list comprehension dictionary

# Another version of reading point data from a file using re

import re

import datetime, time

data_file = r'H:\TEMP\temsys\sample_points3.txt'

def convert_data(s):

for func in (int, float):

try:

n = func(s)

return n

except:

pass

return s

def read_file_data(f):

ptDict = {}

wireDict = {}

fList = open(f).readlines()

in_pnt = False

patt = re.compile(r'''\d+\.\d+E\+\d+| # engineering notation ++

-\d+\.\d+E\+\d+| # engineering notation -+

-\d+\.\d+E-\d+| # engineering notation --

\d+\.\d+E-\d+| # engineering notation +-

\d+\.\d+| # positive float format

-\d+\.\d+| # negative float format

\d+\.| # positive float format

-\d+\.| # negative float format

\d+ # positive integer

''', re.X

)

for line in fList:

lineList = [x.lower().strip() for x in line.strip().split(' ', 1) if x != '']

if in_pnt:

pt_data += re.findall(patt, line)

in_pnt = False

ptDict[convert_data(pt_data[0])] = [convert_data(x) for x in pt_data[1:] if x != '0']

elif 'pnt' in line.lower() and '*' in line:

in_pnt = True

pt_data = re.findall(patt, line)

elif 'rect' in lineList or 'tria' in lineList:

sLst = re.findall(r'\d{8}', line)

wireDict[convert_data(sLst[0])] = [convert_data(x) for x in sLst[1:]]

elif 'pnt' in lineList:

sLst = re.findall(r'\d.{7}|(?<=\d.{7}).+', line)

ptDict[convert_data(sLst[0])] = [convert_data(x.strip()) for x in sLst[1:]]

return ptDict,wireDict

if __name__ == '__main__':

ptDict,wireDict = read_file_data(data_file)

print 'Point dictionary:'

for key in ptDict:

print '%s = %s' % (key, ptDict[key])

print '\nWire dictionary:'

for key in wireDict:

print '%s = %s' % (key, wireDict[key])

print time.time()

print time.clock()

print time.ctime(time.time())

print time.strftime('%H:%M:%S', time.localtime())

'''

>>> Point dictionary:

30400000 = [10.0, 0.0, 0.0]

40000000 = [15.0, 0.0, 0.0]

2 = [5.0, 0.0, 0.0]

3 = [10.0, 0.0, 0.0]

4 = [15.0, 0.0, 0.0]

5 = [20.0, 0.0, 0.0]

6 = [0.0, 5.0, 0.0]

1 = [0.0, 0.0, 0.0]

8 = [10.0, 5.0, 0.0]

9 = [15.0, 5.0, 0.0]

10270106 = [1097.15002, 250.67631499999999, 140.78933699999999]

10270107 = [1115.47864, 271.83373999999998, 144.698837]

10010012 = [3.0, 0.0, 0.0]

60000800 = [0.0, 5.0, 0.0]

20020003 = [5.0, 0.0, 0.0]

10260209 = [1156.2659900000001, 313.99282799999997, 155.018463]

80009000 = [10.0, 5.0, 0.0]

7 = [5.0, 5.0, 0.0]

3280311 = [1365.6743200000001, -371.22653200000002, 201.031464]

50050000 = [20.0, 0.0, 0.0]

90009000 = [15.0, 5.0, 0.0]

70000000 = [5.0, 5.0, 0.0]

3280502 = [1254.3385000000001, -142.613068, 180.20266699999999]

3280503 = [1270.5728799999999, -175.843582, 184.23608400000001]

3280504 = [1282.8614500000001, -201.004501, 187.21845999999999]

Wire dictionary:

10000000 = [10000000, 20000000, 70000000]

20000000 = [20000000, 30000000, 80000000]

30000000 = [30000000, 40000000, 90000000]

10000071 = [10000101, 20000022, 70000000, 60000055]

40000000 = [40000000, 50000000]

30000088 = [30000208, 40000002, 90005000, 80003000]

20000092 = [20000105, 30000004, 80004000, 71111167]

40000094 = [40000304, 50000071, 90000600]

'''

''' Data File:

Rect 1000007110000101200000227000000060000055

Rect 2000009220000105300000048000400071111167

Rect 3000008830000208400000029000500080003000

Tria 40000094400003045000007190000600

Pnt 100100123. 0. 0.

Pnt 200200035. 0. 0.

Pnt 3040000010. 0. 0.

Pnt 4000000015. 0. 0.

Pnt 5005000020. 0. 0.

Pnt 600008000. 5. 0.

Pnt 700000005. 5. 0.

Pnt 8000900010. 5. 0.

Pnt 9000900015. 5. 0.

Pnt 100100123. 0. 0.

Pnt 200200035. 0. 0.

Pnt 3040000010. 0. 0.

Pnt 4000000015. 0. 0.

Pnt 5005000020. 0. 0.

Pnt 600008000. 5. 0.

