alexey.zholtikov/mozc-devices
1
0
Fork 0
mirror of https://github.com/google/mozc-devices.git synced 2025-11-08 16:53:28 +03:00
Code Issues Packages Projects Releases Wiki Activity

Update README.rst

Browse Source 
- Explicitly install nazoru-input from PyPi.
- Remove unnecessary indents
- apt install for necessary packages (tested on a fresh Debian 9 image on GCE).
This commit is contained in:
Makoto Shimazu
2018-04-01 14:46:31 +09:00
parent 77453137ed
commit 15ea07071c
2 changed files with 80 additions and 61 deletions
Download Patch File Download Diff File Expand all files Collapse all files

139
mozc-nazoru/README.rst
View File

@@ -12,15 +12,31 @@ character.
Software Usage Software Usage
-------------- --------------
Input Characters Install and Try
~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~
Please make sure your local environemnt has packages as follows.
For Ubuntu/Debian:
.. code:: shell .. code:: shell
$ pip install . $ sudo apt install libffi-dev libcairo2 # For Ubuntu/Debian
$ nazoru-input
By running the commands above, you can make your own machine into an For macOS:
.. code:: shell
$ brew install cairo # For macOS
After that, you can try nazoru-input after installing it from pip.
.. code:: shell
$ pip install nazoru-input
$ nazoru-input
Using ``nazoru-input``, You can make your own machine into an
input device which accepts scribbles on the connected keyboard and send input device which accepts scribbles on the connected keyboard and send
characters via bluetooth. At the beginning, this script scans characters via bluetooth. At the beginning, this script scans
connected keyboards and starts listening to inputs from one of the connected keyboards and starts listening to inputs from one of the
@@ -28,20 +44,24 @@ keyboards. Then it translates a sequence of keydowns into a predicted
character considering pressed timings, and send the character to the character considering pressed timings, and send the character to the
target device paired by bluetooth. target device paired by bluetooth.
If you want to try it for development, you can use ``-e`` option. If you want to try it for development, you can download the source from
github and use ``-e`` option.
.. code:: shell .. code:: shell
$ pip install -e . $ sudo apt install libffi-dev libcairo2 # For Ubuntu/Debian
$ nazoru-input $ git clone https://github.com/google/mozc-devices.git
$ cd mozc-devices/mozc-nazoru
$ pip install -e .
$ nazoru-input
Training Model Training Model
~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~
.. code:: shell .. code:: shell
$ pip install . $ curl -LO https://github.com/google/mozc-devices/raw/master/mozc-nazoru/data/strokes.zip
$ nazoru-training ./data/strokes.zip $ nazoru-training ./strokes.zip
We have a script to generate a trained model which recognizes input We have a script to generate a trained model which recognizes input
characters from scribbles. This script renders input stroke data into characters from scribbles. This script renders input stroke data into
@@ -51,7 +71,7 @@ neural network model and the optimizer tunes the model to fit the data.
Once the training is done, the script outputs the trained graph, which Once the training is done, the script outputs the trained graph, which
you can use for your own device. In the case where you install you can use for your own device. In the case where you install
``nazoru-training`` from pip, you can find ``strokes.zip`` at here: ``nazoru-training`` from pip, you can find ``strokes.zip`` at here:
https://github.com/google/mozc-devices/mozc-nazoru/data/strokes.zip https://github.com/google/mozc-devices/blob/master/mozc-nazoru/data/strokes.zip
You can change some configurations by passing command line flags (e.g. You can change some configurations by passing command line flags (e.g.
path to the input/output files, hyper-parameters). Run path to the input/output files, hyper-parameters). Run
@@ -79,72 +99,73 @@ Raspberry Pi Setup
**Step 0 - Prepare your Raspberry Pi** **Step 0 - Prepare your Raspberry Pi**
Please prepare your Raspberry Pi, SD card initialized by RASPBIAN Please prepare your Raspberry Pi, SD card initialized by RASPBIAN
image, and RN42 module. Connect your Raspberry Pi with RN42 as the image, and RN42 module. Connect your Raspberry Pi with RN42 as the
schematic shows. Please make sure you can have access to the internet schematic shows. Please make sure you can have access to the internet
and also it has enough disk space to install packages on the following and also it has enough disk space to install packages on the following
steps. steps.
**Step 1 - Setup UART to RN42** **Step 1 - Setup UART to RN42**
If you try it on Raspberry Pi Zero W or Raspberry Pi 3, you need to If you try it on Raspberry Pi Zero W or Raspberry Pi 3, you need to
have additional settings for the serial communication because they have additional settings for the serial communication because they
