Skip to content
Some features of this SDK are still in development. Consult with our [online support team](mailto:onlineteam@sinch.com) if you run into issues using this SDK in a production environment.

Sinch Python SDK for Conversation

The Sinch Python SDK allows you to quickly interact with the Conversation API from inside your Python applications. When using the Python SDK, the code representing requests and queries sent to and responses received from the Conversation API are structured similarly to those that are sent and received using the Conversation API.

The fastest way to get started with the SDK is to check out our getting started guides. There you'll find all the instructions necessary to download, install, set up, and start using the SDK.

Syntax

Note:

This guide describes the syntactical structure of the Python SDK for the Conversation API, including any differences that may exist between the API itself and the SDK. For a full reference on Conversation API calls and responses, see the Conversation API Reference.

The code sample below is an example of how to use the Python SDK to send an SMS text message. The Conversation API call that accomplishes the same task is displayed below for reference.

send-message.py
"""
Sinch Python Snippet

This snippet is available at https://github.com/sinch/sinch-sdk-python/tree/main/examples/snippets
"""

import os
from dotenv import load_dotenv
from sinch import SinchClient

load_dotenv()

sinch_client = SinchClient(
    project_id=os.environ.get("SINCH_PROJECT_ID") or "MY_PROJECT_ID",
    key_id=os.environ.get("SINCH_KEY_ID") or "MY_KEY_ID",
    key_secret=os.environ.get("SINCH_KEY_SECRET") or "MY_KEY_SECRET",
    conversation_region=os.environ.get("SINCH_CONVERSATION_REGION") or "MY_CONVERSATION_REGION"
)

# The ID of the Conversation App to send the message from
app_id = "CONVERSATION_APP_ID"
# The phone number of the recipient in E.164 format (e.g. +46701234567)
recipient_identities = [
    {
        "channel": "SMS",
        "identity": "RECIPIENT_PHONE_NUMBER"
    }
]

response = sinch_client.conversation.messages.send_text_message(
    app_id=app_id,
    text="[Python SDK: Conversation] Sample text message",
    recipient_identities=recipient_identities
)

print(f"Successfully sent text message.\n{response}")

This example highlights the following required to successfully make a Conversation API call using the Sinch Python SDK:

Client

When using the Sinch Python SDK, you initialize communication with the Sinch backend by initializing the Python SDK's main client class. This client allows you to access the the functionality of the Sinch Python SDK.

Initialization

To successfully initialize the Sinch client class, you must provide a valid access key ID and access key secret combination. You must also provide your Project ID. For example:


from sinch import SinchClient

sinch_client = SinchClient(
    key_id="YOUR_key_id",
    key_secret="YOUR_key_secret",
    project_id="YOUR_project_id"
)

Configuration

After intializing the client, you can modify the following Conversation properties using the configuration class:

PropertyDescription
conversation_regionThe geographical location in which the conversation server is located. Must set to either us or eu.
conversation_domainThe URL (excluding the region) of the conversation server. Do not change unless advised by your account manager.
templates_regionThe geographical location in which the templates server is located. Must set to either us or eu.
templates_domainThe URL (excluding the region) of the templates server. Do not change unless advised by your account manager.

You will mostly use the configuration class to detail the location of the server you'd like to interact with. For example:


sinch_client.configuration.conversation_region="us"

Conversation domain

The Sinch Python SDK organizes different functionalities in the Sinch product suite into domains. These domains are accessible through the client. For example, sinch_client.conversation.[endpoint_category].[method]. You can also create a domain-specific client from a general client. For example:


from sinch import SinchClient

sinch_client = SinchClient(key_id="YOUR_key_id", key_secret="YOUR_key_secret",
project_id="YOUR_project_id")

from sinch.domains.conversation import Conversation

conversation_client = Conversation(sinch_client)

Endpoint categories

In the Sinch Python SDK, Conversation API endpoints are accessible through the client (either a general client or a Conversation-specific client). The naming convention of the endpoint's representation in the SDK matches the API:

  • conversation.apps
  • conversation.capability
  • conversation.consents
  • conversation.contacts
  • conversation.conversations
  • conversation.event_destinations
  • conversation.events
  • conversation.messages
  • conversation.project_settings
  • conversation.templates
  • conversation.transcoding

To validate and parse incoming events, use conversation.sinch_events(). For more information, see Sinch Events.

For example:


app_id = "CONVERSATION_APP_ID"
recipient_identities = [
    {
        "channel": "SMS",
        "identity": "RECIPIENT_PHONE_NUMBER"
    }
]

response = sinch_client.conversation.messages.send_text_message(
    app_id=app_id,
    text="[Python SDK: Conversation] Sample text message",
    recipient_identities=recipient_identities
)

conversation.apps endpoint category

The conversation.apps category of the Python SDK corresponds to the apps endpoint. The mapping between the API operations and corresponding Python methods are described below:

conversation.capability endpoint category

The conversation.capability category of the Python SDK corresponds to the capability endpoint. The mapping between the API operations and corresponding Python methods are described below:

API operationSDK method
Capability lookuplookup

conversation.consents endpoint category

The conversation.consents category of the Python SDK corresponds to the consent endpoint. The mapping between the API operations and corresponding Python methods are described below:

