To match the API URLs, we should name our objects/files/etc. the same as our API URLs.

You can supply a friendly from in the existing from_email field. For example the following are both valid:

sp.transmissions.send(..., from_email='[email protected]', ...)

sp.transmissions.send(..., from_email='Test Email <[email protected]>', ...)

The _extract_recipients (see this line) method handles converting either of these into the proper format for the API. We just need to document it.

See https://www.sparkpost.com/api#/reference/recipient-lists for docs for all resources.

When you run make test it doesn't make it past the flake8 validation. If you bypass that and run venv/bin/activate; py.test --cov sparkpost test/, import errors are returned. We must get tests running and passing before moving forward.

Sending attachment isn't supported at this moment. We need to add that support for attachment in python (including in Django backend).

@bizob2828 reviewed the transmissions real quick - just a couple of thoughts/questions around the _translate_keys method:

• When the incoming values in kwargs fallback to None, how does that pass through to the API? I'm wondering if we should just never add them to the model if they are being passed on.
• What are your thoughts on removing email_rfc822? I really don't think we should encourage users to go this route. It's just not user-friendly.

Caveat to this issue: I maybe misreading the code and there is another way to do this that I am not seeing.

While working on the Java SDK I see this project and PHP SDK require the API Key to be set inside a checked in file. Example

I believe things like that should not be placed inside checked in code because it is too easy to accidentally checkin you key.

I would prefer to see a name something like "config.something.example" that gets checked in. Put "config.something" would be put in the .gitignore so ti is not checked in. The example file would have the place holder properties.

The above filenames should use something appropriate for the language. In Java it would probably be "config.properties.example".

When we have an email in the format Example Person <[email protected] supplied in the from_email we should automatically parse it into the longer form dict `from: { name: 'Example Person', email: '[email protected]`` format. This gets around some issues related to unicode and how our system parses the long form.

Tox is a testing tool that allows local testing of various combination of Python versions and dependencies. It is similar to running the Travis build locally, which can be very helpful as the number of version combinations grows.

I'll make a pull request to demonstrate its usage.

Using the Python SDK, a POST to /transmissions with the BODY.recipients set to an array of strings, shown in example below, returns a mixed 200 response.

recipients = ['[email protected]', '[email protected]']

Response:

{u'total_accepted_recipients': 1, u'id': u'30010207866232273', u'rcpt_to_errors': [{u'message': u'invalid data format/type', u'code': u'1300', u'description': u'address  is not JSON'}], u'total_rejected_recipients': 1}

The example code shows this as a valid use case:
https://github.com/SparkPost/python-sparkpost/blob/master/examples/transmissions/send_transmission.py#L7

Need to research if this is a valid bug (and resolve) or user error.

Using the django allauth package to do my auth.

Clicking the forget password and sending by email, I get below.
No idea how to debug. How do i event add stuff to the content.html?

sparkpost.exceptions.SparkPostAPIException
SparkPostAPIException: Call to https://api.sparkpost.com/api/v1/transmissions returned 422, errors:

    required field is missing: content.inline_images requires content.html

Using this works
EMAIL_HOST = 'smtp.sparkpostmail.com'
EMAIL_PORT = 587
EMAIL_HOST_USER = 'SMTP_Injection'
EMAIL_HOST_PASSWORD = '348b25528c6b64abc0b423307314292c108d8899'
EMAIL_USE_TLS = True

See https://www.sparkpost.com/api#/reference/sending-domains for docs for all resources.

I've been two days behind a bug in the Django Backend.

Some of the text-only emails were arriving in HTML and with the content of other totally different email. Only viewing the email in raw you can see the correct text content.

The problem was the backend was modifing the global settings dict.
#87 should fix this bug.

I don't see any option in the transmission api to schedule a transmission. I'm missing something?

https://www.sparkpost.com/api#/reference/message-events

Let's get our code coverage showing up in https://coveralls.io/.

The REST API Documentation has "campaign_id" listed as the property to use in all the human readable documentation.

Shouldn't the SparkPost Python SDK POST to /transmission use the same property name instead of "campaign" as you can see here:

https://github.com/SparkPost/python-sparkpost/blob/master/examples/transmissions/send_transmission.py#L26

• Implement following methods:
  • GET
  • POST
  • PUT
  • DELETE
• Add User-Agent header and set it to 'python-sparkpost' (add /version if possible)
• Add/Update any unit tests

A few items based on some feedback:

• Update docs to show example transmission with substitution data
• Update docs to explain precedence of substitution data - recipient > transmission > template
• Use transmissions instead of transmission

People may want to use the SDK to offer an easy "download as .csv" format for responses.

We should make that simple and bind it to the following SparkPost API services:

• Templates
• Recipient Lists
• Suppression Lists
• Sending Domains

https://www.sparkpost.com/api#/reference/webhooks

Include the tests and code to incorporate Suppression List in the SDK.

https://www.sparkpost.com/api#/reference/suppression-list

I have a weird issue, preventing me from sending emails out with sparkpost==1.1.0 and Django==1.8.12. I traced the steps, starting with a simple:

email = EmailMessage('TEST SUBJECT', 'TEST MSG', '[email protected]', ['[email protected]'])
email.send()

This causes to execute (.../python2.7/site-packages/sparkpost/transmissions.py", line 232):

results = self.request('POST', self.uri, data=json.dumps(payload))

The payload here (as per django_extensions) looks like this:

{'description': None, 'recipients': [{'address': {'email': '[email protected]'}}], 'campaign_id': None, 'return_path': '[email protected]', 'content': {'from': {'email': '[email protected]'}, 'attachments': [], 'text': 'TEST MSG', 'headers': {}, 'html': None, 'reply_to': None, 'use_draft_template': False, 'inline_images': [], 'template_id': None, 'subject': 'TEST SUBJECT'}, 'substitution_data': None, 'options': {'skip_suppression': None, 'open_tracking': None, 'start_time': None, 'inline_css': None, 'transactional': None, 'sandbox': None, 'click_tracking': None, 'ip_pool': None}, 'metadata': None}

Error returned:

SparkPostAPIException: Call to https://api.sparkpost.com/api/v1/transmissions returned 422, errors: required field is missing: content.inline_images requires content.html

However, when I substitute [] with None in payload at this stage, the message goes out.

