lattice/pyerrors
1
0
Fork 0
mirror of https://github.com/fjosw/pyerrors.git synced 2025-06-30 16:59:27 +02:00
Code Issues Projects Releases Packages Wiki Activity

merge with develop

Browse source
This commit is contained in:
jkuhl-uni 2022-02-08 11:16:20 +01:00
commit 71fe86b8ba
59 changed files with 5367 additions and 1798 deletions
Download patch file Download diff file Expand all files Collapse all files

20
pyerrors/input/hadrons.py
View file

@ -1,6 +1,7 @@
import os
import h5py
import numpy as np
from collections import Counter
from ..obs import Obs, CObs
from ..correlators import Corr
@ -32,6 +33,10 @@ def _get_files(path, filestem, idl):
filtered_files.append(line)
cnfg_numbers.append(no)
if idl:
if Counter(list(idl)) != Counter(cnfg_numbers):
raise Exception("Not all configurations specified in idl found, configurations " + str(list(Counter(list(idl)) - Counter(cnfg_numbers))) + " are missing.")
# Check that configurations are evenly spaced
dc = np.unique(np.diff(cnfg_numbers))
if np.any(dc < 0):
@ -58,16 +63,13 @@ def read_meson_hd5(path, filestem, ens_id, meson='meson_0', tree='meson', idl=No
meson : str
label of the meson to be extracted, standard value meson_0 which
corresponds to the pseudoscalar pseudoscalar two-point function.
tree : str
Label of the upmost directory in the hdf5 file, default 'meson'
for outputs of the Meson module. Can be altered to read input
from other modules with similar structures.
idl : range
If specified only configurations in the given range are read in.
"""
files, idx = _get_files(path, filestem, idl)
tree = meson.rsplit('_')[0]
corr_data = []
infos = []
for hd5_file in files:
@ -155,7 +157,7 @@ def read_ExternalLeg_hd5(path, filestem, ens_id, idl=None):
raw_data = file['ExternalLeg/corr'][0][0].view('complex')
corr_data.append(raw_data)
if mom is None:
mom = np.array(str(file['ExternalLeg/info'].attrs['pIn'])[3:-2].strip().split(' '), dtype=int)
mom = np.array(str(file['ExternalLeg/info'].attrs['pIn'])[3:-2].strip().split(), dtype=float)
file.close()
corr_data = np.array(corr_data)
@ -200,9 +202,9 @@ def read_Bilinear_hd5(path, filestem, ens_id, idl=None):
raw_data = file['Bilinear/Bilinear_' + str(i) + '/corr'][0][0].view('complex')
corr_data[name].append(raw_data)
if mom_in is None:
mom_in = np.array(str(file['Bilinear/Bilinear_' + str(i) + '/info'].attrs['pIn'])[3:-2].strip().split(' '), dtype=int)
mom_in = np.array(str(file['Bilinear/Bilinear_' + str(i) + '/info'].attrs['pIn'])[3:-2].strip().split(), dtype=float)
if mom_out is None:
mom_out = np.array(str(file['Bilinear/Bilinear_' + str(i) + '/info'].attrs['pOut'])[3:-2].strip().split(' '), dtype=int)
mom_out = np.array(str(file['Bilinear/Bilinear_' + str(i) + '/info'].attrs['pOut'])[3:-2].strip().split(), dtype=float)
file.close()
@ -265,9 +267,9 @@ def read_Fourquark_hd5(path, filestem, ens_id, idl=None, vertices=["VA", "AV"]):
raw_data = file[tree + str(i) + '/corr'][0][0].view('complex')
corr_data[name].append(raw_data)
if mom_in is None:
mom_in = np.array(str(file[tree + str(i) + '/info'].attrs['pIn'])[3:-2].strip().split(' '), dtype=int)
mom_in = np.array(str(file[tree + str(i) + '/info'].attrs['pIn'])[3:-2].strip().split(), dtype=float)
if mom_out is None:
mom_out = np.array(str(file[tree + str(i) + '/info'].attrs['pOut'])[3:-2].strip().split(' '), dtype=int)
mom_out = np.array(str(file[tree + str(i) + '/info'].attrs['pOut'])[3:-2].strip().split(), dtype=float)
file.close()

313
pyerrors/input/json.py
View file

@ -6,8 +6,11 @@ import socket
import datetime
import platform
import warnings
import re
from ..obs import Obs
from ..covobs import Covobs
from ..correlators import Corr
from ..misc import _assert_equal_properties
from .. import version as pyerrorsversion
@ -18,8 +21,8 @@ def create_json_string(ol, description='', indent=1):
Parameters
----------
ol : list
List of objects that will be exported. At the moments, these objects can be
either of: Obs, list, numpy.ndarray.
List of objects that will be exported. At the moment, these objects can be
either of: Obs, list, numpy.ndarray, Corr.
All Obs inside a structure have to be defined on the same set of configurations.
description : str
Optional string that describes the contents of the json file.
