##// END OF EJS Templates
Start describing notebooks for IPython architecture doc
Thomas Kluyver -
Show More
@@ -1,56 +1,71 b''
1 1 How IPython Works
2 2 =================
3 3
4 4 Terminal IPython
5 5 ----------------
6 6
7 7 When you type ``ipython``, you get the original IPython interface, running in
8 8 the terminal. It does something like this::
9 9
10 10 while True:
11 11 code = input(">>> ")
12 12 exec(code)
13 13
14 14 Of course, it's much more complicated, because it has to deal with multi-line
15 15 code, tab completion using :mod:`readline`, magic commands, and so on. But the
16 16 model is like that: prompt the user for some code, and when they've entered it,
17 17 exec it in the same process.
18 18
19 19 The IPython Kernel
20 20 ------------------
21 21
22 22 All the other interfaces—the Notebook, the Qt console, ``ipython console`` in
23 23 the terminal, and third party interfaces—use the IPython Kernel. This is a
24 24 separate process which is responsible for running user code, and things like
25 25 computing possible completions. Frontends communicate with it using JSON
26 26 messages sent over ZeroMQ sockets; the protocol they use is described in
27 27 :doc:`messaging`.
28 28
29 29 The core execution machinery for the kernel is shared with terminal IPython:
30 30
31 31 .. image:: figs/ipy_kernel_and_terminal.png
32 32
33 33 A kernel process can be connected to more than one frontend simultaneously. In
34 34 this case, the different frontends will have access to the same variables.
35 35
36 36 .. TODO: Diagram illustrating this?
37 37
38 38 This design was intended to allow easy development of different frontends based
39 39 on the same kernel, but it also made it possible to support new languages in the
40 40 same frontends, by developing kernels in those languages, and we are refining
41 41 IPython to make that more practical.
42 42
43 43 Today, there are two ways to develop a kernel for another language. Wrapper
44 44 kernels reuse the communications machinery from IPython, and implement only the
45 45 core execution part. Native kernels implement execution and communications in
46 46 the target language:
47 47
48 48 .. image:: figs/other_kernels.png
49 49
50 50 .. seealso::
51 51
52 52 :doc:`kernels`
53 53
54 54 :doc:`wrapperkernels`
55 55
56
56 Notebooks
57 ---------
58
59 The Notebook frontend does something extra. In addition to running your code, it
60 stores code and output, together with markdown notes, in an editable document
61 called a notebook. When you save it, this is sent from your browser to the
62 notebook server, which saves it on disk as a JSON file with a ``.ipynb``
63 extension.
64
65 .. TODO: Diagram of how these pieces fit together.
66
67 The notebook server, not the kernel, is responsible for saving and loading
68 notebooks, so you can edit notebooks even if you don't have the kernel for that
69 language—you just won't be able to run code. The kernel doesn't know anything
70 about the notebook document: it just gets sent cells of code to execute when the
71 user runs them.
General Comments 0
You need to be logged in to leave comments. Login now