Using pip > 1.4.x there is no longer support for pre-release packages (like the SparkPost Python SDK) unless you use the --pre option on the pip command.

Adding this issue from this question on Stackoverflow:
http://stackoverflow.com/questions/29871778/sparkpost-python-sdk-install-failing/29871893

Here's the reference I found that helped me resolve this issue:
http://stackoverflow.com/questions/18230956/could-not-find-a-version-that-satisfies-the-requirement-pytz

Right now, we have to do the following:

sp.transmission.create(
    recipients=[{'address': {'email': '[email protected]'}}],
    description='test',
    campaign='rich-testing',
    metadata={},
    substitution_data={},
    reply_to='[email protected]',
    subject='test messsage',
    envelope_from='[email protected]',
    text='test message'
)

The recipients part seems especially over-complicated to me. Maybe just allow an array of addrs but also support the more complex object.

Additionally, we should change envelope_from to from as this field is actually the From header (aka Friendly From).

It is confusing for developers to use a flattened structure of their request payloads, especially when the property names vary from the API documentation on SparkPost.com/api.

Use a idiomatic structure passthrough that maps to the API docs for all first class citizens of the SparkPost API.

The current Transmissions examples use sp.transmission and should be changed to sp.transmissions

Need to change this to say it's a list of dicts.

• Transmission is failing to autoload after name change
• Templates are missing from API tree

• Create docs in /docs
• Set up auto-build on readthedocs.org

there are some files that point to the old www.sparkpost.com/docs - need to update to use the www.sparkpost.com/api/whatever.

When running the following code

    sub_data = {'content': html}

    result = sparkpost_client.templates.preview(
        template_id='newsletter',
        substitution_data=sub_data,
        draft=False
    )

I get the following error:

required field is missing: field 'substitution_data' is required

I think it may be because, in templates.py, line 181:

data=json.dumps(substitution_data))

You can't just simply dump the json, need to append it to a new dict with the "substitution_data" key before doing that.

I'm fairly sure this is the issue because if i change my code like so:

    sub_data = {'substitution_data': {'content': html}}

    result = sparkpost_client.templates.preview(
        template_id=config.item['sparkpost_template'],
        substitution_data=sub_data,
        draft=False
    )

It works.

Examples that are in there were quick and dirty and untested. Build a set of examples, similar to the others that exercise all the capabilities

Currently these are not supported.

Once these are added in the python library, we'll have to add the support those in the upcoming Django backend too.

Engagement tracking flags always get set on transmissions no matter if they are sent or not. This prevents the ability to fall back to the settings set on a saved template for these values.

When running make test it would be nice to see the coverage report in the terminal. That would inform the user if they needed to open the html report for more detail.

Hey @richleland,

So I have my environment setup, and I tried to use the following command to install the SDK into a Django app:

pip install -e <path/to/python-sparkpost/>

Here's the output I'm receiving (forgive me if this is simple...my Python hasn't been touched since 1.x versions). It very well could be an environment issue, and I'm going to keep on digging, but I thought I'd share what I hit.

Here's some environment specs:
Python Version: 2.7.6
Django Version: 1.8
Latest PIP installed

Obtaining file:///Users/benjamin/GitForks/python-sparkpost
Traceback (most recent call last):
File "", line 20, in
File "/Users/benjamin/GitForks/python-sparkpost/setup.py", line 1, in
import sparkpost
File "sparkpost/init.py", line 4, in
from .metrics import Metrics
File "sparkpost/metrics.py", line 1, in
from .base import Resource
File "sparkpost/base.py", line 1, in
import requests
ImportError: No module named requests
Complete output from command python setup.py egg_info:
Traceback (most recent call last):

  File "<string>", line 20, in <module>

  File "/Users/benjamin/GitForks/python-sparkpost/setup.py", line 1, in <module>

    import sparkpost

  File "sparkpost/__init__.py", line 4, in <module>

    from .metrics import Metrics

  File "sparkpost/metrics.py", line 1, in <module>

    from .base import Resource

  File "sparkpost/base.py", line 1, in <module>

    import requests

ImportError: No module named requests

----------------------------------------
Command "python setup.py egg_info" failed with error code 1 in /Users/benjamin/GitForks/python-sparkpost

See https://www.sparkpost.com/api#/reference/templates for docs for all resources.

Install tox, create tox.ini, set up to run in multiple environments including Python3

We need to add some options that are available via API:

• inline_images
• ip_pool

The metrics API only includes two very basic classes right now - Domains and Campaigns. Let's get the rest of the metrics functionality set up.

See https://www.sparkpost.com/docs/metrics-api for docs for all resources.

• See https://help.github.com/articles/setting-guidelines-for-repository-contributors/ for how this works with Github
• See https://github.com/atom/atom/blob/master/CONTRIBUTING.md for an example

We currently have tox tests to run against Python 2.7, 3.4, and 3.5, but Travis only runs against 2.7 and 3.4. Need to look into adding 3.5 to the list of Pythons Travis runs tests against.