Pnt 700000005. 5. 0.

Pnt 8000900010. 5. 0.

Pnt 9000900015. 5. 0.

Rect 100000001000000020000000700000006

Rect 200000002000000030000000800000007

Rect 300000003000000040000000900000008

Tria 4000000040000000500000009

Pnt 1 0. 0. 0.

Pnt 2 5. 0. 0.

Pnt 3 10. 0. 0.

Pnt 4 15. 0. 0.

Pnt 5 20. 0. 0.

Pnt 6 0. 5. 0.

Pnt 7 5. 5. 0.

Pnt 8 10. 5. 0.

Pnt 9 15. 5. 0.

Pnt * 3280311 0 1.36567432E+03 -3.71226532E+02

* 2.01031464E+02 0

Pnt * 3280502 0 1.25433850E+03 -1.42613068E+02

* 1.80202667E+02 0

Pnt * 3280503 0 1.27057288E+03 -1.75843582E+02

* 1.84236084E+02 0

Pnt * 3280504 0 1.28286145E+03 -2.01004501E+02

* 1.87218460E+02 0

Pnt* 10260209 1156.26599 313.992828

* 155.018463

Pnt* 10270106 1097.15002 250.676315

* 140.789337

Pnt* 10270107 1115.47864 271.83374

* 144.698837

'''

>>> d = {100: [1,0,0],102: [0,1,0],202: [0,1,1]}

>>> dict(zip([str(x) for x in d.values()], d.keys()))

{'[0, 1, 1]': 202, '[1, 0, 0]': 100, '[0, 1, 0]': 102}

>>>

'''

>>> d = {100: [1,0,0],102: [0,1,0],202: [0,1,1]}

>>> dict(zip([tuple(x) for x in d.values()], d.keys()))

{(0, 1, 0): 102, (0, 1, 1): 202, (1, 0, 0): 100}

>>>

'''

>>> f = open(r'H:\TEMP\temsys\sample_points4.txt').readlines()

>>> in_pt = False

>>> ptDict = {}

>>> for line in f:

... lineList = [x.lower().strip() for x in line.strip().split(' ', 1) if x != '']

... if in_pnt:

... pt_data += [s.strip('* ') for s in lineList[1].split() if (s != '*' and s != '0')]

... in_pnt = False

... ptDict[convert_data(pt_data[1])] = [convert_data(x) for x in pt_data[2:]]

... elif 'pnt' in lineList and '*' in lineList[1]:

... in_pnt = True

... pt_data = [lineList[0], ] + [s.strip('* ') for s in lineList[1].split() if (s != '*' and s != '0')]

...

>>> ptDict

{3280504: [1282.8614500000001, -201.004501, 187.21845999999999], 3280503: [1270.5728799999999, -175.843582, 184.23608400000001], 3280502: [1254.3385000000001, -142.613068, 180.20266699999999], 3280311: [1365.6743200000001, -371.22653200000002, 201.031464]}

>>>

'''

>>> s = 'Pnt * 3280504 0 1.28286145E+03 -2.01004501E+02'

>>> re.findall('\d+\.\d+E\+\d+|\+\d+\.\d+E\+\d+|-\d+\.\d+E\+\d+|\d+\.\d+|\d+', s)

['3280504', '0', '1.28286145E+03', '-2.01004501E+02']

>>>

'''

"""

>>> patt = re.compile(r'''\d+\.\d+E\+\d+|

... \d+\.\d+E\+\d+|

... -\d+\.\d+E\+\d+|

... -\d+\.\d+E-\d+|

... \d+\.\d+E-\d+|

... \d+\.\d+|

... -\d+\.\d+|

... \d+''', re.X

... )

>>> patt

<_sre.SRE_Pattern object at 0x00DE68D0>

>>> s = 'Pnt * 3280311 0 +1.36567432E+03 -3.71226532E+02'

>>> re.findall(patt,s)

['3280311', '0', '1.36567432E+03', '-3.71226532E+02']

>>> dd = {}

>>> lst = re.findall(patt,s)

>>> dd[int(lst[0])] = [float(i) for i in lst[1:] if i != '0']

>>> dd

{3280311: [1365.6743200000001, -371.22653200000002]}

>>> s1 = '* 155.018463'

>>> lst1 = re.findall(patt,s)

>>> dd[int(lst[0])] = dd[int(lst[0])]+[float(i) for i in lst1 if i != '0']

>>> dd

{3280311: [1365.6743200000001, -371.22653200000002, 155.018463]}

>>>

"""

# ghostdog74

data = open('H:/TEMP/temsys/strdata.txt').read()

pat = re.compile("(\d+) THRU (\d+)",re.M|re.DOTALL)

for items in pat.findall(data):

print items

'''

>>> ('880', '897')

('1001343', '1001349')

('921', '932')

('1001323', '1001331')

('1001343', '1001349')

('1001359', '1001365')

…………………………………..

"""