Driver):
+ """Handle Asylum Research's MFP3D data format.
+ """
+ def __init__(self):
+ super(MFP3DDriver, self).__init__(name='mfp3d')
- #Construction and other special methods
+ def is_me(self, path):
+ """Look for identifying fields in the IBW note.
+ """
+ if not path.endswith('.ibw'):
+ return False
+ targets = ['Version:', 'XOPVersion:', 'ForceNote:']
+ found = [False]*len(targets)
+ for line in open(path, 'rU'):
+ for i,ft in enumerate(zip(found, targets)):
+ f,t = ft
+ if f == False and line.startswith(t):
+ found[i] = True
+ if min(found) == True:
+ return True
+ return False
- def __init__(self,filename):
- '''
- constructor method
- '''
-
- self.textfile =file(filename)
- self.binfile=file(filename,'rb')
- #unnecesary, but some other part of the program expects these to be open
-
- self.forcechunk=0
- self.distancechunk=1
- #TODO eliminate the need to set chunk numbers
-
- self.filepath=filename
- self.debug=True
-
- self.data = []
- self.note = []
- self.retract_velocity = None
- self.spring_constant = None
- self.filename = filename
-
- self.filedata = open(filename,'rU')
- self.lines = list(self.filedata.readlines())
- self.filedata.close()
+ def read(self, path):
+ data,bin_info,wave_info = loadibw(path)
+ approach,retract = self._translate_ibw(data, bin_info, wave_info)
- self.filetype = 'mfp3d'
- self.experiment = 'smfs'
-
+ info = {'filetype':self.name, 'experiment':experiment.VelocityClamp}
+ return ([approach, retract], info)
- def _get_data_chunk(self,whichchunk):
-
- data = None
- f = open(self.filename, 'rb')
- ####################### ORDERING
- # machine format for IEEE floating point with big-endian
- # byte ordering
- # MacIgor use the Motorola big-endian 'b'
- # WinIgor use Intel little-endian 'l'
- # If the first byte in the file is non-zero, then the file is a WinIgor
- firstbyte = struct.unpack('b', f.read(1))[0]
- if firstbyte == 0:
- format = '>'
- else:
- format = '<'
- ####################### CHECK VERSION
- f.seek(0)
- version = struct.unpack(format+'h', f.read(2))[0]
- ####################### READ DATA AND ACCOMPANYING INFO
- if version == 2 or version == 3:
- # pre header
- wfmSize = struct.unpack(format+'i', f.read(4))[0] # The size of the WaveHeader2 data structure plus the wave data plus 16 bytes of padding.
- noteSize = struct.unpack(format+'i', f.read(4))[0] # The size of the note text.
- if version==3:
- formulaSize = struct.unpack(format+'i', f.read(4))[0]
- pictSize = struct.unpack(format+'i', f.read(4))[0] # Reserved. Write zero. Ignore on read.
- checksum = struct.unpack(format+'H', f.read(2))[0] # Checksum over this header and the wave header.
- # wave header
- dtype = struct.unpack(format+'h', f.read(2))[0]
- if dtype == 2:
- dtype = numpy.float32(.0).dtype
- elif dtype == 4:
- dtype = numpy.double(.0).dtype
- else:
- assert False, "Wave is of type '%i', not supported" % dtype
- dtype = dtype.newbyteorder(format)
-
- ignore = f.read(4) # 1 uint32
- bname = self._flatten(struct.unpack(format+'20c', f.read(20)))
- ignore = f.read(4) # 2 int16
- ignore = f.read(4) # 1 uint32
- dUnits = self._flatten(struct.unpack(format+'4c', f.read(4)))
- xUnits = self._flatten(struct.unpack(format+'4c', f.read(4)))
- npnts = struct.unpack(format+'i', f.read(4))[0]
- amod = struct.unpack(format+'h', f.read(2))[0]
- dx = struct.unpack(format+'d', f.read(8))[0]
- x0 = struct.unpack(format+'d', f.read(8))[0]
- ignore = f.read(4) # 2 int16
- fsValid = struct.unpack(format+'h', f.read(2))[0]
- topFullScale = struct.unpack(format+'d', f.read(8))[0]
- botFullScale = struct.unpack(format+'d', f.read(8))[0]
- ignore = f.read(16) # 16 int8
- modDate = struct.unpack(format+'I', f.read(4))[0]
- ignore = f.read(4) # 1 uint32
- # Numpy algorithm works a lot faster than struct.unpack
- data = numpy.fromfile(f, dtype, npnts)
-
- elif version == 5:
- # pre header
- checksum = struct.unpack(format+'H', f.read(2))[0] # Checksum over this header and the wave header.