@ -103,20 +106,6 @@ def create_json_string(ol, description='', indent=1):
dl.append(ed)
return dl
def _assert_equal_properties(ol, otype=Obs):
for o in ol:
if not isinstance(o, otype):
raise Exception("Wrong data type in list.")
for o in ol[1:]:
if not ol[0].is_merged == o.is_merged:
raise Exception("All Obs in list have to be defined on the same set of configs.")
if not ol[0].reweighted == o.reweighted:
raise Exception("All Obs in list have to have the same property 'reweighted'.")
if not ol[0].e_content == o.e_content:
raise Exception("All Obs in list have to be defined on the same set of configs.")
if not ol[0].idl == o.idl:
raise Exception("All Obs in list have to be defined on the same set of configurations.")
def write_Obs_to_dict(o):
d = {}
d['type'] = 'Obs'
@ -173,12 +162,44 @@ def create_json_string(ol, description='', indent=1):
d['cdata'] = cdata
return d
def _nan_Obs_like(obs):
samples = []
names = []
idl = []
for key, value in obs.idl.items():
samples.append([np.nan] * len(value))
names.append(key)
idl.append(value)
my_obs = Obs(samples, names, idl)
my_obs.reweighted = obs.reweighted
my_obs.is_merged = obs.is_merged
return my_obs
def write_Corr_to_dict(my_corr):
first_not_none = next(i for i, j in enumerate(my_corr.content) if np.all(j))
dummy_array = np.empty((my_corr.N, my_corr.N), dtype=object)
dummy_array[:] = _nan_Obs_like(my_corr.content[first_not_none].ravel()[0])
content = [o if o is not None else dummy_array for o in my_corr.content]
dat = write_Array_to_dict(np.array(content, dtype=object))
dat['type'] = 'Corr'
corr_meta_data = str(my_corr.tag)
if 'tag' in dat.keys():
dat['tag'].append(corr_meta_data)
else:
dat['tag'] = [corr_meta_data]
taglist = dat['tag']
dat['tag'] = {} # tag is now a dictionary, that contains the previous taglist in the key "tag"
dat['tag']['tag'] = taglist
if my_corr.prange is not None:
dat['tag']['prange'] = my_corr.prange
return dat
if not isinstance(ol, list):
ol = [ol]
d = {}
d['program'] = 'pyerrors %s' % (pyerrorsversion.__version__)
d['version'] = '0.1'
d['version'] = '0.2'
d['who'] = getpass.getuser()
d['date'] = datetime.datetime.now().astimezone().strftime('%Y-%m-%d %H:%M:%S %z')
d['host'] = socket.gethostname() + ', ' + platform.platform()
@ -193,6 +214,10 @@ def create_json_string(ol, description='', indent=1):
d['obsdata'].append(write_List_to_dict(io))
elif isinstance(io, np.ndarray):
d['obsdata'].append(write_Array_to_dict(io))
elif isinstance(io, Corr):
d['obsdata'].append(write_Corr_to_dict(io))
else:
raise Exception("Unkown datatype.")
jsonstring = json.dumps(d, indent=indent, cls=my_encoder, ensure_ascii=False)
@ -212,6 +237,7 @@ def create_json_string(ol, description='', indent=1):
return '\n'.join(split)
jsonstring = remove_quotationmarks(jsonstring)
jsonstring = jsonstring.replace('nan', 'NaN')
return jsonstring
@ -221,8 +247,8 @@ def dump_to_json(ol, fname, description='', indent=1, gz=True):
Parameters
----------
ol : list
List of objects that will be exported. At the moments, these objects can be
either of: Obs, list, numpy.ndarray.
List of objects that will be exported. At the moment, these objects can be
either of: Obs, list, numpy.ndarray, Corr.
All Obs inside a structure have to be defined on the same set of configurations.
fname : str
Filename of the output file.
@ -255,7 +281,7 @@ def dump_to_json(ol, fname, description='', indent=1, gz=True):
def import_json_string(json_string, verbose=True, full_output=False):
"""Reconstruct a list of Obs or structures containing Obs from a json string.
The following structures are supported: Obs, list, numpy.ndarray
The following structures are supported: Obs, list, numpy.ndarray, Corr
If the list contains only one element, it is unpacked from the list.
Parameters
@ -374,6 +400,33 @@ def import_json_string(json_string, verbose=True, full_output=False):
ret[-1].tag = taglist[i]
return np.reshape(ret, layout)
def get_Corr_from_dict(o):
if isinstance(o.get('tag'), list): # supports the old way
taglist = o.get('tag') # This had to be modified to get the taglist from the dictionary
temp_prange = None
elif isinstance(o.get('tag'), dict):
tagdic = o.get('tag')
taglist = tagdic['tag']
if 'prange' in tagdic:
temp_prange = tagdic['prange']
else:
temp_prange = None
else:
raise Exception("The tag is not a list or dict")
corr_tag = taglist[-1]
tmp_o = o
tmp_o['tag'] = taglist[:-1]
if len(tmp_o['tag']) == 0:
del tmp_o['tag']
dat = get_Array_from_dict(tmp_o)
my_corr = Corr([None if np.isnan(o.ravel()[0].value) else o for o in list(dat)])
if corr_tag != 'None':
my_corr.tag = corr_tag
my_corr.prange = temp_prange
return my_corr
json_dict = json.loads(json_string)
prog = json_dict.get('program', '')
@ -400,6 +453,10 @@ def import_json_string(json_string, verbose=True, full_output=False):
ol.append(get_List_from_dict(io))
elif io['type'] == 'Array':
ol.append(get_Array_from_dict(io))
elif io['type'] == 'Corr':
ol.append(get_Corr_from_dict(io))
else:
raise Exception("Unkown datatype.")
if full_output:
retd = {}
@ -420,9 +477,9 @@ def import_json_string(json_string, verbose=True, full_output=False):
def load_json(fname, verbose=True, gz=True, full_output=False):
"""Import a list of Obs or structures containing Obs from a .json.gz file.
