deepomatic-client-python

Deepomatic API Client for Python.

This client have been made in order to help you integrating our services within your apps in python.

A demo code can be run in demo.py. It will explain you how to :

• Play with dbs.
• Make batched queries.
• Searching based on similarity into your dbs.
• Find a perfect match.
• Use our fashion detection API.

API Documentation

https://api.deepomatic.com/

Requirements

sudo pip install -r requirements.txt

Client

Initialize a client. Does not make any call to the server.

# You should find your appID and apiKey in your account on developpers.deepomatic.com
client = deepomatic.Client(appID, apiKey)

Tasks

Since most of server calls need some processing time, you will often receive a Task (json dictionary). Task are very simple to use.

Fields:

• id
• data
• status
• error
• date_created
• date_updated

When the task "status" is equal to "succes" you can find the request result in the "data" field.

Models

Point

    pt = deepomatic.Point(0,0)

Polygons

You can create a polygon with the Polygon class :

    poly = deepomatic.Polygon()

To add points to the polygon, just use the addPoint method :

    poly.addPoint(deepomatic.Point(1,2))

If you want to add multiple points at once, just use the addPoints methode:

    poly.addPoints([deepomatic.Point(1,2), deepomatic.Point(2,2), deepomatic.Point(1,1)])

For a polygon to be valid, it must have at least 3 points.

Bbox

The bbox class takes 2 points, min and max.

    bbox = deepomatic.Bbox(deepomatic.Point(1,2), deepomatic.Point(2,3))

Images

The class ImgsSend, allows you to create an image with multiple fields :

    imgs = deepomatic.ImgsSend(sourceType, source, polygon = None, bbox = None)

You can add images by using the addImg method with the same parameters:

    imgs.addImg(sourceType, source, polygon = None, bbox = None)

You can use url, base64 or file for the sourceType.

Batch

You can send multiple request type on a batch : PUT, POST, GET, DELETE.

batch = deepomatic.BatchObject(dbname)
batch.deleteObject(id)
batch.getObject(id)
batch.addObject(imgs, id = None)

To send the batch to the API, just use the client batchRequest() methode :

client.batchRequest(batch, true)

Database commands

The database is created when the first object is indexed

List all databases

response = client.getDBs()

Add an object

With url

imgs = deepomatic.ImgsSend("url", "my_url", polygon = None, bbox = None)
data = { "myData": "beautiful shoes" }
id = "myID"
# if id is not specified the server will generate an id returned it in the response
# if the id already exists on the server side, it will be overwritten
response = client.saveObject("db_name", imgs, data = data, id = id)

With file

imgs = deepomatic.ImgsSend("file", "my_file.jpg", polygon = None, bbox = None)
data = { "myData": "beautiful shoes" }
id = "myID"
# if id is not specified the server will generate an id returned it in the response
# if the id already exists on the server side, it will be overwritten
response = client.saveObject("db_name", imgs, data = data, id = id)

Get an object

response = client.getObject("db_name", "myObjectID")

Delete an object

# if wait = True it will wait until the task is completed
response = client.deleteObject("db_name", "myObjectID", wait = True)

Clear a database (delete all objects)

# if wait = True it will wait until the task is completed else, it will return the task id
response = client.clearDB("db_name", wait = True)

Delete a database (delete the db and all the objects)

# if wait = True it will wait until the task is completed
response = client.deleteDB("db_name", wait = True)

Search in a database

Perform a visual Search

# if wait = True it will wait until the task is completed else it will return the task id
img_url = {"url" : "url_img"}
response = client.search("db_name", img_url)

# you can also use file upload
img_file = {"file" : "path_to_file"}
response = client.search("db_name", img_file)

You can use url, base64 or file as source.

Count number of objects in a database

img = "url_img"
response = client.getCount("db_name")

Batch operations

requests = deepomatic.BatchObject(dbname)
url_base = "http://example.com/image_%d.jpg"
for i in range(10):
    # refer to https://api.deepomatic.com for the content of the requests
    imgs = deepomatic.ImgsSend(url, url_base % i)
    requests.addObject(imgs, name)
# if wait = True it will wait until the task is completed
response = client.batchRequest(requests, wait = True)

Fashion detection

# if wait = True it will wait until the task is completed else it will send you back a task id
response = client.detect("fashion", "img_url", wait = True)

Retrieve the task status

task = client.retrieveTask(task_id)

Bugs

support@deepomatic.com

clothes