- wfmSize = struct.unpack(format+'i', f.read(4))[0] # The size of the WaveHeader2 data structure plus the wave data plus 16 bytes of padding.
- formulaSize = struct.unpack(format+'i', f.read(4))[0]
- noteSize = struct.unpack(format+'i', f.read(4))[0] # The size of the note text.
- dataEUnitsSize = struct.unpack(format+'i', f.read(4))[0]
- dimEUnitsSize = struct.unpack(format+'4i', f.read(16))
- dimLabelsSize = struct.unpack(format+'4i', f.read(16))
- sIndicesSize = struct.unpack(format+'i', f.read(4))[0]
- optionSize1 = struct.unpack(format+'i', f.read(4))[0]
- optionSize2 = struct.unpack(format+'i', f.read(4))[0]
-
- # header
- ignore = f.read(4)
- CreationDate = struct.unpack(format+'I',f.read(4))[0]
- modData = struct.unpack(format+'I',f.read(4))[0]
- npnts = struct.unpack(format+'i',f.read(4))[0]
- # wave header
- dtype = struct.unpack(format+'h',f.read(2))[0]
- if dtype == 2:
- dtype = numpy.float32(.0).dtype
- elif dtype == 4:
- dtype = numpy.double(.0).dtype
- else:
- assert False, "Wave is of type '%i', not supported" % dtype
- dtype = dtype.newbyteorder(format)
-
- ignore = f.read(2) # 1 int16
- ignore = f.read(6) # 6 schar, SCHAR = SIGNED CHAR? ignore = fread(fid,6,'schar'); #
- ignore = f.read(2) # 1 int16
- bname = self._flatten(struct.unpack(format+'32c',f.read(32)))
- ignore = f.read(4) # 1 int32
- ignore = f.read(4) # 1 int32
- ndims = struct.unpack(format+'4i',f.read(16)) # Number of of items in a dimension -- 0 means no data.
- sfA = struct.unpack(format+'4d',f.read(32))
- sfB = struct.unpack(format+'4d',f.read(32))
- dUnits = self._flatten(struct.unpack(format+'4c',f.read(4)))
- xUnits = self._flatten(struct.unpack(format+'16c',f.read(16)))
- fsValid = struct.unpack(format+'h',f.read(2))
- whpad3 = struct.unpack(format+'h',f.read(2))
- ignore = f.read(16) # 2 double
- ignore = f.read(40) # 10 int32
- ignore = f.read(64) # 16 int32
- ignore = f.read(6) # 3 int16
- ignore = f.read(2) # 2 char
- ignore = f.read(4) # 1 int32
- ignore = f.read(4) # 2 int16
- ignore = f.read(4) # 1 int32
- ignore = f.read(8) # 2 int32
-
- data = numpy.fromfile(f, dtype, npnts)
- note_str = f.read(noteSize)
- note_lines = note_str.split('\r')
- self.note = {}
- for line in note_lines:
- if ':' in line:
- key, value = line.split(':', 1)
- self.note[key] = value
- self.retract_velocity = float(self.note['Velocity'])
- self.spring_constant = float(self.note['SpringConstant'])
- else:
- assert False, "Fileversion is of type '%i', not supported" % dtype
- data = []
-
- f.close()
- if len(data) > 0:
- #we have 3 columns: deflection, LVDT, raw
- #TODO detect which is each one
- count = npnts / 3
- lvdt = data[:count]
- deflection = data[count:2 * count]
- #every column contains data for extension and retraction
- #we assume the same number of points for each
- #we could possibly extract this info from the note
- count = npnts / 6
-
- forcechunk=deflection*self.spring_constant
- distancechunk=lvdt
-
- if whichchunk==self.forcechunk:
- return forcechunk
- if whichchunk==self.distancechunk:
- return distancechunk
- else:
- return None
-
- def _force(self):
- #returns force vector
- Kspring=self.spring_constant
- return DataChunk([(meter*Kspring) for meter in self._deflection()])
-
- def _deflection(self):
- #for internal use (feeds _force)
- deflect=self.data_chunks[self.forcechunk]/self.spring_constant
- return deflect
-
- def _flatten(self, tup):
- out = ''
- for ch in tup:
- out += ch
- return out
-
- def _Z(self):
- return DataChunk(self.data_chunks[self.distancechunk])
-
- def is_me(self):
- if len(self.lines) < 34:
- return False
-
- name, extension = os.path.splitext(self.filename)
- if extension == '.ibw':
- for line in self.lines:
- if line.startswith('ForceNote:'):
- self.data_chunks=[self._get_data_chunk(num) for num in [0,1,2]]
- return True
+ def _translate_ibw(self, data, bin_info, wave_info):
+ if bin_info['version'] != 5:
+ raise NotImplementedError('IBW version %d (< 5) not supported'
+ % bin_info['version'])
+ # We need version 5 for multidimensional arrays.