"""Import a list of Obs or structures containing Obs from a .json(.gz) file.
The following structures are supported: Obs, list, numpy.ndarray
The following structures are supported: Obs, list, numpy.ndarray, Corr
If the list contains only one element, it is unpacked from the list.
Parameters
@ -451,3 +508,215 @@ def load_json(fname, verbose=True, gz=True, full_output=False):
d = fin.read()
return import_json_string(d, verbose, full_output)
def _ol_from_dict(ind, reps='DICTOBS'):
"""Convert a dictionary of Obs objects to a list and a dictionary that contains
placeholders instead of the Obs objects.
Parameters
----------
ind : dict
Dict of JSON valid structures and objects that will be exported.
At the moment, these object can be either of: Obs, list, numpy.ndarray, Corr.
All Obs inside a structure have to be defined on the same set of configurations.
reps : str
Specify the structure of the placeholder in exported dict to be reps[0-9]+.
"""
obstypes = (Obs, Corr, np.ndarray)
if not reps.isalnum():
raise Exception('Placeholder string has to be alphanumeric!')
ol = []
counter = 0
def dict_replace_obs(d):
nonlocal ol
nonlocal counter
x = {}
for k, v in d.items():
if isinstance(v, dict):
v = dict_replace_obs(v)
elif isinstance(v, list) and all([isinstance(o, Obs) for o in v]):
v = obslist_replace_obs(v)
elif isinstance(v, list):
v = list_replace_obs(v)
elif isinstance(v, obstypes):
ol.append(v)
v = reps + '%d' % (counter)
counter += 1
elif isinstance(v, str):
if bool(re.match(r'%s[0-9]+' % (reps), v)):
raise Exception('Dict contains string %s that matches the placeholder! %s Cannot be savely exported.' % (v, reps))
x[k] = v
return x
def list_replace_obs(li):
nonlocal ol
nonlocal counter
x = []
for e in li:
if isinstance(e, list):
e = list_replace_obs(e)
elif isinstance(e, list) and all([isinstance(o, Obs) for o in e]):
e = obslist_replace_obs(e)
elif isinstance(e, dict):
e = dict_replace_obs(e)
elif isinstance(e, obstypes):
ol.append(e)
e = reps + '%d' % (counter)
counter += 1
elif isinstance(e, str):
if bool(re.match(r'%s[0-9]+' % (reps), e)):
raise Exception('Dict contains string %s that matches the placeholder! %s Cannot be savely exported.' % (e, reps))
x.append(e)
return x
def obslist_replace_obs(li):
nonlocal ol
nonlocal counter
il = []
for e in li:
il.append(e)
ol.append(il)
x = reps + '%d' % (counter)
counter += 1
return x
nd = dict_replace_obs(ind)
return ol, nd
def dump_dict_to_json(od, fname, description='', indent=1, reps='DICTOBS', gz=True):
"""Export a dict of Obs or structures containing Obs to a .json(.gz) file
Parameters
----------
od : dict
Dict of JSON valid structures and objects that will be exported.
At the moment, these objects can be either of: Obs, list, numpy.ndarray, Corr.
All Obs inside a structure have to be defined on the same set of configurations.
fname : str
Filename of the output file.
description : str
Optional string that describes the contents of the json file.
indent : int
Specify the indentation level of the json file. None or 0 is permissible and
saves disk space.
reps : str
Specify the structure of the placeholder in exported dict to be reps[0-9]+.
gz : bool
If True, the output is a gzipped json. If False, the output is a json file.
"""
if not isinstance(od, dict):
raise Exception('od has to be a dictionary. Did you want to use dump_to_json?')
infostring = ('This JSON file contains a python dictionary that has been parsed to a list of structures. '
'OBSDICT contains the dictionary, where Obs or other structures have been replaced by '
'' + reps + '[0-9]+. The field description contains the additional description of this JSON file. '
'This file may be parsed to a dict with the pyerrors routine load_json_dict.')
desc_dict = {'INFO': infostring, 'OBSDICT': {}, 'description': description}
ol, desc_dict['OBSDICT'] = _ol_from_dict(od, reps=reps)
dump_to_json(ol, fname, description=desc_dict, indent=indent, gz=gz)
def _od_from_list_and_dict(ol, ind, reps='DICTOBS'):
"""Parse a list of Obs or structures containing Obs and an accompanying
dict, where the structures have been replaced by placeholders to a
dict that contains the structures.
