1
00:00:08,160 --> 00:00:10,480
You'll recall that we import

2
00:00:10,480 --> 00:00:12,805
a library using
the import keyword.

3
00:00:12,805 --> 00:00:15,760
So just import PIL,
and we run this.

4
00:00:15,760 --> 00:00:19,600
Documentation is a big help
in learning a library.

5
00:00:19,600 --> 00:00:22,600
There exists standards that
make this process easier.

6
00:00:22,600 --> 00:00:24,960
For example, most of
the libraries let you

7
00:00:24,960 --> 00:00:27,825
check their version using
the version attribute.

8
00:00:27,825 --> 00:00:29,370
So we go PIL.

9
00:00:29,370 --> 00:00:31,330
and then double underscore.

10
00:00:31,330 --> 00:00:34,560
So this is called
dander version dander.

11
00:00:34,560 --> 00:00:39,205
We see the version of the one
that I'm using is 4.2.1.

12
00:00:39,205 --> 00:00:40,960
You might be using
a different version

13
00:00:40,960 --> 00:00:42,850
because we might
have upgraded it.

14
00:00:42,850 --> 00:00:46,490
Let's figure out how to
open an image with Pillow.

15
00:00:46,490 --> 00:00:49,555
Python provides some built-in
functions to help us

16
00:00:49,555 --> 00:00:50,905
understand the functions and

17
00:00:50,905 --> 00:00:53,530
objects which are
available in libraries.

18
00:00:53,530 --> 00:00:55,570
For instance, the help function

19
00:00:55,570 --> 00:00:57,160
when called on any object,

20
00:00:57,160 --> 00:01:00,310
returns the objects
built-in documentation.

21
00:01:00,310 --> 00:01:03,580
Let's try it with
our new library and module, PIL.

22
00:01:03,580 --> 00:01:07,915
So help PIL and this
renders nicely in line

23
00:01:07,915 --> 00:01:10,870
a bit of help file that
actually comes from

24
00:01:10,870 --> 00:01:14,545
the documentation of
the source code itself.

25
00:01:14,545 --> 00:01:16,690
This shows us that
there's a host of

26
00:01:16,690 --> 00:01:19,060
classes available to
us in the module,

27
00:01:19,060 --> 00:01:21,700
as well as version information
and even the file

28
00:01:21,700 --> 00:01:26,725
called dander init dander.py,

29
00:01:26,725 --> 00:01:29,300
which has the source code
for the module itself.

30
00:01:29,300 --> 00:01:31,370
We can look up
the source code for this

31
00:01:31,370 --> 00:01:33,575
in the Jupyter console
if we wanted to.

32
00:01:33,575 --> 00:01:36,350
These documentation standards
make it easy to

33
00:01:36,350 --> 00:01:40,030
poke around an
unexplored a library.

34
00:01:40,030 --> 00:01:43,215
Python also has
a function called dir,

35
00:01:43,215 --> 00:01:45,605
which will list
the contents of an object.

36
00:01:45,605 --> 00:01:48,530
This is especially useful
with modules where you

37
00:01:48,530 --> 00:01:51,650
might want to see which classes
you might interact with.

38
00:01:51,650 --> 00:01:54,820
Let's list the details
of the PIL module.

39
00:01:54,820 --> 00:01:57,500
So dir PIL, and we can see

40
00:01:57,500 --> 00:02:00,725
here a list comes back
with a bunch of strings.

41
00:02:00,725 --> 00:02:04,010
Most of these are intended
to be internal functions.

42
00:02:04,010 --> 00:02:06,635
So they've got dander
before and dander after,

43
00:02:06,635 --> 00:02:08,300
and that's just
double underscore,

44
00:02:08,300 --> 00:02:11,385
that's a fancy way in
Python that they say.

45
00:02:11,385 --> 00:02:13,820
We see that there's
a couple that,

46
00:02:13,820 --> 00:02:16,135
do not have a dander and

47
00:02:16,135 --> 00:02:19,025
are expected to be
used more generally.

48
00:02:19,025 --> 00:02:20,690
At the top of this list,

49
00:02:20,690 --> 00:02:22,115
there's something called Image.

50
00:02:22,115 --> 00:02:24,365
This sounds like it could
be interesting for us.

51
00:02:24,365 --> 00:02:25,970
So let's import it directly,

52
00:02:25,970 --> 00:02:27,685
and run the help command on it.

53
00:02:27,685 --> 00:02:32,590
So from PIL we'll Import
image, then help Image.

54
00:02:34,310 --> 00:02:37,170
Running help on
the Image tells us that

55
00:02:37,170 --> 00:02:39,875
this object is
the Image class wrapper.

