qharv.reel package¶
Subpackages¶
Submodules¶
qharv.reel.ascii_out module¶
-
qharv.reel.ascii_out.block_text(mm, header, trailer, **kwargs)[source]¶ find a block of text in between header and trailer header and trailer are not included by default
e.g. given text in mm ‘’’ begin important data 1 2 3 4 5 6 7 8 9 end important data ‘’’
mm.block_text(mm, ‘begin’, ‘end’) returns ‘’’1 2 3 4 5 6 7 8 9 ‘’’
- Parameters
mm (mmap.mmap) – memory map to text file
header (str) – string indicating the beginning of block
trailer (str) – string indicating the end of block
-
qharv.reel.ascii_out.lr_mark(line, lmark, rmark)[source]¶ - read a string segment from line, which is enclosed between l&rmark
e.g. extract the contents in parenteses
- Parameters
line (str) – text line
lmark (str) – left marker, e.g. ‘(‘
rmark (str) – right marker, e.g. ‘)’
- Returns
text in between left and right markers
- Return type
str
-
qharv.reel.ascii_out.name_val_table(text, dtype=<class 'float'>)[source]¶ - designed to parse optVariables text block
e.g. ‘’’uu_0 1.0770e+00 1 1 ON 0
uu_1 6.7940e-01 1 1 ON 1 uu_2 4.3156e-01 1 1 ON 2 ud_0 1.6913e+00 1 1 ON 5 ud_1 1.0443e+00 1 1 ON 6 ud_2 6.1912e-01 1 1 ON 7 ‘’’
return variable-value map, only the first two columns are parsed.
- Args:
text (str): text block such as given in the example dtype (type): data type of value, default is float
- Return:
dict: variable name -> value map
qharv.reel.config_h5 module¶
-
qharv.reel.config_h5.extract_checkpoint_walkers(fconfig)[source]¶ extract the checkpoint walkers from config.h5 file
- Parameters
fconfig (str) – path to config.h5 file
- Returns
a list of walkers of shape (nconf, nptcl, ndim)
- Return type
np.array
-
qharv.reel.config_h5.loadh5(fname, path='/data')[source]¶ load matrix from h5 file, mimic np.loadtxt
- Parameters
fname (str) – name of hdf5 to read
- Returns
matrix of data
- Return type
np.array
-
qharv.reel.config_h5.save_dict(arr_dict, h5file, slab=None)[source]¶ - save a dictionary of numpy arrays into h5file
each entry will create its own sub-slab using key as name
- Parameters
arr_dict (dict) – dictionary of numpy arrays
h5file (tables.file.File) – pytables File
slab (tables.Group, optional) – HDF5 slab, if None, then use root
-
qharv.reel.config_h5.save_vec(vec, h5file, slab, name)[source]¶ save numpy array into an h5 slab under name
- Parameters
vec (np.array) – numpy ndarray of arbitrary dimension and type
h5file (tables.file.File) – pytables File
slab (tables.Group) – HDF5 slab
name (str) – name of CArray to create
-
qharv.reel.config_h5.saveh5(fname, mat, name='data')[source]¶ save matrix at root of h5 file, mimic call signature of np.savetxt
e.g. mat = np.eye(3) saveh5(‘mat.h5’, mat) $ h5ls mat.h5 data Dataset {3/2730, 3}
- Parameters
fname (str) – name of hdf5 file to write
mat (np.array) – 2D numpy array of floats
name (str, optional) – CArray name at the root of the hdf5 file
qharv.reel.mole module¶
-
qharv.reel.mole.build_qmcpack_fname(entry)[source]¶ inverse of interpret_qmcpack_fname
- Parameters
entry (dict) – a dictionary of meta data, must include
in key (['id','grouped','group','series','category','ext']) –
- Returns
filename
- Return type
str
-
qharv.reel.mole.clean_path(path)[source]¶ remove . and .. from path
- Parameters
path (str) – file or folder path
- Returns
clean path
- Return type
str
-
qharv.reel.mole.files_with_regex(regex, rundir, case=True, ftype='f', **kwargs)[source]¶ find files with the given suffix in folder rundir rely on bash find command
- Parameters
regex (str) – regular expression for file names
rundir (str) – directory containing the files to be found
case (bool, optional) – case sensity, default is True
ftype (str, optional) – files type, default is regular file ‘f’
may be directory 'd' (,) –
- Returns
flist, a list of filenames matching the given regular expression
- Return type
list
-
qharv.reel.mole.find(regex, rundir, **kwargs)[source]¶ find the first file that matches the given regular expression RuntimeError will be raised unless exactly one file is found
- Parameters
regex (str) – regular expression for file names
rundir (str) – directory containing the files to be found
- Returns
filename
- Return type
str
-
qharv.reel.mole.findall(regex, rundir, prepend_star=True, **kwargs)[source]¶ find all matching files with regular expression in rundir
- Parameters
regex (str) – regular expression for file names
rundir (str) – directory containing the files to be found
prepend_star (bool, optional) – add ‘*’ to regex, default True
- Returns
flist, list of all matching files
- Return type
list
-
qharv.reel.mole.interpret_qmcpack_fname(fname)[source]¶ extract metadata regarding the contents of a file based on its filename. QMCPACK generates files having a pre-determined suffix structure. This function will interpret the last 4 period-separated segments of the suffix.