The following structures are supported: Obs, list, numpy.ndarray, Corr
Parameters
----------
ol : list
List of objects -
At the moment, these objects can be either of: Obs, list, numpy.ndarray, Corr.
All Obs inside a structure have to be defined on the same set of configurations.
ind : dict
Dict that defines the structure of the resulting dict and contains placeholders
reps : str
Specify the structure of the placeholder in imported dict to be reps[0-9]+.
"""
if not reps.isalnum():
raise Exception('Placeholder string has to be alphanumeric!')
counter = 0
def dict_replace_string(d):
nonlocal counter
nonlocal ol
x = {}
for k, v in d.items():
if isinstance(v, dict):
v = dict_replace_string(v)
elif isinstance(v, list):
v = list_replace_string(v)
elif isinstance(v, str) and bool(re.match(r'%s[0-9]+' % (reps), v)):
index = int(v[len(reps):])
v = ol[index]
counter += 1
x[k] = v
return x
def list_replace_string(li):
nonlocal counter
nonlocal ol
x = []
for e in li:
if isinstance(e, list):
e = list_replace_string(e)
elif isinstance(e, dict):
e = dict_replace_string(e)
elif isinstance(e, str) and bool(re.match(r'%s[0-9]+' % (reps), e)):
index = int(e[len(reps):])
e = ol[index]
counter += 1
x.append(e)
return x
nd = dict_replace_string(ind)
if counter == 0:
raise Exception('No placeholder has been replaced! Check if reps is set correctly.')
return nd
def load_json_dict(fname, verbose=True, gz=True, full_output=False, reps='DICTOBS'):
"""Import a dict of Obs or structures containing Obs from a .json(.gz) file.
The following structures are supported: Obs, list, numpy.ndarray, Corr
Parameters
----------
fname : str
Filename of the input file.
verbose : bool
Print additional information that was written to the file.
gz : bool
If True, assumes that data is gzipped. If False, assumes JSON file.
full_output : bool
If True, a dict containing auxiliary information and the data is returned.
If False, only the data is returned.
reps : str
Specify the structure of the placeholder in imported dict to be reps[0-9]+.
"""
indata = load_json(fname, verbose=verbose, gz=gz, full_output=True)
description = indata['description']['description']
indict = indata['description']['OBSDICT']
ol = indata['obsdata']
od = _od_from_list_and_dict(ol, indict, reps=reps)
if full_output:
indata['description'] = description
indata['obsdata'] = od
return indata
else:
return od

17
pyerrors/input/misc.py
View file

@ -1,6 +1,3 @@
#!/usr/bin/env python
# coding: utf-8
import os
import fnmatch
import re
@ -12,11 +9,12 @@ from ..obs import Obs
def read_pbp(path, prefix, **kwargs):
"""Read pbp format from given folder structure. Returns a list of length nrw
Keyword arguments
-----------------
r_start -- list which contains the first config to be read for each replicum
r_stop -- list which contains the last config to be read for each replicum
Parameters
----------
r_start : list
list which contains the first config to be read for each replicum
r_stop : list
list which contains the last config to be read for each replicum
"""
extract_nfct = 1
@ -66,7 +64,6 @@ def read_pbp(path, prefix, **kwargs):
tmp_array = []
with open(path + '/' + ls[rep], 'rb') as fp:
# header
t = fp.read(4) # number of reweighting factors
if rep == 0:
nrw = struct.unpack('i', t)[0]
@ -74,7 +71,7 @@ def read_pbp(path, prefix, **kwargs):
deltas.append([])
else:
if nrw != struct.unpack('i', t)[0]:
raise Exception('Error: different number of reweighting factors for replicum', rep)
raise Exception('Error: different number of factors for replicum', rep)
for k in range(nrw):
tmp_array.append([])

294
pyerrors/input/openQCD.py
View file

@ -1,11 +1,11 @@
#!/usr/bin/env python
# coding: utf-8
import os
import fnmatch
import re
import struct
import numpy as np # Thinly-wrapped numpy
import warnings
import matplotlib.pyplot as plt
from matplotlib import gridspec
from ..obs import Obs
from ..fits import fit_lin
@ -15,16 +15,31 @@ def read_rwms(path, prefix, version='2.0', names=None, **kwargs):
Parameters
----------
path : str
path that contains the data files
prefix : str
all files in path that start with prefix are considered as input files.
May be used together postfix to consider only special file endings.
Prefix is ignored, if the keyword 'files' is used.
version : str
version of openQCD, default 2.0
names : list
list of names that is assigned to the data according according
to the order in the file list. Use careful, if you do not provide file names!
r_start : list
list which contains the first config to be read for each replicum
r_stop : list
list which contains the last config to be read for each replicum
r_step : int
integer that defines a fixed step size between two measurements (in units of configs)
If not given, r_step=1 is assumed.
postfix : str
postfix of the file to read, e.g. '.ms1' for openQCD-files
idl_offsets : list
offsets to the idl range of obs. Useful for the case that the measurements of rwms are only starting at cfg. 20
files : list
list which contains the filenames to be read. No automatic detection of
files performed if given.
print_err : bool
Print additional information that is useful for debugging.
