Jastec France is a company that provides image recognition as a service. This Python module is a client that allows you to access the HTTP API to perform image recognition tasks. For more information on image recognition, please visit http://www.jastec.fr.
Note that this client cannot be used without a valid application key. Please contact our support department to get your application key today (support@jastec.fr).
The client comes in the form of a multiplatform python package. Although the package has been tested on linux platforms only, it should run fine under Windows.
Note that python 2.6.+ or later is required. To know if python is already installed or to check the version use the following command:
python --versionIn first, install virtualenv by typing the folling command in a terminal:
sudo apt-get install virtualenvThen, create and activate your dev environment in the project folder.
cd ltuengine-python-client
virtualenv env
source env/bin/activateThe package can be installed along with dependencies by running:
python setup.py installYou can check everything is fine by running the unit tests:
./ltu/engine/testsunit.pyAdding images to the application is done through the add_image() function of a ModifyClient instance:
my_application_key = "replace by your own key"
from ltu.engine.client import ModifyClient
modify_client = ModifyClient(my_application_key)
print(modify_client.add_image("my_image_id", "/home/user/image.jpg"))Once you have at least one image in your application, you can start making search queries by using the search_image() function of a QueryClient instance:
from ltu.engine.client import QueryClient
query_client = QueryClient(my_application_key)
print(query_client.search_image("/home/user/image.jpg"))For advanced usage, please consult the docstrings for each function.
An example is provided to quickly test the ADD, SEARCH and DELETE API functions. You have to execute the file cli.py by specifing these parameters: Positional arguments: ACTION: add, search, delete or bench(test the 3 functions in one action) APPLICATION_KEY INPUT_DIR
Optional arguments: --help show this help message and exit --host HOST, -h HOST (default: None) --nb-threads NB_THREADS, -n NB_THREADS (default: 1) --force force actions already processed on a file
Result of queries are saved in a Json file in the folder "out_result". Every actions already performed on a file won't be executed a second time (except in bench mode) unless the parameter --force is specified.
Here, an example for the ADD feature. A folder containing 5 images is available in this client: ./data
python ./ltu/engine/cli.py add application_key --force ./data -n 4If the parameter host is not specified, the script will hit the LTU OnDemand API. Otherwise, you can specify your own server by adding the domain name or the IP address after --host. Don't forget to specify http or https.
python ./ltu/engine/cli.py add application_key ./data --host http://10.5.20.56 -n 4
python ./ltu/engine/cli.py add application_key ./data --host http://mydomainname.com -n 4This software is licensed under the terms of the Apache License 2.0. In particular, you are free to distribute it, modify it and to distribute modified versions as long as you include the attached NOTICE file with your software. Read the attached LICENSE file for more information.
Feel free to use any of the resources provided in this client.
EMAIL: support@jastec.fr