56
00:02:39,875 --> 00:02:42,140
We see from the top
level documentation

57
00:02:42,140 --> 00:02:43,490
about the image object that

58
00:02:43,490 --> 00:02:45,560
there's hardly ever any reason

59
00:02:45,560 --> 00:02:47,765
to call the image
constructor directly,

60
00:02:47,765 --> 00:02:49,430
and they suggest that we use

61
00:02:49,430 --> 00:02:51,260
the open function and that's what

62
00:02:51,260 --> 00:02:54,090
we should be using to get images.

63
00:02:54,170 --> 00:02:57,040
Let's call help on
the open function

64
00:02:57,040 --> 00:02:58,540
to see what it's all about.

65
00:02:58,540 --> 00:03:00,100
Remember that since we want to

66
00:03:00,100 --> 00:03:02,004
pass in the function reference,

67
00:03:02,004 --> 00:03:03,940
and not run the function itself,

68
00:03:03,940 --> 00:03:07,060
we don't put parentheses
behind the function name.

69
00:03:07,060 --> 00:03:10,330
So help Image.open, we're passing

70
00:03:10,330 --> 00:03:11,950
an object but that object is

71
00:03:11,950 --> 00:03:15,470
actually a reference
to the function.

72
00:03:16,340 --> 00:03:19,295
So it looks like Image.open is

73
00:03:19,295 --> 00:03:21,280
a function that loads
an image from a file,

74
00:03:21,280 --> 00:03:23,920
and returns an instance
of the image class.

75
00:03:23,920 --> 00:03:25,325
Let's give it a try.

76
00:03:25,325 --> 00:03:27,160
In the read_only
directory there's

77
00:03:27,160 --> 00:03:28,780
an image I've provided which is

78
00:03:28,780 --> 00:03:30,310
from our Master's of Information

79
00:03:30,310 --> 00:03:32,175
program recruitment flyer.

80
00:03:32,175 --> 00:03:34,105
Let's try and load that now.

81
00:03:34,105 --> 00:03:36,655
So file we'll make it a string,

82
00:03:36,655 --> 00:03:39,595
read_only directory and then

83
00:03:39,595 --> 00:03:43,490
MSI_recruitment.gif
and we'll call

84
00:03:43,490 --> 00:03:47,045
image.open and pass it
this path to the file.

85
00:03:47,045 --> 00:03:48,950
That should return to
us an image object,

86
00:03:48,950 --> 00:03:51,320
which we're going to put it
into the image variable.

87
00:03:51,320 --> 00:03:54,190
Let's just print out this image.

88
00:03:54,190 --> 00:03:56,250
So we see it printed a

89
00:03:56,250 --> 00:03:59,840
PIL.GifImagePlugin.GifImageFile
and it gives

90
00:03:59,840 --> 00:04:01,775
us some other information there.

91
00:04:01,775 --> 00:04:04,250
So we see that this returns to us

92
00:04:04,250 --> 00:04:07,645
a kind of
PIL.GifImagePlugin.GifImageFile.

93
00:04:07,645 --> 00:04:10,360
At first this might
seem a bit confusing,

94
00:04:10,360 --> 00:04:11,450
because we were told by

95
00:04:11,450 --> 00:04:13,369
the docs that we
should be expecting

96
00:04:13,369 --> 00:04:16,715
a PIL.Image.Image object back.

97
00:04:16,715 --> 00:04:19,925
But this is actually
just object inheritance working.

98
00:04:19,925 --> 00:04:21,690
In fact, the object returned is

99
00:04:21,690 --> 00:04:24,385
both an Image and a GifImageFile.

100
00:04:24,385 --> 00:04:28,025
We can use the Python
inspect module to see this,

101
00:04:28,025 --> 00:04:31,640
as the getmro function
will return a list of all

102
00:04:31,640 --> 00:04:32,870
of the classes that are being

103
00:04:32,870 --> 00:04:35,100
inherited by a given object.

104
00:04:35,100 --> 00:04:38,405
Let's give it a try. So
we'll import inspect.

105
00:04:38,405 --> 00:04:40,120
Now this is not
a third party library,

106
00:04:40,120 --> 00:04:41,825
it comes with Python.

107
00:04:41,825 --> 00:04:43,640
Then we'll write a function.

108
00:04:43,640 --> 00:04:46,220
We'll just print
the type of the image.

109
00:04:46,220 --> 00:04:50,520
So a type will tell us
the type of the image,

110
00:04:51,400 --> 00:04:54,605
but we want to convert
that to a string.

111
00:04:54,605 --> 00:04:57,725
But then we're going
to call inspect.getmro

112
00:04:57,725 --> 00:04:58,790
and pass it the type

113
00:04:58,790 --> 00:05:00,440
of the image and see

114
00:05:00,440 --> 00:05:03,875
what that inheritance
chain looks like.

