To learn a new programming language with specific goals can speed up the learning. It’s a little guide book for such purpose. When acquiring a new language, it’s very nature to make some subconscious comparisons between the previous ones and the one you are learning. It’s helpful to follow the specific sub-goals to get familiar to the basic data strucutres, for-loop, string, I/O, so on and so forth.
To generate a markdown file from a jupyter notebook.ipynb, run:
jupyter nbconvert guide.ipynb --to guide.md
initial value, conversion between different types - upcast or downcast
for, while, do, if-else, continue, break - range, xrange
initialize, length, assign value, add, delete, insert, visit, index, slice, compare (ref, empty), sort
import packages
import os
import sys
import pickle
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import tensorflow as tf
%matplotlib inline
## string concat
s = ""
for i in range(5):
s += str(i) ## int -> string
print(s) ## print will auto-append a new line char '\n'
## tuple
a = s[0], s[-5], s[-2]
type(a) ## gives the type of a variable
01234
tuple
int(a[2]) + 1 ## string -> int
4
** How to create empty {list, tuple, dict, string}? **
b = ([], (), {}, 0, 0.2, '', True)
for i in range(len(b)):
print(type(b[i]))
<class 'list'>
<class 'tuple'>
<class 'dict'>
<class 'int'>
<class 'float'>
<class 'str'>
<class 'bool'>
** how to evaluate whether a list, tuple or dict, string is empty? **
for i in range(len(b)):
if not b[i]: ## special way to know whether it's empty
print(str(b[i]) + " is empty.")
else:
print(str(b[i]) + " is not empty.")
[] is empty.
() is empty.
{} is empty.
0 is empty.
0.2 is not empty.
is empty.
True is not empty.
** Useful numpy.linespace to generate a sequence with a fixed width **
# np.linspace(start, stop, num=50, endpoint=True, retstep=False, dtype=None)[source]
x = np.linspace(10, 100, num=10, endpoint=True)
for i in x:
print(int(i))
type(x)
10
20
30
40
50
60
70
80
90
100
numpy.ndarray
** Hashable in Python **
All of Python’s immutable built-in objects are hashable, while no mutable containers (such as lists or dictionaries) are. Objects which are instances of user-defined classes are hashable by default; they all compare unequal, and their hash value is their id().
s1 = "it's great"
s2 = "it's great" ## s2 = s1
print(hash(s1) == hash(s2))
False
t = {'i':hash('i'), 'u':hash('u')}
t
{'i': 433611154635708497, 'u': 3049833569042992159}
for k,v in t.items():
print(k, v)
i 433611154635708497
u 3049833569042992159
l = [1, 9, '2', True]
l[1:4:2]
[9, True]
l.reverse()
l
[1, 9, '2', True]
b = [2, 9, 0]
b.sort()
b.reverse()
b
[9, 2, 0]
print('it' in s1, 'we' in s1, s1.find('he'), s1.rfind('g'), s1.count('t'), s1.split(' '))
True False -1 5 2 ["it's", 'great']
s3 = " welcome to python"
## trim() function
print("before:" + s3)
print("after:" + s3.strip())
before: welcome to python
after:welcome to python
print('{1} and {0}'.format('spam', 'eggs'))
eggs and spam
source = "/Users/chjiang/GitHub/csc/soc-3-hardcase/m10n10-1.csv"
with open(source, 'r') as f:
for line in f:
print(line, end = "") ## without newline
10
0,c0
1,c1
2,c2
3,c3
4,c4
5,c5
6,c6
7,c7
8,c8
9,c9
10,10,10
1,7,9,4,2,1,0,6,5,8,3
1,5,2,7,1,4,9,3,8,0,6
1,5,2,7,8,6,0,3,4,9,1
1,9,7,6,2,5,8,4,1,0,3
1,3,8,4,2,6,5,9,7,0,1
1,0,4,3,5,7,2,8,6,9,1
1,9,7,5,0,2,4,1,6,8,3
1,3,2,6,1,5,7,4,8,9,0
1,7,0,3,9,6,4,2,1,5,8
1,5,4,8,7,9,2,1,3,6,0
** Reading lines and skip some of them **
with open(source, 'r') as f:
m = int(f.readline())
for _ in range(m + 1):
next(f) ## skip lines, f is an iterator
for line in f:
print(line, end = "") ## without newline
