search

levylabpitt / Instrument-Framework

An object-oriented framework for LabVIEW based on JKI SMOs.
BSD 3-Clause "New" or "Revised" License
4 stars 7 forks source link

Read Multiple Section Clusters or Modified Section Clusters #32

Closed Monkeymerlot closed 9 months ago

Monkeymerlot commented 3 years ago

https://github.com/levylabpitt/AFM-Lithography/issues/16#issuecomment-780109581 https://github.com/levylabpitt/AFM-Lithography/issues/16#issuecomment-780122280 https://github.com/levylabpitt/AFM-Lithography/issues/16#issuecomment-780195806 So I thought about the scenarios that we could have pertaining to the issue that I had in the other project and laid out what I thought would be the best way to go about this. image

Replacing the default data with the existing data from file when the cluster size has grown (ie the section is missing an element) was complicated greatly by the fact that the returned data from the read ini file VIs does not have labelled elements. Here is the test setup that I did:

  1. I created a file and initialized it with the following content:

    [Test Cluster]
Hello = "World"
Number = "1"
Number 2 = "3"

  2. Then I requested that it read from the file a cluster containing 1 more element than what was in the file. To make sure that the variant to data wasn't coercing the data I changed the values in the type cluster before I used it in variant to data: image

  3. This is what I came up with that correctly wrote to the file and returned the correct values: image And the file contents: image returned data: image

Files: Configuration Prototype.zip

ciozi137 commented 3 years ago

@Monkeymerlot thanks I will think about this. After first read can we just test if variant is empty?

Monkeymerlot commented 3 years ago

@ciozi137 Good question. I will check but I have a suspicion that it will return something like this:

" " =
" " =
" " =

Also I think that the fact that there are no names on the returned cluster is actually a bug in the openG VIs. I am thinking of opening an issue and doing a pull request with the fix (I fixed it on my machine already) but there are pull requests from 2013 that haven't been reviewed or merged yet so IDK if it will do any good.

ciozi137 commented 3 years ago

@Monkeymerlot I had to do a quick security update related to instrument/postgreSQL. I can look at the config stuff now.

Monkeymerlot commented 3 years ago

So testing for an empty variant will not work. Without my fix for the variant configuration palette: image With Fix: image and test VI: image

ciozi137 commented 3 years ago

@Monkeymerlot thanks for checking. I also just checked and the result was that my config file was erased completely. I am going to take your zip file above and merge

ciozi137 commented 3 years ago

@Monkeymerlot looking good. Update will be included in next build (1.8.3 I think)

Here is my test:

Screen Shot 2021-02-24 at 12 27 52 PM
Monkeymerlot commented 3 years ago

The test looks good but I think you might have to delete the temp file once youre done with it.

ciozi137 commented 3 years ago

@Monkeymerlot I delete it at the beginning of the test

Monkeymerlot commented 3 years ago

Oops missed that

Monkeymerlot commented 3 years ago

I was having some issues getting everything to work because it was looking for the Configuration Prototype file, but it was also not part of the class. So I put it in the class then deleted it from the project and pushed the changes. Hopefully that will fix it.

ciozi137 commented 3 years ago

@Monkeymerlot no I took your code and added it to my VIs. So just update the instrument framework on vipm

Monkeymerlot commented 3 years ago

@ciozi137 I understand. I'm saying the project was looking for it probably when you had it open to copy the code. Like a ghost dependency. And it was making the class have a broken run arrow. It should be good now though.

Monkeymerlot commented 3 years ago

@ciozi137 I am not seeing a new package (current package is 1.7.0.95). Were you saying that you wanted me to build the package or did you not get to it yet or is there a problem on my end?

ciozi137 commented 3 years ago

@Monkeymerlot refresh the repo. Latest version is 1.8.3

Monkeymerlot commented 3 years ago

@ciozi137 That was the issue *slaps forehead*. Thanks

ciozi137 commented 3 years ago

@Monkeymerlot I just reverted your commit. Instrument Framework is programmed in LV2016 and you should be installing via VIPM. Can you elaborate what issue you were having with configuration dependencies? I have no issue on my system nvm maybe I do. Let me check...

ciozi137 commented 3 years ago

@Monkeymerlot sorry I see the problem now. I will build a new version (will be v1.8.3.100)

Monkeymerlot commented 3 years ago

No worries. LabVIEW just loves bringing us back to dependency hell...

Monkeymerlot commented 2 years ago

I was wondering if we could modify this ever so slightly to this: image I have been writing things to variant attributes and LabVIEW will save those as a binary section in the file. There are two issues with this:

  1. It's not human readable or changeable.
  2. If you change LabVIEW versions, it can no longer read the variant.

So I want to dynamically change them to a cluster and write them that way. This actually works quite well. However, currently upon reading, if they already exist, I don't get the names back. This will make it difficult to set them as variant attributes. Basically I want to move the case structure to only encapsulate the write section cluster part so all clusters/sections read will be named.

ciozi137 commented 9 months ago

bump