"""
known_oqcd_versions = ['1.4', '1.6', '2.0']
if not (version in known_oqcd_versions):
@ -44,7 +59,6 @@ def read_rwms(path, prefix, version='2.0', names=None, **kwargs):
if 'files' in kwargs:
ls = kwargs.get('files')
else:
# Exclude files with different names
for exc in ls:
if not fnmatch.fnmatch(exc, prefix + '*' + postfix + '.dat'):
ls = list(set(ls) - set([exc]))
@ -56,8 +70,7 @@ def read_rwms(path, prefix, version='2.0', names=None, **kwargs):
r_start = kwargs.get('r_start')
if len(r_start) != replica:
raise Exception('r_start does not match number of replicas')
# Adjust Configuration numbering to python index
r_start = [o - 1 if o else None for o in r_start]
r_start = [o if o else None for o in r_start]
else:
r_start = [None] * replica
@ -68,16 +81,22 @@ def read_rwms(path, prefix, version='2.0', names=None, **kwargs):
else:
r_stop = [None] * replica
if 'r_step' in kwargs:
r_step = kwargs.get('r_step')
else:
r_step = 1
print('Read reweighting factors from', prefix[:-1], ',',
replica, 'replica', end='')
# Adjust replica names to new bookmarking system
if names is None:
rep_names = []
for entry in ls:
truncated_entry = entry.split('.')[0]
idx = truncated_entry.index('r')
rep_names.append(truncated_entry[:idx] + '|' + truncated_entry[idx:])
else:
rep_names = names
print_err = 0
if 'print_err' in kwargs:
@ -86,11 +105,14 @@ def read_rwms(path, prefix, version='2.0', names=None, **kwargs):
deltas = []
configlist = []
r_start_index = []
r_stop_index = []
for rep in range(replica):
tmp_array = []
with open(path + '/' + ls[rep], 'rb') as fp:
# header
t = fp.read(4) # number of reweighting factors
if rep == 0:
nrw = struct.unpack('i', t)[0]
@ -99,7 +121,6 @@ def read_rwms(path, prefix, version='2.0', names=None, **kwargs):
for k in range(nrw):
deltas.append([])
else:
# little weird if-clause due to the /2 operation needed.
if ((nrw != struct.unpack('i', t)[0] and (not version == '2.0')) or (nrw != struct.unpack('i', t)[0] / 2 and version == '2.0')):
raise Exception('Error: different number of reweighting factors for replicum', rep)
@ -112,8 +133,6 @@ def read_rwms(path, prefix, version='2.0', names=None, **kwargs):
for i in range(nrw):
t = fp.read(4)
nfct.append(struct.unpack('i', t)[0])
# print('nfct: ', nfct) # Hasenbusch factor,
# 1 for rat reweighting
else:
for i in range(nrw):
nfct.append(1)
@ -126,13 +145,13 @@ def read_rwms(path, prefix, version='2.0', names=None, **kwargs):
if not struct.unpack('i', fp.read(4))[0] == 0:
print('something is wrong!')
# body
configlist.append([])
while 0 < 1:
t = fp.read(4)
if len(t) < 4:
break
if print_err:
config_no = struct.unpack('i', t)
config_no = struct.unpack('i', t)[0]
configlist[-1].append(config_no)
for i in range(nrw):
if(version == '2.0'):
tmpd = _read_array_openQCD2(fp)
@ -163,8 +182,32 @@ def read_rwms(path, prefix, version='2.0', names=None, **kwargs):
print('Partial factor:', tmp_nfct)
tmp_array[i].append(tmp_nfct)
if r_start[rep] is None:
r_start_index.append(0)
else:
try:
r_start_index.append(configlist[-1].index(r_start[rep]))
except ValueError:
raise Exception('Config %d not in file with range [%d, %d]' % (
r_start[rep], configlist[-1][0], configlist[-1][-1])) from None
if r_stop[rep] is None:
r_stop_index.append(len(configlist[-1]) - 1)
else:
try:
r_stop_index.append(configlist[-1].index(r_stop[rep]))
except ValueError:
raise Exception('Config %d not in file with range [%d, %d]' % (
r_stop[rep], configlist[-1][0], configlist[-1][-1])) from None
for k in range(nrw):
deltas[k].append(tmp_array[k][r_start[rep]:r_stop[rep]])
deltas[k].append(tmp_array[k][r_start_index[rep]:r_stop_index[rep]][::r_step])
if np.any([len(np.unique(np.diff(cl))) != 1 for cl in configlist]):
raise Exception('Irregular spaced data in input file!', [len(np.unique(np.diff(cl))) for cl in configlist])
stepsizes = [list(np.unique(np.diff(cl)))[0] for cl in configlist]
if np.any([step != 1 for step in stepsizes]):
warnings.warn('Stepsize between configurations is greater than one!' + str(stepsizes), RuntimeWarning)
print(',', nrw, 'reweighting factors with', nsrc, 'sources')
if "idl_offsets" in kwargs:
@ -172,15 +215,9 @@ def read_rwms(path, prefix, version='2.0', names=None, **kwargs):
else:
idl_offsets = np.ones(nrw, dtype=int)
result = []
idl = [range(configlist[rep][r_start_index[rep]], configlist[rep][r_stop_index[rep]], r_step) for rep in range(replica)]
for t in range(nrw):
idl = []
for rep in range(replica):
idl.append(range(idl_offsets[rep], len(deltas[t][rep] + idl_offsets[rep])))
if names is None:
result.append(Obs(deltas[t], rep_names, idl=idl))
else:
print(names)
result.append(Obs(deltas[t], names, idl=idl))
result.append(Obs(deltas[t], rep_names, idl=idl))
return result
@ -193,7 +230,12 @@ def extract_t0(path, prefix, dtr_read, xmin,
The data around the zero crossing of t^2<E> - 0.3
is fitted with a linear function
from which the exact root is extracted.