1,7,9,4,2,1,0,6,5,8,3
1,5,2,7,1,4,9,3,8,0,6
1,5,2,7,8,6,0,3,4,9,1
1,9,7,6,2,5,8,4,1,0,3
1,3,8,4,2,6,5,9,7,0,1
1,0,4,3,5,7,2,8,6,9,1
1,9,7,5,0,2,4,1,6,8,3
1,3,2,6,1,5,7,4,8,9,0
1,7,0,3,9,6,4,2,1,5,8
1,5,4,8,7,9,2,1,3,6,0
profile = {}
with open(source, 'r') as f:
m = int(f.readline())
for _ in range(m + 1):
next(f) ## skip lines, f is an iterator
for line in f:
fields = [int(c) for c in line.split(",")]
pref = tuple(fields[1:]) ## change list to tuple such that it's hashable
profile[pref] = fields[0] ## add item to dict
profile
{(0, 4, 3, 5, 7, 2, 8, 6, 9, 1): 1,
(3, 2, 6, 1, 5, 7, 4, 8, 9, 0): 1,
(3, 8, 4, 2, 6, 5, 9, 7, 0, 1): 1,
(5, 2, 7, 1, 4, 9, 3, 8, 0, 6): 1,
(5, 2, 7, 8, 6, 0, 3, 4, 9, 1): 1,
(5, 4, 8, 7, 9, 2, 1, 3, 6, 0): 1,
(7, 0, 3, 9, 6, 4, 2, 1, 5, 8): 1,
(7, 9, 4, 2, 1, 0, 6, 5, 8, 3): 1,
(9, 7, 5, 0, 2, 4, 1, 6, 8, 3): 1,
(9, 7, 6, 2, 5, 8, 4, 1, 0, 3): 1}
next(iter(profile.items())) ## iteratively get access to all items
((0, 4, 3, 5, 7, 2, 8, 6, 9, 1), 1)
profile
{(0, 4, 3, 5, 7, 2, 8, 6, 9, 1): 1,
(3, 2, 6, 1, 5, 7, 4, 8, 9, 0): 1,
(3, 8, 4, 2, 6, 5, 9, 7, 0, 1): 1,
(5, 2, 7, 1, 4, 9, 3, 8, 0, 6): 1,
(5, 2, 7, 8, 6, 0, 3, 4, 9, 1): 1,
(5, 4, 8, 7, 9, 2, 1, 3, 6, 0): 1,
(7, 0, 3, 9, 6, 4, 2, 1, 5, 8): 1,
(7, 9, 4, 2, 1, 0, 6, 5, 8, 3): 1,
(9, 7, 5, 0, 2, 4, 1, 6, 8, 3): 1,
(9, 7, 6, 2, 5, 8, 4, 1, 0, 3): 1}
a, b, c = "123"; d, e, f = [2, 'w', True]
c, e
('3', 'w')
ll = [2,1,9]
ll.append(43)
print(ll)
[2, 1, 9, 43]
import glob
import os
source = "./*"
for file in glob.glob(source):
print(os.path.basename(file)) ## glob.glob: absolute path of files -> os.path.basename: file name
import os
# file name
print(os.path.basename(__file__))
# absolute path
path = os.path.abspath(__file__)
print(path)
# directory name of ABSOLUTE PATH
print(os.path.dirname(path))
** write string to file, read lines from file, remove file **
source = "./temp.txt"
with open(source, "w+") as file:
file.write("hello\n")
file.write("what?\n")
with open(source, "r") as file:
for line in file:
print(line, end = "")
if os.path.exists(source):
# print(os.remove(source))
os.rename(source, source + "-2")
hello
what?
a = [1, 4, 9, 0, 9]
len(a)
5
** Tranform a list to a set **
set(a) ## sorted set
{0, 1, 4, 9}
len(a) == 5 and len(set(a)) == 4
True
** Set operations **
s1 = set(['a', 'c', 'd', 'b']) # list -> set
s2 = set(('b', 'c', 'e', 'b')) # tuple -> set
s3 = set(['b', 'c'])
s4 = set(('b', 'c', 'e'))
# union, intersection, difference, symmetric difference, subset, equal, true subset
(s1 | s2, s1 & s2, s1 - s2, s1 ^ s2, s3 <= s1, s4 == s2, s4 < s2, s4 <= s2)
({'a', 'b', 'c', 'd', 'e'},
{'b', 'c'},
{'a', 'd'},
{'a', 'd', 'e'},
True,
True,
False,
True)
# equivalent functions
print((s1.union(s2), s1.intersection(s2), s1.difference(s2), s1.symmetric_difference(s2)))
print((s3.issubset(s1), s4.isdisjoint(s2), s4.issubset(s2)))
({'c', 'a', 'e', 'd', 'b'}, {'b', 'c'}, {'a', 'd'}, {'e', 'd', 'a'})
(True, False, True)
## membership
'a' in s1, 'ag' in s2
(True, False)
** Priority Queue **
from queue import PriorityQueue
# if maxsize=0, the priority queue will be infinity
q1 = PriorityQueue(maxsize=2)
q1.put((1, "a"))
q1.put((0, "b"))
print(q1.empty(), q1.full(), q1.maxsize)
False True 2
# add elements into quque
q2 = PriorityQueue()
q2.put((1, "a"))
q2.put((0, "b"))
q2.put((9, 'e'))
q2.put((-1, 'e'))
print(q2.empty(), q2.full(), q2.maxsize)
False False 0
# gets acess to elements in queue, it also remove elements from queue
while not q2.empty():
print(q2.get())
(-1, 'e')
(0, 'b')
(1, 'a')
(9, 'e')
class State(object):
def __init__(self, val):
self.val = val
def __lt__(self, other):
