Spring / Summer 2018 and onwards software

[Penri]

Hello

I won't be around for the next couple of days so I can't be there to help, good luck.

Hwyl

Penri

[hamishmb]

Hello,

I came in and fixed the issue today. It was nothing major. Here's a description of the problem (excerpt from the email):

The issue was an accidental delay I introduced a long time ago when queuing readings from the sump probe, but it only became an issue as of yesterday, because of changes to the software's internals to enable future works.

I will now commit the fix to our code repository, to avoid this issue in future. The issue is still present on the buttspi side, but it won't be a problem because of the way the software starts:

• On buttspi the software starts and connects to sumppi when/if it can.
• On sumppi, because of temporary changes, the software will now wait for buttspi before it starts to do anything.
• ^ I will revert this back to the old behaviour of starting anyway when possible. For now it will be fine - our network connection between pis seems just fine.

I'm happy this wasn't another occurrence of our previous, non-diagnosable issue.

Everything else seems to be working well, so all is good .

Hamish

[hamishmb]

To clarify, the software would eventually have reacted, but just several minutes late, which isn't much use.

NB: Terry, DHCP on your router works fine, I realised I configured a static IP for the ethernet port on my laptop

[hamishmb]

UPDATE:

Several more changes have been made. Mostly these are more enabling works for the future, notably:

• All config for all sites is now stored in config.py (previously called universal_standalone_monitor_config.py).
  • Hence, there are no longer any commandline options for main.py.
  • There will ideally be just one commandline option for universal_monitor.py (--site).
  • ^ Currently, we still have -n to specify the number of readings to take (usually unspecified to run forever).
  • The -t, --type, option has been removed, because config.py now contains this info.
• More refactoring has been done between the two scripts to make them as flexible as possible from a monitoring point of view.
• ^ This is now largely complete, and will make it easy to implement the overall site log.
• The logic for controlling gate valves will probably be in core tools, in case we want a remote pi to handle that.

Quick question: we have never specified the number of readings to take - it's always been indefinite. Shall I remove this option and the supporting code?

There is a fair bit of refactoring being done here, as our design needed to change, but we're getting pretty high levels of flexibility now. Yay for real-world software development XD. Thanks for being patient waiting for the new features to roll in

Hamish

[hamishmb]

As noted elsewhere, I have reverted the readings file format to make it easier to plot results on a graph and do other analysis.

Some other changes are in the pipeline to help add support for multiple pis - more centralised config in config.py.

[hamishmb]

The software has been updated.

Summary of changes:

• The readings file format was reverted to make charts easier to produce / allow easier analysis.
• Multiple sockets can now be created at the master pi end using the configuration in config.py.
  • This is (I think) that last thing that was stopping us from adding more pis to the system with relative ease.
  • Each remote pi will communicate over its own dedicated socket to the master pi, counting up from port no. 30000.
  • ^ eg 30000 - buttspi, 30001 - stagepi, 30002 - railwaypi etc.
• Probe objects now have an id attribute instead of a name one.
  • This makes it easier for us to identify different probes at different sites for making decisions.

Now all we need is support for gate valves and PWM I guess I'll have to wait for that, unless we've decided how to implement these things.

Hamish

[Penri]

Good job.

[hamishmb]

I fixed a few issues with the new software today:

The circulation pump was configured for pin 5 for some reason - now pin 18 as specified.
A software mechanism I used at home for looping over a dictionary didn't work at WMT. I will investigate on my pi, but it's weird because I already tested it...

If we run out of stuff to do on Thursday (if I make it in), I'll have a look then, or I'll come in at some weird time when I'm free to have a look.

Also:

I have edited the .bashrc files on the pis:

• Sumppi's script now takes no options - "python3 main.py".
• Buttspi's script now only takes one option - "python3 universal_monitor.py --id "G4".

[hamishmb]

I updated the software again and updated the pis today:

• Magnetic probe levels can now be logged immediately when a level change is detected, but before the reading interval has elapsed.
  • This is a convenience feature so we can check the probes are working more easily.
  • Currently disabled: needs debugging log mode enabled to prevent huge file sizes.
• Remove dot in filenames to making charting w/ libreoffice easier.
• Remove old config in config.py (new format).
• Make the main files executable for convenience (been meaning to do this for months).
• Update docs for config.py