Only works with openQCD v 1.2.
Only works with openQCD
It is assumed that one measurement is performed for each config.
If this is not the case, the resulting idl, as well as the handling
of r_start, r_stop and r_step is wrong and the user has to correct
this in the resulting observable.
Parameters
----------
@ -215,10 +257,26 @@ def extract_t0(path, prefix, dtr_read, xmin,
crossing to be included in the linear fit. (Default: 5)
r_start : list
list which contains the first config to be read for each replicum.
r_stop: list
r_stop : list
list which contains the last config to be read for each replicum.
r_step : int
integer that defines a fixed step size between two measurements (in units of configs)
If not given, r_step=1 is assumed.
plaquette : bool
If true extract the plaquette estimate of t0 instead.
names : list
list of names that is assigned to the data according according
to the order in the file list. Use careful, if you do not provide file names!
files : list
list which contains the filenames to be read. No automatic detection of
files performed if given.
plot_fit : bool
If true, the fit for the extraction of t0 is shown together with the data.
assume_thermalization : bool
If True: If the first record divided by the distance between two measurements is larger than
1, it is assumed that this is due to thermalization and the first measurement belongs
to the first config (default).
If False: The config numbers are assumed to be traj_number // difference
"""
ls = []
@ -229,20 +287,21 @@ def extract_t0(path, prefix, dtr_read, xmin,
if not ls:
raise Exception('Error, directory not found')
# Exclude files with different names
for exc in ls:
if not fnmatch.fnmatch(exc, prefix + '*.ms.dat'):
ls = list(set(ls) - set([exc]))
if len(ls) > 1:
ls.sort(key=lambda x: int(re.findall(r'\d+', x[len(prefix):])[0]))
if 'files' in kwargs:
ls = kwargs.get('files')
else:
for exc in ls:
if not fnmatch.fnmatch(exc, prefix + '*.ms.dat'):
ls = list(set(ls) - set([exc]))
if len(ls) > 1:
ls.sort(key=lambda x: int(re.findall(r'\d+', x[len(prefix):])[0]))
replica = len(ls)
if 'r_start' in kwargs:
r_start = kwargs.get('r_start')
if len(r_start) != replica:
raise Exception('r_start does not match number of replicas')
# Adjust Configuration numbering to python index
r_start = [o - 1 if o else None for o in r_start]
r_start = [o if o else None for o in r_start]
else:
r_start = [None] * replica
@ -253,14 +312,31 @@ def extract_t0(path, prefix, dtr_read, xmin,
else:
r_stop = [None] * replica
if 'r_step' in kwargs:
r_step = kwargs.get('r_step')
else:
r_step = 1
print('Extract t0 from', prefix, ',', replica, 'replica')
if 'names' in kwargs:
rep_names = kwargs.get('names')
else:
rep_names = []
for entry in ls:
truncated_entry = entry.split('.')[0]
idx = truncated_entry.index('r')
rep_names.append(truncated_entry[:idx] + '|' + truncated_entry[idx:])
Ysum = []
configlist = []
r_start_index = []
r_stop_index = []
for rep in range(replica):
with open(path + '/' + ls[rep], 'rb') as fp:
# Read header
t = fp.read(12)
header = struct.unpack('iii', t)
if rep == 0:
@ -279,12 +355,13 @@ def extract_t0(path, prefix, dtr_read, xmin,
Ysl = []
# Read body
configlist.append([])
while 0 < 1:
t = fp.read(4)
if(len(t) < 4):
break
nc = struct.unpack('i', t)[0]
configlist[-1].append(nc)
t = fp.read(8 * tmax * (nn + 1))
if kwargs.get('plaquette'):
@ -302,6 +379,38 @@ def extract_t0(path, prefix, dtr_read, xmin,
current + tmax - xmin])
for current in range(0, len(item), tmax)])
diffmeas = configlist[-1][-1] - configlist[-1][-2]
configlist[-1] = [item // diffmeas for item in configlist[-1]]
if kwargs.get('assume_thermalization', True) and configlist[-1][0] > 1:
warnings.warn('Assume thermalization and that the first measurement belongs to the first config.')