return self.val[0] < other.val[0]
def __str__(self):
return self.val.__str__()
val1 = [0, 2, 4]
state1 = State(val1)
val2 = [90, 110]
state2 = State(val2)
val3 = [19, 100]
state3 = State(val3)
print("Before state3: " + str(state3))
q = PriorityQueue()
q.put(state1)
q.put(state2)
q.put(state3)
print("peel:" + str(q.get()))
val3[0] = 100
## the change affects the state, but does not changes the priority queue
print("After state3: " + str(state3))
print("peel:" + str(q.get()))
Before state3: [19, 100]
peel:[0, 2, 4]
After state3: [100, 100]
peel:[100, 100]
v = [0, 9, 0, 9, 2, 91, 2]
v.index(2, 5)
6
** Timing ** It’s an example to compute the $\pi$ value from a series approximation: $\pi = \sqrt{12} \sum\limits_{i=0}^\infty \frac{1}{(2 i + 1) 3^i}$
import time
import math
import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline
def compute_pi(max_i):
diff = []
value = 0.0
sgn = -1
fact = 1.0
for i in range(max_i):
sgn *= -1
value += sgn * fact / (2 * i + 1)
fact /= 3.0
diff.append(2 * math.sqrt(3) * value - math.pi)
plt.plot(np.array(range(max_i)), np.array(diff))
return 2 * math.sqrt(3) * value
begin = time.perf_counter()
max_i = 100
pi = compute_pi(max_i)
end = time.perf_counter()
print("PI = ", str(pi))
print("Elapsed: " + str((end - begin)) + "s")
PI = 3.141592653589794
Elapsed: 0.11220011999830604s
** argmin, argmax, get_rank **
def argmax(x):
"""
retrieval all positions with element = max value in x
:param x: list or tuple
:return: all positions with element = max element in x
"""
max_v = max(x)
ret = []
for i in range(len(x)):
if x[i] == max_v:
ret.append(i)
return ret
def argmin(x):
"""
retrieval all positions with element = min value in x
:param x: list or tuple
:return: all positions with element = min element in x
"""
min_v = min(x)
ret = []
for i in range(len(x)):
if x[i] == min_v:
ret.append(i)
return ret
def get_rank(x, ascend=True):
"""
ranking positions of elements in x in ascending or descending order
:param x: list or tuple
:param ascend: boolean
:return: ranking positions of elements in x
"""
inv = 1
if not ascend:
inv = -1
return sorted(range(len(x)), key=lambda i : inv * x[i])
x = [1,2,4,4,1,3]
get_rank(x)
[0, 4, 1, 5, 2, 3]
argmin(x)
[0, 4]
argmax(x)
[2, 3]
** init method is used when the class is called to initialize the instance, while the call method is called when the instance is called **
class CLS(object):
def __init__(self, a, b):
self.a = a
self.b = b
def __call__(self, a, b):
self.a = a
self.b = b
print((a,b))
cls = CLS(2, 1) # initialize a CLS instance
cls(6, 7) # call instance with parameters
(6, 7)
** difference between python2 and python3 in division operation between two integers **
a = 1
b = 2
a/b = 0 # python2
a/b = 0.5 # python3
To avoid possible issue, just multiply the numerator (number above the line in a fraction) by 1.0, with no need to change the denominator (number below the line in a fraction), i.e.
a * 1.0 / b = 0.5 # works for both python2 and python3
** Different ways to remove an item from a list **
lst = [8, 29, -1, 0, 89]
lst.remove(0) # remove 0 from lst
print("rmv 0:" + str(lst))
lst.pop() # remove from lst the last element
print("pop:" + str(lst))
lst.pop(1) # remove from lst the element with the index
print("pop(1): " + str(lst))
del lst[0] # remove from lst the first element
print("del lst[0]: " + str(lst))
rmv 0:[8, 29, -1, 89]
pop:[8, 29, -1]
pop(1): [8, -1]
del lst[0]: [-1]
** With ** block is used to close some stream to release the assigned resource, e.g. file stream when open a file, or Tensorflow Session, the stream will be closed automatically at the end of the with block.