115
00:05:03,875 --> 00:05:06,845
So we'd see the result is

116
00:05:06,845 --> 00:05:10,205
a tuple that's returned
to us which is actually

117
00:05:10,205 --> 00:05:11,990
all of the different
objects that are

118
00:05:11,990 --> 00:05:13,860
being inherited from here with

119
00:05:13,860 --> 00:05:16,690
GifImagePlugin at the very

120
00:05:16,690 --> 00:05:19,420
usually bottom we would call
this of the inheritance,

121
00:05:19,420 --> 00:05:21,380
the most specific version

122
00:05:21,380 --> 00:05:23,795
up to an ImageFile
up to an Image,

123
00:05:23,795 --> 00:05:26,260
and then finally up to an object.

124
00:05:26,260 --> 00:05:28,620
Now that we're comfortable
with the object,

125
00:05:28,620 --> 00:05:30,465
how do we view the image?

126
00:05:30,465 --> 00:05:33,740
It turns out that the image
object has a show function.

127
00:05:33,740 --> 00:05:35,360
You can find this by looking at

128
00:05:35,360 --> 00:05:38,150
the properties of the object
if you wanted to,

129
00:05:38,150 --> 00:05:39,785
using the dir function.

130
00:05:39,785 --> 00:05:42,030
So we'll call image.show,

131
00:05:42,030 --> 00:05:46,295
and that didn't seem to
have the intended effect.

132
00:05:46,295 --> 00:05:47,930
The problem is
the image is stored

133
00:05:47,930 --> 00:05:50,195
remotely on Coursera's server,

134
00:05:50,195 --> 00:05:52,900
but show tries to show
it locally to you.

135
00:05:52,900 --> 00:05:55,670
So if the Coursera
server software it was running

136
00:05:55,670 --> 00:05:58,625
on someone's workstation in
Mountain View California,

137
00:05:58,625 --> 00:06:00,710
where Coursera has its offices,

138
00:06:00,710 --> 00:06:01,970
then you just popped up

139
00:06:01,970 --> 00:06:04,055
a picture of our
recruitment materials.

140
00:06:04,055 --> 00:06:06,410
Thanks for that. Instead though,

141
00:06:06,410 --> 00:06:09,475
we want to render the image
in the Jupyter notebook.

142
00:06:09,475 --> 00:06:10,940
It turns out Jupyter has

143
00:06:10,940 --> 00:06:13,175
a function which
can help with this.

144
00:06:13,175 --> 00:06:15,440
So from IPython.display,

145
00:06:15,440 --> 00:06:17,280
we'll talk a little
bit more about that,

146
00:06:17,280 --> 00:06:18,650
but IPython was one of

147
00:06:18,650 --> 00:06:22,250
the early terms for
Jupyter and it started as

148
00:06:22,250 --> 00:06:25,625
just an interactive
Python interpreter

149
00:06:25,625 --> 00:06:28,640
before moving into
a much larger project.

150
00:06:28,640 --> 00:06:30,740
We want to import the
display function,

151
00:06:30,740 --> 00:06:33,685
and then let's call display
and pass it the image.

152
00:06:33,685 --> 00:06:36,950
Okay. So there we see
our inline display of

153
00:06:36,950 --> 00:06:41,550
happy Masters of Science
in Information students.

154
00:06:42,730 --> 00:06:45,110
For those who would
like to understand

155
00:06:45,110 --> 00:06:46,235
this in more detail,

156
00:06:46,235 --> 00:06:49,010
the Jupyter environment is
running a special wrapper

157
00:06:49,010 --> 00:06:52,240
across the Python interpreter
called IPython.

158
00:06:52,240 --> 00:06:54,950
IPython allows
the kernel back end to

159
00:06:54,950 --> 00:06:56,945
communicate with
the browser front end

160
00:06:56,945 --> 00:06:58,595
among other things.

161
00:06:58,595 --> 00:07:01,550
The IPython package has
a display function,

162
00:07:01,550 --> 00:07:03,230
which can take objects and use

163
00:07:03,230 --> 00:07:06,755
custom formatters in order to
render them to the screen.

164
00:07:06,755 --> 00:07:10,250
There's a lot of different
formatters provided including

165
00:07:10,250 --> 00:07:11,270
ones which know how to handle

166
00:07:11,270 --> 00:07:13,565
image types and
different image types,

167
00:07:13,565 --> 00:07:15,785
and that's what we're using here.

168
00:07:15,785 --> 00:07:18,290
That's a quick overview
of how to read

169
00:07:18,290 --> 00:07:20,525
and display images using pillow.

170
00:07:20,525 --> 00:07:22,310
In the next lecture,
we're going to jump

171
00:07:22,310 --> 00:07:23,990
in a bit more into

172
00:07:23,990 --> 00:07:25,520
detail to understand how to

173
00:07:25,520 --> 00:07:28,500
use pillow to manipulate images.