offset = configlist[-1][0] - 1
configlist[-1] = [item - offset for item in configlist[-1]]
if r_start[rep] is None:
r_start_index.append(0)
else:
try:
r_start_index.append(configlist[-1].index(r_start[rep]))
except ValueError:
raise Exception('Config %d not in file with range [%d, %d]' % (
r_start[rep], configlist[-1][0], configlist[-1][-1])) from None
if r_stop[rep] is None:
r_stop_index.append(len(configlist[-1]) - 1)
else:
try:
r_stop_index.append(configlist[-1].index(r_stop[rep]))
except ValueError:
raise Exception('Config %d not in file with range [%d, %d]' % (
r_stop[rep], configlist[-1][0], configlist[-1][-1])) from None
if np.any([len(np.unique(np.diff(cl))) != 1 for cl in configlist]):
raise Exception('Irregular spaced data in input file!', [len(np.unique(np.diff(cl))) for cl in configlist])
stepsizes = [list(np.unique(np.diff(cl)))[0] for cl in configlist]
if np.any([step != 1 for step in stepsizes]):
warnings.warn('Stepsize between configurations is greater than one!' + str(stepsizes), RuntimeWarning)
idl = [range(configlist[rep][r_start_index[rep]], configlist[rep][r_stop_index[rep]], r_step) for rep in range(replica)]
t2E_dict = {}
for n in range(nn + 1):
samples = []
@ -309,8 +418,8 @@ def extract_t0(path, prefix, dtr_read, xmin,
samples.append([])
for cnfg in rep:
samples[-1].append(cnfg[n])
samples[-1] = samples[-1][r_start[nrep]:r_stop[nrep]]
new_obs = Obs(samples, [(w.split('.'))[0] for w in ls])
samples[-1] = samples[-1][r_start_index[nrep]:r_stop_index[nrep]][::r_step]
new_obs = Obs(samples, rep_names, idl=idl)
t2E_dict[n * dn * eps] = (n * dn * eps) ** 2 * new_obs / (spatial_extent ** 3) - 0.3
zero_crossing = np.argmax(np.array(
@ -323,32 +432,62 @@ def extract_t0(path, prefix, dtr_read, xmin,
[o.gamma_method() for o in y]
fit_result = fit_lin(x, y)
if kwargs.get('plot_fit'):
plt.figure()
gs = gridspec.GridSpec(2, 1, height_ratios=[3, 1], wspace=0.0, hspace=0.0)
ax0 = plt.subplot(gs[0])
xmore = list(t2E_dict.keys())[zero_crossing - fit_range - 2: zero_crossing + fit_range + 2]
ymore = list(t2E_dict.values())[zero_crossing - fit_range - 2: zero_crossing + fit_range + 2]
[o.gamma_method() for o in ymore]
ax0.errorbar(xmore, [yi.value for yi in ymore], yerr=[yi.dvalue for yi in ymore], fmt='x')
xplot = np.linspace(np.min(x), np.max(x))
yplot = [fit_result[0] + fit_result[1] * xi for xi in xplot]
[yi.gamma_method() for yi in yplot]
ax0.fill_between(xplot, y1=[yi.value - yi.dvalue for yi in yplot], y2=[yi.value + yi.dvalue for yi in yplot])
retval = (-fit_result[0] / fit_result[1])
retval.gamma_method()
ylim = ax0.get_ylim()
ax0.fill_betweenx(ylim, x1=retval.value - retval.dvalue, x2=retval.value + retval.dvalue, color='gray', alpha=0.4)
ax0.set_ylim(ylim)
ax0.set_ylabel(r'$t^2 \langle E(t) \rangle - 0.3 $')
xlim = ax0.get_xlim()
fit_res = [fit_result[0] + fit_result[1] * xi for xi in x]
residuals = (np.asarray([o.value for o in y]) - [o.value for o in fit_res]) / np.asarray([o.dvalue for o in y])
ax1 = plt.subplot(gs[1])
ax1.plot(x, residuals, 'ko', ls='none', markersize=5)
ax1.tick_params(direction='out')
ax1.tick_params(axis="x", bottom=True, top=True, labelbottom=True)
ax1.axhline(y=0.0, ls='--', color='k')
ax1.fill_between(xlim, -1.0, 1.0, alpha=0.1, facecolor='k')
ax1.set_xlim(xlim)
ax1.set_ylabel('Residuals')
ax1.set_xlabel(r'$t/a^2$')
plt.show()
return -fit_result[0] / fit_result[1]
def _parse_array_openQCD2(d, n, size, wa, quadrupel=False):
arr = []
if d == 2:
tot = 0
for i in range(n[d - 1] - 1):
for i in range(n[0]):
tmp = wa[i * n[1]:(i + 1) * n[1]]
if quadrupel:
tmp = wa[tot:n[d - 1]]
tmp2 = []
for i in range(len(tmp)):
if i % 2 == 0:
tmp2.append(tmp[i])
for j in range(0, len(tmp), 2):
tmp2.append(tmp[j])
arr.append(tmp2)
else:
arr.append(np.asarray(wa[tot:n[d - 1]]))
arr.append(np.asarray(tmp))
else:
raise Exception('Only two-dimensional arrays supported!')