with open('stv.csv', 'w+') as file:
file.write('stv is a popular voting rule.')
import Tensorflow as tf
x = tf.constant([[3., 3.]])
y = tf.constant([[2.],[2.]])
product = tf.matmul(x, y)
with tf.Session() as sess:
result = sess.run([product])
print(result)
** Difference between list append and extend **
a = [1, 2, 4]
a.append([5, 0])
print(a)
a.extend([8, 9])
print(a)
[1, 2, 4, [5, 0]]
[1, 2, 4, [5, 0], 8, 9]
** Evaluate whether a key exists in dict using in **
d = {(2, 4, 5) : 1, (2, 5) : 2}
(2, 5, 3) in d
False
** Merging of two dict **
e = {(1, 3) : 2}
d.update(e)
d
{(1, 3): 2, (2, 4, 5): 1, (2, 5): 2}
e = {(1, 2) : 4}
d
{(1, 3): 2, (2, 4, 5): 1, (2, 5): 2}
a = {}
'a' in a
False
**module import and init.py ** init.py file is trying to direct the Python importer to a module under a directory. If you want to include the module under certain directory, you have to give an init.py file, even the file is empty, otherwise, the module under the directory may fail to import.
source = "/users/chjiang/github/csc/temp.txt"
with open(source, "w+") as output:
output.write("lhel" + "\n")
output.write("lskdf" + "\n")
s = set()
s.add(1)
s.add(2)
s
{1, 2}
type(s)
set
sorted(s)
[1, 2]
** shell basic: kill background process nohup **
** serialization and deserialization using pickle **
import pickle
objs = []
while 1:
try:
objs.append(pickle.load(f))
except EOFError:
break
import pickle
file = "/users/chjiang/github/csc/temp.pkl"
profile = {"C" : 100, "D" : 90}
pickle.dump(profile, open(file, "wb")) # write in binary
prof = pickle.load(open(file, "rb"))
print(prof)
{'a': 100, 'b': 90}
#*** record log using logging ***
import logging
# create logger with name `HoRseHouR`
logger = logging.getLogger('HoRseHouR')
logger.setLevel(logging.DEBUG)
# create file handler which logs even debug messages
file = "/users/chjiang/github/csc/temp.log"
fh = logging.FileHandler(file)
fh.setLevel(logging.DEBUG)
# create formatter
formatter = logging.Formatter('%(asctime)s - %(name)s - %(levelname)s - %(message)s')
# add the formmatter to the handlers
fh.setFormatter(formatter)
# add the handlers to the logger
logger.addHandler(fh)
# add information to the logger
logger.info("I am riding and running")
logger.info("Look man, he fell off the horse. How hilarious! ;-)")
# open file and write its name and all lines
with open(file, "r") as output:
print(output.name)
print(output.readline())
print(output.readline())
print(output.readline())
# delete file
import os
file = "/users/chjiang/github/csc/temp.pkl"
os.path.exists(file)
# os.remove(file)
True
# basename
os.path.basename(file)
'temp.log'
import random
# random sample without replacement, shuffle
import random
a = [1, 3, 4, 20]
random.sample(a, 2)
random.shuffle(a)
a
[20, 4, 3, 1]
a = [1, 23, 4]
b = [1, 3, 4]
a == b
False
c = np.array(a) + np.array(b)
[sum(x)/2 for x in zip(a,b)]
[1.0, 13.0, 4.0]
zip(a,b)
<zip at 0x1073181c8>
a[1:3] = b
a
[1, 1, 3, 4]
b = [5,6]
a.extend([1,2,1,4])
sum([e == 1 for e in a])
4
def onehot(x, k=10):
"""
Encode input elements with a one-hot vector
:param x: nonnegative input elements
:param int k: length of the code
:return: onehot code of length k
:rtype: list
"""
code = [0] * k
for i in x:
code[i] = 1
return code
x = [0, 1, 4, 5]
onehot(x)
[1, 1, 0, 0, 1, 1, 0, 0, 0, 0]
a = {1:2, 1:4}
a
{1: 4}
print("{0:>2} {1:<30} {2:<2} {3:<10}".format(str([1]), str(x), str(x), str([0,9])))
print("{0:>2} {1:<30} {2:<2} {3:<10}".format(str([1]), str(x+[2]), str(x), str([1,9])))
[1] [0, 1, 4, 5] [0, 1, 4, 5] [0, 9]
[1] [0, 1, 4, 5, 2] [0, 1, 4, 5] [1, 9]
import sys
-sys.float_info.min < 0
True
# minimum float
-float('inf') < -sys.float_info.max
True
a, b, c = 1, 2, 0
not 2
False