Request for assistance about WeeWx config for a water level sensor

[marco rainone]

Hello everyone.

This month I started the development of an opensource application for the acquisition of meteorological data in developing countries.

The datalogger is composed of a Raspberry PI with Ubuntu and WEEWX installed on using the Python tool setup.py.

The developed software periodically read the ambient data  (temperature, pressure and so on ...) from different sensors, prepare a text file read by WEEWX (for flexibility reasons, the acquisition software pass data to WEEWX, using text files read by fileparse driver).

Finally, WEEWX collect the physical quantities in a Sqlite3 database and create reports.

During the development period I have already solved various problems encountered, by carefully reading the documentation and several posts in this forum.

Unfortunately, I'm finding problems about the ability  through WEEWX to manage an optional sensor (a commercial water sensor level, used to acquire the water level in a river) and then ask for help.

Actually, my software collect data from the sensor level correctly, and I want to pass these values to WEEWX.

Since it is an optional sensor, I had thought of managing the sensor data using a different WEEWX process, passing the data to the fileparse driver to store them in the weewx SQLite3 database.

I have an even limited experience on WEEWX, so I still don't know how you could correctly configure the program to manage the 3 variables read by the sensor:

1. "temp": the external temperature measured, in celsius degrees
2. "'Battery_SOH": battery state of health, a percentage value that represents the charge status of a non-rechargeable battery. It is necessary to monitor the charge status to check when physically change the battery.
3. "'level": distance of sensor from water, in centimeters

For example, In base of weewx documentation (particolarly, [WeeWX: Customization Guide](https://www.weewx.com/docs/customizing.htm#aggregation_types)), I see that I could assign the sensor temperature to outTemp, but I don't know what standard weewx variable (corrispondent to a column of the sqlite database) I can assign to "Battery_SOH" and "level"  and to also draw the graphs using the configured skin.

What are the steps I need to follow to configure weewx correctly in this case?

In particular, there are numerous WEEWX variables that can be used to monitor sensor power supplies, but the measurement units are expressed in Volts. In my case, Battery_SOH is expressed with a percentage value of full charge.

Thank's for collaboration.

Marco Rainone - ICTP Wireless Lab

Wireless | T/ICT4D Lab (ictp.it)

[matthew wall]

hello marco!

this sounds like a wonderful project!  i am working on a community-oriented water level project that measures tide depths using maxbotix sensors for sea level rise analysis - we are hoping to get accuracies that are acceptable for use by the US NOAA in their research.

short answer:

assuming that you use a single database, and assuming that you emit all of the sensor data to a single text file that is read by the fileparse driver: you should use names that correspond directly to the names in the schema.  in the case of observations related to water level, i would suggest the following:

temperature - use 'extraTemp1' (or extraTemp2 or 3 or ... if 1 is already in use)

Battery_SOH - use 'batteryStatus1' (or 2 or 3 or ... corresponding to the temperature)

level - use 'snowDepth'

in the weewx config file, add label entries in a [Defaults] section of [StdReport] so they show up in every skin.  you might have to set the units on snowDepth - try it first as-is, then if the units are not right in the Seasons skins you can adjust.

long answer:

let's look at your problem from two points of view.  first, how to collect the data.  it sounds like you are doing that already use the fileparse driver: your software collects data from sensors, saves the data to a single file, then the fileparse driver reads the file.  with this configuration, the names used in fileparse map directly to database field names.  so if there is a name in fileparse that does not exist in the database schema, that observation will not be retained (but you can still see it in LOOP/archive data when you run weewx directly).

a more advanced approach (but not necessarily better!) would be to write your own weewx driver.  in this case, the driver would poll the sensors map the observations directly to database field names.  this approach eliminates the need for saving to file and reading from file on every LOOP/archive.  you can write the drive so that it can be tested directly, such as when you need to test the code that communicates with the sensors (see other weewx drivers such as weewx-maxbotix, weewx-gw1000, weewx-sdr, for examples of how to do this).

second point of view is how to store the data.  for this part there are two approaches: (a) overload existing fields in the standard schema, or (b) define your own schema.

if you do not have many additional observations, then overloading existing fields is the easiest, yet still robust and maintainable option.  however, beware that you might have to override the units, and you will almost certainly have to define your own labels.  do the units overrides in your bin/user/extensions.py file.  do the label overrides in the weewx.conf file, preferably in a [Defaults] section so that they show up in every skin, whether it is a standard weewx skin or something you create.

defining your own schema can be done within a driver file (see the weewx-maxbotix driver as an example), or in the bin/user/extensions.py file, or a standalone schema file in the bin/user directory.  this is a good option if you have more than 10 observations that do not correspond well to existing fields, or if you are already using all of the existing fields. i'm not sure if we have a section in the customization guide yet about using an entirely different schema...