Commissioned Plugin Work -- Seeking help

[Lee Prevost]

I'm interested in commissioning a small development project as follows. The project would have two main deliverables:

1. SignalK-to-influxdb updade -- Update plugin to support Influx new version 2 java client (see below).   New plug in will be submitted to SignalK github as a PR and release as open source.  gh repo
2. Extend new plugin as follows (for private use-- rights to commissioner of the work) -- 
   
   1. Allow for multiple subscriptions (ie. 2-3) on varying time intervals such as "high frequency (200ms), mid-freq (1 sec), and low freq (15 min).
   2. Allow for user defined multiple points (ie. navigation sog, cog, angle of turn) via list or wildcard sequence to be written as a single time record into Influx (ie. timestamp, field name 1, field value 1, field name 2, field value 2,  ...  3, tag key 1, tag value 1)

If interested, please contact me via email: l...@prevost.net.  Provide questions, some information about your background and experience with the project.  

Discussion from gh issues on upgrade 

Lee

[Bradford Harley]

Lee--Did you get any traction on this?   I would be willing to chip in if needed for 1.

As an aside, assuming I understand correctly, couldn't you do 2a. via a continuous query in the influxdb and doesn't 2b. kind of conflict with how the data flows on the NMEA bus?  Every datapoint will have a unique timestamp, so by combining them you'll actually have to change the time.   You could also manually do this in the influxdb if you accept the loss of time resolution.

Good luck and keep us updated!

--

You received this message because you are subscribed to the Google Groups "Signal K" group.

To unsubscribe from this group and stop receiving emails from it, send an email to signalk+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/signalk/113e9439-72c7-4290-9cb5-963eb6162aa3n%40googlegroups.com.

[Lee Prevost]

I have gotten interest and am discussing that on sidebar with some excellent options for how to advance the plug in.  That sidebar has resulted in raising some of your same questions.   Thanks for agreeing to chip in on 1.   

I agree with both of your other points.  

I think the term continuous query has gone away in v2 and is now called ‘tasks’.  I don’t understand what is different.  

My main driver for the combination on 2 was the ability to do math across datapoints.  As an example, the derivative of temp divided by speed to get temp change per meter traveled.    I agree it could be done with resampling without too much of a loss in resolution. 

To your point about the time stamp being driven by underlying nmea time stamp, I’m realizing that I don’t fully understand the current time sample rates of the points I’m interested in so am going to dig in to understand that.  

Sent from my iPad

On Aug 16, 2022, at 1:59 PM, Bradford Harley <bradfor...@gmail.com> wrote:



You received this message because you are subscribed to a topic in the Google Groups "Signal K" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/signalk/osR2vHvCnkM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to signalk+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/signalk/e97c5f20-f772-4994-bf8e-4f6f86f23f1f%40www.fastmail.com.

[l...@prevost.net]

I believe this has been announced elsewhere but wanted to get back to this and to Brad Harley’s offer to chip in with me on this:

 

 

I have been working in background with Teppo Kurki, the original author of the SignalK-to-influxdb plug-in to upgrade it to support Influx OSS version > 2 and am reporting the results with a plan to make a contribution to the Open Collective.

 

Brad, would welcome you and I having a sidebar to coordinate that.

 

Here is an update on the project as I see it:

 

1. Teppo has rewritten the plug-in from the ground up and its better than it was before.   https://github.com/tkurki/signalk-to-influxdb2
2. I’ve been running tests on it and it works great.  Its currently publishing my N2k data at native bus refresh rates directly to an influx db version 2.3 database. 
3. Queries work better (you can do math across measurements and joins now).   And also the built in UI may be good enough to avoid needing Grafana for some use cases.
4. Have put in a request for programmatically identifying “self” mmsi so as to avoid having to pass keys between SignalK and Influx. 
5. I’m planning on making a donation to Open Collective per Teppo’s request:
1. Open Collective -- The use of the donations is completely transparent and can be viewed by each of you at https://opencollective.com/signalk (the same address can also be used for donations).
2. According to 9/21 slack update, 37 people have contributed.

 

Much thanks to Teppo for another solid contribution to the community.

 

Brad, feel free to reach out to me at l...@prevost.net or on the slack signal-k channel.

 

Thanks,

 

Lee Prevost