API operationSDK method
Get identities from consent listlist_identities
Get audit records from consent listslist_audit_records

conversation.contacts endpoint category

The conversation.contacts category of the Python SDK corresponds to the contacts endpoint. The mapping between the API operations and corresponding Python methods are described below:

API operationSDK method
List contactslist
Create a contactcreate
Get a contactget
Update a contactupdate
Delete a contactdelete
Merge two contactsmerge_contact
Get channel profileget_channel_profile
List contact identity conflictslist_identity_conflicts

Additionally, there are two helper methods for getting a channel profile:

API operationSDK method
Get channel profile by contact IDget_channel_profile_by_contact_id
Get channel profile by channel identityget_channel_profile_by_channel_identity

conversation.conversations endpoint category

The conversation.conversations category of the Python SDK corresponds to the conversations endpoint. The mapping between the API operations and corresponding Python methods are described below:

conversation.event_destinations endpoint category

The conversation.event_destinations category of the Python SDK corresponds to the webhooks endpoint. The mapping between the API operations and corresponding Python methods are described below:

conversation.events endpoint category

The conversation.events category of the Python SDK corresponds to the events endpoint. The mapping between the API operations and corresponding Python methods are described below:

API operationSDK method
Send an eventsend

conversation.messages endpoint category

The conversation.messages category of the Python SDK corresponds to the messages endpoint. The mapping between the API operations and corresponding Python methods are described below:

API operationSDK method
Send a messagesend
Get a messageget
Delete a messagedelete
List messageslist
Update message metadataupdate
List messages by channel identitylist_last_messages_by_channel_identity

Additionally, there are several helper methods for sending messages of the different media types:

API operationSDK method
Send a card messagesend_card_message
Send a carousel messagesend_carousel_message
Send a choice messagesend_choice_message
Send a contact info messagesend_contact_info_message
Send a list messagesend_list_message
Send a location messagesend_location_message
Send a media messagesend_media_message
Send a template messagesend_template_message
Send a text messagesend_text_message

conversation.project_settings endpoint category

The conversation.project_settings category of the Python SDK corresponds to the project settings endpoint. The mapping between the API operations and corresponding Python methods are described below:

conversation.templates endpoint category

The conversation.templates category of the Python SDK corresponds to the Templates V2 endpoint. The mapping between the API operations and corresponding Python methods are described below:

conversation.transcoding endpoint category

The conversation.transcoding category of the Python SDK corresponds to the transcoding endpoint. The mapping between the API operations and corresponding Python methods are described below:

API operationSDK method
Transcode a messagetranscode_message

Sinch Events

The conversation.sinch_events() method returns a handler for the events that the Conversation API sends to your event destination. Pass the secret configured for your event destination to validate signatures:

OperationSDK method
Validate the event signaturevalidate_authentication_header
Parse an event into a typed modelparse_event

For example:


sinch_events_service = sinch_client.conversation.sinch_events("YOUR_callback_secret")

if not sinch_events_service.validate_authentication_header(
    headers=headers,
    json_payload=raw_body,
):
    # Reject the request
    ...

event = sinch_events_service.parse_event(raw_body, headers)

Request and query parameters

Requests and queries made using the Python SDK are similar to those made using the Conversation API. Many of the fields are named and structured similarly. In most cases, they are the same. For example, consider the representations of a Conversation API app ID below. One field is represented in JSON, and the other is using our Python SDK:

app_id = "{APP_ID}"

Note that the fields have the same name. Additionally, path parameters, request body parameters, and query parameters that are used in the API are all passed as arguments to the corresponding python method. For example, consider this example in which the get method of the messages category is invoked:


conversation_response = sinch_client.conversation.messages.get(
                    message_id="YOUR_message_id",
                    messages_source="CONVERSATION_SOURCE")

When using the Conversation API, message_id would be included as a path parameter, and messages_source would be included as a query parameter in the JSON payload. With the Python SDK, both parameters are included as arguments in the get method.

For update methods, the SDK builds the update_mask from the arguments you pass. Arguments you omit are left unchanged on the server. Passing None explicitly clears the field.

Field name differences

The SDK uses the same field names as the Conversation API, with one exception. from is a reserved word in Python, so the SDK uses from_:

API field nameSDK field name
fromfrom_

Nested objects

When making calls directly to the API, we use JSON objects, including (in some cases) nested JSON objects. When using the Python SDK, we use dictionaries instead of nested JSON objects. For example, consider the message objects below. One is represented in JSON, the other as a Python dictionary:

message = {
    "text_message": {
        "text": "Text message from Sinch Conversation API."
    }
}

Note that, in both cases, the text_message and text objects are structured in exactly the same way as they would be in a normal Python call to the Conversation API. When using the Python SDK, any argument that represents a nested JSON object will be represented as a Python dictionary at the top level, but the contents of that dictionary must be represented as JSON objects. This is also highlighted in the below example:


recipient={
    "identified_by" : {
        "channel_identities" : [
            {"identity":"RECIPIENT_number","channel" : "SMS"}
        ]
    }
}

Responses

Response fields match the API responses. They are delivered as typed Python objects (Pydantic models), with each field, including nested objects, represented as a property.