Announce: squeeze-alexa - DIY secure Alexa integration

[pippin]

Oh, is latency a problem with MQTT?
I'd primarily thought websockets might be much simpler to integrate

[nickb]

squeeze alexa 1.1 was released a while ago.

Includes various fixes and minor improvements, plus improved guides and HOWTOs.

Also, the development version is now internationalised, with UK English and German as the two translations provided (help wanted...)

[Freddy]

Great to see an own thread about this, I totally missed that before.

I have a question.
I´m currently running NGINX for connecting secure to Home-Assistant (HTTP) and connecting to my MQTT broker (stream)
Alexa integration works fine with HA using HAASKA + let´s encrypt cert

I´m a bit confused about this:
"Nginx doesn't seem to support non-HTTP traffic over TLS."

So I need to install stunnel anyhow or should it be possible to use NGINX?

[nickb]

Great to see an own thread about this, I totally missed that before.

I have a question.
I'm currently running NGINX for connecting secure to Home-Assistant (HTTP) and connecting to my MQTT broker (stream)
Alexa integration works fine with HA using HAASKA + let´s encrypt cert

I'm a bit confused about this:
"Nginx doesn't seem to support non-HTTP traffic over TLS."

So I need to install stunnel anyhow or should it be possible to use NGINX?

Stunnel is recommended, or at least documented and tested in detailed.

Good news though - on searching just now it looks like nginx R6 supports SSL termination for TCP protocols , in theory at least The docs make it look quite easy (more modern than stunnel too), and it supports speed-increasing things like SSL session caching and SSL session tickets (should be fast enough already though).

You'd need a separate (secret) cert but I've listed the steps on creating this. If you have issues, try the Github issues, or if you get it working and fancy documenting...

[Aidem]

Hi nickb, sorry, I realised I was on the wrong thread and felt discussion should be switched this thread.
Thank you your reply.

In your installation guide, after clone the repo the guide has the sentence. “Note you will have to run a release process now to get the translations “
What does this mean?
In your reply to me on the other thread, again on this part of the set up you wrote ‘The codebase is primarily to build and deploy the Alexa skill though.’
Should I run some form of compiling process involving Python?
If so, what is the command line routine to do this?
This would explain why I could not find any src/settings.py file. There is no src directory.

Finally you were correct, my home set up will eventually be: RPi as NAS / home server (to install, say, stunnel etc) . I have used the RPi as a Home server since they were first introduced. A RPi is easily up to the task.
However, at present I am using a separate RPi for this project so as not to interfere with my home server’s daily routines. If I muck it up it is easy to start again.

[Freddy]

Stunnel is recommended, or at least documented and tested in detailed.

Good news though - on searching just now it looks like nginx R6 supports SSL termination for TCP protocols , in theory at least The docs make it look quite easy (more modern than stunnel too), and it supports speed-increasing things like SSL session caching and SSL session tickets (should be fast enough already though).

You'd need a separate (secret) cert but I've listed the steps on creating this. If you have issues, try the Github issues, or if you get it working and fancy documenting...

Thank you, I´ll try to find time to test it. I already got Stream configured in NGINX so it would be nice include this into it as well.
I´m not running NGINX Plus however.

[nickb]

Hi nickb, sorry, I realised I was on the wrong thread and felt discussion should be switched this thread.
Thank you your reply.

In your installation guide, after clone the repo the guide has the sentence. “Note you will have to run a release process now to get the translations “
What does this mean?

That was updated recently actually now that it's a multilingual app (currently British English and erm German German). It's aimed mainly at developer I guess but I'll try and clean that up, plus add some automation around this before the next release.

But if you're downloading v1.1 (as I'd recommend) - it predates this change and you should refer to that v1.1 README) (also included in the zip).

In your reply to me on the other thread, again on this part of the set up you wrote ‘The codebase is primarily to build and deploy the Alexa skill though.’
Should I run some form of compiling process involving Python?
If so, what is the command line routine to do this?

Python doesn't need (ahead-of-time) compilation so no. Either way the packaging and deployment to AWS is the harder bit, but all this is in the documentation - I've built scripts (or you can use third-party tooling, also detailed) to help build and deploy as necessary.

This would explain why I could not find any src/settings.py file. There is no src directory.

Sorry, yes, I meant squeezealexa/settings.py.

Finally you were correct, my home set up will eventually be: RPi as NAS / home server (to install, say, stunnel etc). I have used the RPi as a Home server since they were first introduced. A RPi is easily up to the task.
However, at present I am using a separate RPi for this project so as not to interfere with my home server’s daily routines. If I muck it up it is easy to start again.

Great. That should work fine then.

BTW best way to get detailed support is either via the squeeze-alexa chatroom or filing a Github issue - it's too hard on forums

[nickb]

Squeeze Alexa 1.2 is now out.

• The skill is now internationalised with en_GB as default, and German (de_DE). American English (en_US) is supported by default now too.
• Better timeout handling
• Errors with the skill can now (configurably) be reported back by voice / card. It's as English as possible, and due the earlier diagnostics work usually contains useful hints about what the problem might be e.g. firewall issues / certificate mismatches, etc.
• Various setup HOWTO improvements

[Aidem]

Squeeze Alexa 1.2 is now out.

• The skill is now internationalised with en_GB as default, and German (de_DE). American English (en_US) is supported by default now too.
• Better timeout handling
• Errors with the skill can now (configurably) be reported back by voice / card. It's as English as possible, and due the earlier diagnostics work usually contains useful hints about what the problem might be e.g. firewall issues / certificate mismatches, etc.
• Various setup HOWTO improvements

Thank you for continuing with your work on this project. I was just at the AWS stage of my installation when I noticed there had been newer release. I have now updated to v1.2 and it is now back to the joys of AWS.
Fingers crossed.

[nickb]

squeeze-alexa 1.3 is now out with various internal improvements. It's now Python 3.x only, for what it's worth.

Also... a 2.0 version is coming soon, with better networking support using switchable transports depending on what type of Internet connection you have (or what you prefer setting up).