- fname examples:
qmc.s000.scalar.dat qmc.g000.s000.stat.h5 qmc.g161.s000.config.h5 qmc.g005.s001.cont.xml
- Parameters
fname (str) – filename, must end in one of [‘dat’,’h5’,’qmc’,’xml’].
- Returns
a dictionary of metadata.
- Return type
dict
qharv.reel.qmcpack_out module¶
-
qharv.reel.qmcpack_out.opt_vars(fout, begin_tag='<optVariables', end_tag='</optVariables')[source]¶ Parse optimizable variables
- Parameters
fout (str) – optimization output
begin_tag (str, optional) – default ‘<optVariables’
end_tag (str, optional) – default ‘</optVariables’
- Returns
a list of dictionary, each a key-value map of opt. var.s
- Return type
list
qharv.reel.scalar_dat module¶
-
qharv.reel.scalar_dat.error(trace, kappa=None)[source]¶ Calculate the error of a trace of scalar data
- Parameters
trace (list) – should be a 1D iterable array of floating point numbers
kappa (float,optional) – auto-correlation time, default is to re-calculate
- Returns
stderr, the error of the mean of this trace of scalars
- Return type
float
-
qharv.reel.scalar_dat.find_header_lines(text)[source]¶ Find line numbers of all headers
- Parameters
text (str) – text to parse
- Returns
a list of integer line numbers
- Return type
list
-
qharv.reel.scalar_dat.get_string_io(text)[source]¶ - Obtain StringIO object from text
compatible with Python 2 and 3
- Parameters
text (str) – text to parse
- Returns
file-like object
- Return type
StringIO
-
qharv.reel.scalar_dat.parse(text)[source]¶ Parse text of a scalar.dat file, should be table format.
- Parameters
text (str) – content of scalar.dat file
- Returns
table data
- Return type
pd.DataFrame
Example
>>> with open('vmc.s001.scalar.dat', 'r') as f: >>> text = f.read() >>> df = parse(text)
-
qharv.reel.scalar_dat.read(dat_fname)[source]¶ - Read the scalar.dat file, should be table format.
The header line should start with ‘#’ and contain column labels.
- Parameters
dat_fname (str) – name of input file
- Returns
df containing the table of data
- Return type
pd.DataFrame
Example
>>> df = read('vmc.s001.scalar.dat')
-
qharv.reel.scalar_dat.read_to_list(dat_fname)[source]¶ Read scalar.dat file into a list of pandas DataFrames
A line is a header if its first column cannot be converted to a float. Many scalar.dat files can be concatenated. A list will be returned.
- Parameters
dat_fname (str) – name of input file
- Returns
list of df(s) containing the table(s) of data
- Return type
list
Example
>>> dfl = read_to_list('gctas.dat') >>> df = pd.concat(dfl).reset_index(drop=True)
-
qharv.reel.scalar_dat.single_column(df, column, nequil)[source]¶ Calculate mean and error of a column
nequil blocks of data are thrown out; autocorrelation time is taken into
account when calculating error. The equilibrated data is assumed to have Gaussian distribution. Error is calculated for one standard deviation (1-sigma error).