+
+ # Parse the note into a dictionary
+ note = {}
+ for line in bin_info['note'].split('\r'):
+ fields = [x.strip() for x in line.split(':', 1)]
+ key = fields[0]
+ if len(fields) == 2:
+ value = fields[1]
else:
- return False
+ value = None
+ note[key] = value
+ bin_info['note'] = note
+ if note['VerDate'] not in ['80501.041', '80501.0207']:
+ raise Exception(note['VerDate'])
+ raise NotImplementedError(
+ '%s file version %s not supported (yet!)\n%s'
+ % (self.name, note['VerDate'], pprint.pformat(note)))
+
+ info = {
+ 'raw info':{'bin':bin_info,
+ 'wave':wave_info},
+ 'time':wave_info['creationDate'],
+ 'spring constant (N/m)':note['SpringConstant'],
+ }
+ # MFP3D's native data dimensions match Hooke's (<point>, <column>) layout.
+ approach = self._scale_block(data[:wave_info['npnts']/2,:], info, 'approach')
+ retract = self._scale_block(data[wave_info['npnts']/2:,:], info, 'retract')
+ return (approach, retract)
+
+ def _scale_block(self, data, info, name):
+ """Convert the block from its native format to a `numpy.float`
+ array in SI units.
+ """
+ shape = 3
+ # raw column indices
+ columns = info['raw info']['bin']['dimLabels'][1]
+ # Depending on your MFP3D version:
+ # VerDate 80501.0207: ['Raw', 'Defl', 'LVDT', 'Time']
+ # VerDate 80501.041: ['Raw', 'Defl', 'LVDT']
+ if 'Time' in columns:
+ n_col = 3
else:
- return False
-
- def close_all(self):
- '''
- Explicitly closes all files
- '''
- self.textfile.close()
- self.binfile.close()
-
- def default_plots(self):
- '''
- creates the default PlotObject
- '''
- force=self._force()
- zdomain=self._Z()
- main_plot=lhc.PlotObject()
- main_plot.vectors=[[zdomain.ext(), force.ext()],[zdomain.ret(), force.ret()]]
- main_plot.normalize_vectors()
- main_plot.units=['meters','newton']
- main_plot.destination=0
- main_plot.title=self.filepath
-
-
- return [main_plot]
-
- def deflection(self):
- #interface for correct plotmanip and others
- deflectionchunk=DataChunk(self._deflection())
- return deflectionchunk.ext(),deflectionchunk.ret()
+ n_col = 2
+ ret = curve.Data(
+ shape=(data.shape[0], n_col),
+ dtype=numpy.float,
+ info=copy.deepcopy(info)
+ )
+ ret.info['name'] = name
+ ret.info['raw data'] = data # store the raw data
+
+ z_rcol = columns.index('LVDT')
+ d_rcol = columns.index('Defl')
+
+ # scaled column indices
+ ret.info['columns'] = ['z piezo (m)', 'deflection (m)']
+ z_scol = ret.info['columns'].index('z piezo (m)')
+ d_scol = ret.info['columns'].index('deflection (m)')
+
+ # Leading '-' because increasing voltage extends the piezo,
+ # moving the tip towards the surface (positive indentation),
+ # but it makes more sense to me to have it increase away from
+ # the surface (positive separation).
+ ret[:,z_scol] = -data[:,z_rcol].astype(ret.dtype)
+
+ # Leading '-' because deflection voltage increases as the tip
+ # moves away from the surface, but it makes more sense to me
+ # to have it increase as it moves toward the surface (positive
+ # tension on the protein chain).
+ ret[:,d_scol] = -data[:,d_rcol]
+
+ if 'Time' in columns:
+ ret.info['columns'].append('time (s)')
+ t_rcol = columns.index('Time')
+ t_scol = ret.info['columns'].index('time (s)')
+ ret[:,t_scol] = data[:,t_rcol]
+
+ return ret
projects / hooke.git / commitdiff