equipped a wireless module connected by the UART. See details at `an equipped a wireless module connected by the UART. See details at `an
official document official document
<https://www.raspberrypi.org/documentation/configuration/uart.md>`_. <https://www.raspberrypi.org/documentation/configuration/uart.md>`_.
In short, you need to add ``enable_uart=1`` to ``/boot/config.txt`` on In short, you need to add ``enable_uart=1`` to ``/boot/config.txt`` on
your Raspberry Pi. your Raspberry Pi.
**Step 2 - Initial setup for RN42** **Step 2 - Initial setup for RN42**
You need to write your initial setup to RN42. At first, install screen You need to write your initial setup to RN42. At first, install screen
and open ``/dev/serial0`` for configuration. and open ``/dev/serial0`` for configuration.
.. code:: shell .. code:: shell
$ sudo apt install screen $ sudo apt install screen
$ sudo screen /dev/serial0 115200 $ sudo screen /dev/serial0 115200
After that, please type the following commands. Note that you need to After that, please type the following commands. Note that you need to
type ENTER after input commands. For example, please type ``$$$`` type ENTER after input commands. For example, please type ``$$$``
and ENTER to execute ``$$$`` command. and ENTER to execute ``$$$`` command.
1. ``$$$`` : Get into the command mode. The green LED will blink 1. ``$$$`` : Get into the command mode. The green LED will blink
faster. faster.
2. ``+`` : You can see what you type. 2. ``+`` : You can see what you type.
3. ``SD,0540`` : Set the device class to keyboard. 3. ``SD,0540`` : Set the device class to keyboard.
4. ``S~,6`` : Set the profile to HID. 4. ``S~,6`` : Set the profile to HID.
5. ``SH,0200`` : Set the HID flag to keyboard. 5. ``SH,0200`` : Set the HID flag to keyboard.
6. ``SN,nazoru-input`` : Set the device name as nazoru-input. You 6. ``SN,nazoru-input`` : Set the device name as nazoru-input. You
can name it as you want. can name it as you want.
7. ``R,1`` : Reboot RN42. 7. ``R,1`` : Reboot RN42.
You can quit the screen by ``C-a k``. You can quit the screen by ``C-a k``.
**Step 3 - Download and install nazoru-input** **Step 3 - Download and install nazoru-input**
We provide a service file at ``data/nazoru.service`` to launch We provide a service file at ``data/nazoru.service`` to launch
``nazoru-input`` when booting. You can install it by uncomment ``nazoru-input`` when booting. You can install it by uncomment
``data_files`` entry in ``setup.py``. Also, before installing this ``data_files`` entry in ``setup.py``. Also, before installing this
package, We'd strongly recommend you to install some package from apt package, We'd strongly recommend you to install some package from apt
repository as follows, so that you can install pre-built packages. repository as follows, so that you can install pre-built packages.
.. code:: shell .. code:: shell
$ sudo apt install git python-pip python-numpy python-cairocffi \ $ sudo apt install git python-pip python-numpy python-cairocffi \
python-h5py python-imaging python-scipy libblas-dev liblapack-dev \ python-h5py python-imaging python-scipy libblas-dev liblapack-dev \
python-dev libatlas-base-dev gfortran python-setuptools \ python-dev libatlas-base-dev gfortran python-setuptools \
python-html5lib python-html5lib
$ sudo pip install http://ci.tensorflow.org/view/Nightly/job/nightly-pi-zero/219/artifact/output-artifacts/tensorflow-1.6.0-cp27-none-any.whl $ sudo pip install http://ci.tensorflow.org/view/Nightly/job/nightly-pi-zero/219/artifact/output-artifacts/tensorflow-1.6.0-cp27-none-any.whl
$ git clone https://github.com/google/mozc-devices $ git clone https://github.com/google/mozc-devices
$ cd mozc-devices/mozc-nazoru $ cd mozc-devices/mozc-nazoru
$ sudo pip install . # If you want to develop nazoru-input, please use 'pip install -e .' instead. $ vi setup.py # Remove '#' for data_files to install nazoru.service.
$ sudo pip install . # If you want to develop nazoru-input, please use 'sudo pip install -e .' instead.
**Step 4 - Enjoy!** **Step 4 - Enjoy!**
.. code:: shell .. code:: shell
$ sudo nazoru-input # If you miss sudo, nazoru-input may use a DummyBluetooth object. $ sudo nazoru-input # If you miss sudo, nazoru-input may use a DummyBluetooth object.
Training Data Format Training Data Format
-------------------- --------------------
@@ -233,12 +254,10 @@ Authors
------- -------
Machine Learning: Machine Learning:
Shuhei Iitsuka <tushuhei@google.com>
Shuhei Iitsuka <tushuhei@google.com>
Hardwares, system setups: Hardwares, system setups:
Makoto Shimazu <shimazu@google.com>
Makoto Shimazu <shimazu@google.com>
License License
------- -------

2
mozc-nazoru/setup.py
View File

@@ -24,7 +24,7 @@ def read_file(name):
setup( setup(
name='nazoru-input', name='nazoru-input',
version='0.1.1', version='0.1.2',
author='Makoto Shimazu', author='Makoto Shimazu',
author_email='shimazu@google.com', author_email='shimazu@google.com',
url='https://landing.google.com/tegaki', url='https://landing.google.com/tegaki',