On Friday, May 8, 2020 at 11:39:14 AM UTC-7, Greg Troxel wrote: > > Do you think there are some schema columns which can be written only by > drivers, and other columns which are the domain of services? Or is it > simply a notion of main vs not main? >
No and no. The database doesn't know or care how the fields got populated. - Some fields are populated by telegraf subscribing to MQTT topic that the weewx MQTT extension publishes - Some fields are written by arduino doing a HTTP POST over the network - Some fields are written by python influxdb calls from a standlone program I wrote > > > If you want infinite flexibility in database element names, just go with > > influxdb and get it over with...but with great flexibility comes great > > confusion sometimes. At some point you'll need to write your skins and > > cheetah is not so flexible. You're going to eventually need a mapping > of > > your database element names somewhere in the actual code that runs. > > when you talk about influxdb, do you mean still using weewx to > > get data from station > put into influxdb > > get from influxdb and make graphs, as in skins > > ? > > I use influxdb to do data aggregation and comparison locally on the LAN, comparing a lot of sensors from multiple vendors. In doing that, weewx does the talking to the Davis console hardware since I have a serial datalogger. If I had a tcp/ip enabled datalogger or equivalent, I wouldn't need weewx for that for most things, although weewx's automatic catchup feature to read the datalogger menu is super slick and very useful if we have power issues or the like. Influxdb is just a time-series database. It requires something else to do graphs+dashboards. I use 'grafana' for that. Other options exist. > - I have multiple weewx stations feeding influxdb > > - a VP2 feeding it via MQTT > > - a WeatherFlow SmartWeather station feeding it via some python > code > > I wrote (one 'Air' and two 'Sky' sensor suites) > > - a WeatherFlow Tempest station feeding it via the same python > code I > > wrote > > - I have multiple arduino temperature sensor rigs feeding influxdb > as > > well > > - each write to influxdb individually over the network > > - the dashboard is grafana which picks the right data elements from > > influxdb and displays it > > So here you are using weewx as a driver, and influxdb/grafana as your > main weather storage/display software? Not 'as a driver' perhaps. More like 'using the appropriate hardware to talk to the VP2 console for me'. I use weewx to connect to the VP2 console in order to get the data the hardware emits in LOOP packets. I then jump through some hoops to get that data into influxdb so I can cook up an integrated dashboard. I later do comparisons of VP2 vs. WF SmartWeather vs. WF Tempest vs. Arduino vs. PurpleAir hardware in one integrate dashboard. To do that I need to get all the data into one place. Influxdb is great for that, at the cost of setting it up and cooking up an integrated dashboard myself. Examples: - I use weewx's v4 extended schema to get me maxSolarRad graphed nicely in grafana, from data in influxdb. I overlay the WeatherFlow Tempest solar sensor readings. I overlay the WeatherFlow SmartWeather station solar sensor readings. I make them available to the WF vendor as part of being a beta tester for tuning their sensors. They use this to do things like make sure we don't exceed the maxSolarRad here in the usually grey Pacific Northwest. You can't have a UV of 12 here :-) - Similarly I overlay 5 sources of outside temperature for comparison and to help the WF vendor tune their gear. I'm a beta tester for a couple of their products. - Similarly I overlay 3 sources of outside pressure. One thing I found is some gear corrects pressure for altitude. Some gear doesn't. Depends on the vendor. (that kind of stuff) If I tried to all this via weewx only, it is technically possible, but I'd pull my few remaining hairs out battling mapping stuff to available fields in the weewx v4 extended schema. I'd then need to battle writing a custom cheetah skin to display it. And I mean battle. Cheetah is kinda brutal if you try to do to much in it. Taking the influxdb + grafana path, I still need to write a custom dashboard, but it's much much simpler to handle. Not free, but way less stressful. But.....a local LAN-only grafana dashboard does 'not' get me the same functionality as weewx uploading to a static website on Internet via rsync. Different tools for different jobs. Hope this helps explain. -- You received this message because you are subscribed to the Google Groups "weewx-development" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/weewx-development/3d9c3061-a8de-4af0-a8e0-d357410398c0%40googlegroups.com.