return arr
# mimic the read_array routine of openQCD-2.0.
# fp is the opened file handle
# returns the dict array
# at this point we only parse a 2d array
# d = 2
# n = [nfct[irw], 2*nsrc[irw]]
def _read_array_openQCD2(fp):
t = fp.read(4)
d = struct.unpack('i', t)[0]
@ -380,42 +519,40 @@ def read_qtop(path, prefix, c, dtr_cnfg=1, version="1.2", **kwargs):
Parameters
----------
path:
path : str
path of the measurement files
prefix:
prefix : str
prefix of the measurement files, e.g. <prefix>_id0_r0.ms.dat
c: double
c : double
Smearing radius in units of the lattice extent, c = sqrt(8 t0) / L
dtr_cnfg: int
dtr_cnfg : int
(optional) parameter that specifies the number of trajectories
between two configs.
if it is not set, the distance between two measurements
in the file is assumed to be
the distance between two configurations.
steps: int
steps : int
(optional) (maybe only necessary for openQCD2.0)
nt step size, guessed if not given
version: str
version : str
version string of the openQCD (sfqcd) version used to create
the ensemble
L: int
L : int
spatial length of the lattice in L/a.
HAS to be set if version != sfqcd, since openQCD does not provide
this in the header
r_start: list
r_start : list
offset of the first ensemble, making it easier to match
later on with other Obs
r_stop: list
r_stop : list
last configurations that need to be read (per replicum)
files: list
files : list
specify the exact files that need to be read
from path, pratical if e.g. only one replicum is needed
names: list
from path, practical if e.g. only one replicum is needed
names : list
Alternative labeling for replicas/ensembles.
Has to have the appropriate length
"""
# one could read L from the header in case of sfQCD
# c = 0.35
known_versions = ["1.0", "1.2", "1.4", "1.6", "2.0", "sfqcd"]
if version not in known_versions:
@ -435,11 +572,9 @@ def read_qtop(path, prefix, c, dtr_cnfg=1, version="1.2", **kwargs):
r_start = kwargs.get("r_start")
if "r_stop" in kwargs:
r_stop = kwargs.get("r_stop")
# if one wants to read specific files with this method...
if "files" in kwargs:
files = kwargs.get("files")
else:
# find files in path
found = []
files = []
for (dirpath, dirnames, filenames) in os.walk(path + "/"):
@ -450,14 +585,12 @@ def read_qtop(path, prefix, c, dtr_cnfg=1, version="1.2", **kwargs):
if fnmatch.fnmatch(f, prefix + "*" + ".ms.dat"):
files.append(f)
print(files)
# now that we found our files, we dechiffer them...
rep_names = []
deltas = []
idl = []
for rep, file in enumerate(files):
with open(path + "/" + file, "rb") as fp:
# header
t = fp.read(12)
header = struct.unpack('<iii', t)
# step size in integration steps "dnms"
@ -486,7 +619,6 @@ def read_qtop(path, prefix, c, dtr_cnfg=1, version="1.2", **kwargs):
Q = []
ncs = []
while 0 < 1:
# int nt
t = fp.read(4)
if(len(t) < 4):
break
@ -533,8 +665,6 @@ def read_qtop(path, prefix, c, dtr_cnfg=1, version="1.2", **kwargs):
if len(Q_round) != len(ncs) // dtr_cnfg:
raise Exception("qtops and ncs dont have the same length")
# replica = len(files)
truncated_file = file[:-7]
print(truncated_file)
idl_start = 1
@ -562,17 +692,24 @@ def read_qtop(path, prefix, c, dtr_cnfg=1, version="1.2", **kwargs):
rep_names = names
deltas.append(np.array(Q_round))
idl.append(range(idl_start, idl_stop))
# print(idl)
result = Obs(deltas, rep_names, idl=idl)
return result
def read_qtop_sector(target=0, **kwargs):
"""target: int
specifies the topological sector to be reweighted to (default 0)
q_top: Obs
alternatively takes args of read_qtop method as kwargs
"""Constructs reweighting factors to a specified topological sector.
Parameters
----------
target : int
Specifies the topological sector to be reweighted to (default 0)
q_top : Obs
Alternatively takes args of read_qtop method as kwargs
"""
if not isinstance(target, int):
raise Exception("'target' has to be an integer.")
if "q_top" in kwargs:
qtop = kwargs.get("q_top")
else:
@ -603,7 +740,6 @@ def read_qtop_sector(target=0, **kwargs):
dtr_cnfg = 1
qtop = read_qtop(path, prefix, c, dtr_cnfg=dtr_cnfg,
version=version, **kwargs)
# unpack to original values, project onto target sector
names = qtop.names
print(names)
print(qtop.deltas.keys())

11
pyerrors/input/utils.py
View file

@ -2,11 +2,20 @@
def check_idl(idl, che):
"""Checks if list of configurations is contained in an idl
Parameters
----------
idl : range or list
idl of the current replicum
che : list
list of configurations to be checked against
"""
missing = []
for c in che:
if c not in idl:
missing.append(c)
# print missing such that it can directly be parsed to slurm terminal
# print missing configurations such that it can directly be parsed to slurm terminal
if not (len(missing) == 0):
print(len(missing), "configs missing")
miss_str = str(missing[0])