- Parameters
df (pd.DataFrame) – table of data (e.g. from scalar_dat.parse)
column (str) – name of column
nequil (int) – number of equilibration blocks
- Returns
- (ymean,yerr,ycorr), where ymean is the mean of column
, yerr is the 1-sigma error of column, and ycorr is the autocorrelation
- Return type
(float,float,float)
-
qharv.reel.scalar_dat.write(dat_fname, df, header_pad='# ', **kwargs)[source]¶ Write dataframe to plain text scalar table format
Lightly wrap around pandas.to_string with defaults to index and float_format
- Parameters
dat_fname (str) – output data file name
df (pd.DataFrame) – data
header_pad (str, optional) – pad beginning of header with comment string, default ‘# ‘
qharv.reel.stat_h5 module¶
-
qharv.reel.stat_h5.afobs(fp, obs_name, nequil, kappa=None, numer='one_rdm')[source]¶ - extract “best” 1RMD output from AFQMC stat.h5 file
BackPropagated (BP)
- Parameters
fp (h5py.File) – h5py handle of stat.h5 file
obs_name (str) – observable name, probably ‘FullOneRDM’
nequil (int) – number of equilibration BP blocks to remove
kappa (float, optional) – auto-correlation, default recalculate
numer (str, optional) – numerator to extract, default ‘one_rdm’
- Returns
a dictionary of 1RDMs, one for each species (eg. u, d)
- Return type
dict
-
qharv.reel.stat_h5.dsk_from_skall(fp, obs_name, nequil, kappa=None)[source]¶ extract fluctuating structure factor dS(k) from skall
- Parameters
fp (h5py.File) – stat.h5 handle
obs_name (str, optional) – name the “skall” estimator, likely “skall”
nequil (int) – equilibration length
kappa (float, optional) – auto-correlation, default recalculate
- Returns
- (kvecs, dskm, dske),
kvectors and S(k) mean and error
- Return type
(np.array, np.array, np.array)
-
qharv.reel.stat_h5.gofr(fp, obs_name, nequil, kappa=None, force=False)[source]¶ extract pair correlation function g(r) from stat.h5 file :param fp: h5py handle of stat.h5 file :type fp: h5py.File :param obs_name: observable name, should start with ‘gofr’, e.g. gofr_e_0_1 :type obs_name: str :param nequil: number of equilibration blocks to remove :type nequil: int :param kappa: auto-correlation, default recalculate :type kappa: float, optional :param force: force execution, i.e. skip all checks :type force: bool, optional
- Returns
(myr, grm, gre): bin locations, g(r) mean, g(r) error
- Return type
tuple
-
qharv.reel.stat_h5.me2d(edata, kappa=None, axis=0)[source]¶ Calculate mean and error of a table of columns
- Parameters
edata (np.array) – 2D array of equilibrated time series data
kappa (float, optional) – pre-calculate auto-correlation, default is to re-calculate on-the-fly
axis (int, optional) – axis to average over, default 0 i.e. columns
- Returns
(mean, error) of each column
- Return type
(np.array, np.array)
-
qharv.reel.stat_h5.mean_and_err(handle, obs_path, nequil, kappa=None)[source]¶ calculate mean and error of an observable from QMCPACK stat.h5 file
- Parameters
handle (h5py.Group) – or h5py.File or h5py.Dataset
obs_path (str) – path to observable, e.g. ‘gofr_e_1_1’
nequil (int) – number of equilibration blocks to throw out
kappa (float, optional) – auto-correlation, default recalculate
- Returns
(mean, err), the mean and error of observable
- Return type
(np.array, np.array)
-
qharv.reel.stat_h5.nofk(fp, obs_name, nequil, kappa=None)[source]¶ extract momentum estimator output n(k) from stat.h5 file
- Parameters
fp (h5py.File) – h5py handle of stat.h5 file
obs_name (str) – observable name, probably ‘nofk’
nequil (int) – number of equilibration blocks to remove
kappa (float, optional) – auto-correlation, default recalculate
- Returns
- (kvecs, nkm, nke),
k-vectors, n(k) mean and error
- Return type
(np.array, np.array, np.array)
-
qharv.reel.stat_h5.rdm1(fp, obs_name, nequil, kappa=None)[source]¶ extract 1RMD output from stat.h5 file
- Parameters
fp (h5py.File) – h5py handle of stat.h5 file
obs_name (str) – observable name, probably ‘1rdms’
nequil (int) – number of equilibration blocks to remove
kappa (float, optional) – auto-correlation, default recalculate
- Returns
a dictionary of 1RDMs, one for each species (eg. u, d)
- Return type
dict
-
qharv.reel.stat_h5.rhok_from_skall(fp, obs_name, nequil, kappa=None)[source]¶ extract electronic density rho(k) from stat.h5 file
- Parameters
fp (h5py.File) – h5py handle of stat.h5 file
obs_name (str, optional) – name the “skall” estimator, likely “skall”
nequil (int) – number of equilibration blocks to remove
kappa (float, optional) – auto-correlation, default recalculate
- Returns
- (kvecs, rhokm, rhoke)
k-vectors, rho(k) mean and error, shape (nk, ndim) notice rhok is the real-view of a complex vector, shape (2*nk,)
- Return type
(np.array, np.array, np.array)