How to managed data on DB and query by ma-point-query

pierjsap

@CraigWeb are you referring to this example?

CraigWeb

yup I think its doing exactly what you want but obliviously a different process....

pierjsap

@CraigWeb How could I create a batch? I do not know where to start. I can not take the time as a point of reference, but I need to give an ID to every series of processes that take place between a start and a stop. I did not understand how to accomplish this

phildunlap

Whoa, let's not get ahead of ourselves too far.

It is true that the Brewer's Dashboard provides means of aggregating information about a product produced in batches, and in that sense it definitely is related and possibly usable for what you're talking about. However, it's definitely got an idiosyncratic implementation that was tailored to use in a brewery, and most specifically their fermentation. Things like automatically advancing between stages in terms of days, or only being able to specify one value to set to all setpoints associated with a fermentation tank in an automatic stage advancement, requiring a 'brewers dashboard monitor' data source with a glycol sensor and a cooling utilization data point, etc may make it cumbersome to adapt to the processes' needs to aggregate data for or control. Also, the brewers dashboard is not included in the Enterprise license.

If you do want to give it a try, you can, there are instructional videos in this category: https://forum.infiniteautomation.com/category/20/brewers-dashboard

Tanks need to be created as Temperature Control Monitor data sources, which are then associated with sensors. Then recipes need to be declared, which are stages composing a process. With that set up, you can begin a process in an asset.

I do think it would be worthwhile to have a further generalized process data aggregation / stage progression tool.

Jared Wiltshire

@pierjsap I would say that Mango contains all the tools necessary to construct such a dashboard. I would not be trying to store the information in a SQL table, you should just leverage data points which keep their data in the time series database.

Data point tags
JSON store
Meta data points
Custom dashboard
User module

I can't give you a complete example as its obviously something a little more complex. If you give us a little more information I can give you some pointers though.

Is there a set of data points x,y,z per machine, and multiple machines?
Is the start/stop for each machine just a binary data point?
What is the "product object" and how is this information obtained/entered?
How is the product id generated?
What sort of reports do you need? We can build a dashboard to display a report like view which could be printed, however using a scheduled/emailed report would be trickier

If you contact support we can help you with a custom tailored solution.

MattFox

If it's not affecting IP, i'd be keen to see the outcome of this job, sounds really interesting and could work for me with a future control related job.

pierjsap

@Jared-Wiltshire

Is there a set of data points x,y,z per machine, and multiple machines?
Yes, my project involves the construction of n-machines that have n-data points. But I would start to implement only one with 3 data points.

Is the start/stop for each machine just a binary data point?
I have inserted a binary button that is read by a scripting data source that when set to true starts reading 3 data points.
In this regard, I wanted to ask you if you can activate and deactivate a data point and / or a data source using a script command. Eg by pressing the active button and deactivating the data source and its data points

What is the "product object" and how is this information obtained/entered?
I'll give you an example. Let's say that I have to create a glass object, this object can be a glass, a plate or a bottle. Each of these can be characterized by the various chemical components, because a glass must be transparent a bottle of green color a dish must be more resistant than both. So to realize this I have to observe how long it takes to make an object? I need to get some energy consumption data for a glass,
derive the consumption for the realization of a dish. These data will then be used for statistical purposes ie to make a stronger glass takes 1 hour of time, I have to melt the components at 1500 ° C instead of 1000 ° C. So every time I start production of an object I can not take into account a period of time. But I have to take into account an ID that identifies a single production process process for both immediate consultation and future data consultation.

How is the product id generated?
I do not know, this was my question, I was trying to understand how to associate a production process with an object.
Example:
ID: 001
XID: dhsflhdsfdhskfhdsfhds001
type: Yellow glass
realization time: 2 hours
realized on: 2018-05-30 9.30 AM
maximum temperature: 1000 ° C
temperature per minute 1: 500 ° C - power per minute 1: 500W
temperature per minute 2: 800 ° C - power per minute 1: 2000W
temperature at minute 3: 1000 ° C - power per minute 1: 3000W
temperature per minute 4: 500 ° C - power per minute 1: 1000W
temperature per minute 5:30 ° C - power per minute 1: 500W

What sort of reports do you need? We can build a dashboard to display a report like view which could be printed, however using a scheduled/emailed report would be trickier
I think reporting is the least of my problems because Mango has these customizable tools

MattFox

From what I'm seeing here, the most logical way of doing this is either using the Json store or an alphanumeric data point storing json. This way you could maintain all of the properties/characteristics applied for a given product generation session. When graphing however, take the list of point values and parse the json inside. Combine the respective values into arrays of the same type, then you can graph it all simultaneously to provide a characteristics chart. Am I getting the right idea here or way out left field?

phildunlap

From what I'm seeing here, the most logical way of doing this is either using the Json store or an alphanumeric data point storing json. This way you could maintain all of the properties/characteristics applied for a given product generation session. When graphing however, take the list of point values and parse the json inside. Combine the respective values into arrays of the same type, then you can graph it all simultaneously to provide a characteristics chart. Am I getting the right idea here or way out left field?

Sounds reasonable to me. I may be over-complicating what is being talked about, but I would think it's going to involve several components, and those could all be done a few ways, I'm sure. My first hack at some of the potential problems:

Getting data about a batch in progress: create batches by tagging points with the batch id, then retrieve this in queries, store tag creation time in the json store (I would also use the JSON store to keep stage history and possible sequences if I had some sense of stage progression). Figure out what batches exist by processing what JSON store entries exist.
Ending batches and maintain a process history: gather point IDs from the tags, start time from the JSON store, and serialize that into an alphanumeric virtual point as JSON with the end time and the id of any process script run on it.
Arbitrary setpoint process control: at batch creation time POST a script to the /rest/v2/script/run endpoint with some RuntimeManager.sleep(millis); handling the delays. Store the whole request model in the JSON store or construct the context points from the tags and store the script. Set values out to some update alphanumeric point if you want a live update feed from running processes. You might also need something like that if you're to be able to resume a process after a power failure, as the script would no longer be running. I would consider making the script somewhat arbitrary by storing process control information into an alphanumeric point as JSON, also, which may give it the ability to resume if it knew where it was and what resuming should be.

Then reviewing batches is a matter of one point's history, and when a value is selected parsing that JSON to make the appropriate requests on the points from the process, as I think Fox is suggesting.

Jared Wiltshire

@pierjsap said in How to managed data on DB and query by ma-point-query:

I have inserted a binary button that is read by a scripting data source that when set to true starts reading 3 data points.
In this regard, I wanted to ask you if you can activate and deactivate a data point and / or a data source using a script command. Eg by pressing the active button and deactivating the data source and its data points

Its pretty easy to enable/disable data points from a dashboard using a point's enable() method. And yes you can also enable/disable data points or data sources from a script. However I'm not sure this is necessary, I would probably just leave the data points logging continuously.

Tag each data point with a set of tags to uniquely identify the machine.

To find and view a set of trends/statistics for a particular product's production I would query the binary data point which indicates the start/stop of the machine and find the time periods where the machine was running. You can then query for the point values / stats for the x, y, z data points for the time period selected. You can collate these time periods with the product ID that corresponds to each by checking the value of the product ID data point (see below) at that point in time.

@pierjsap said in How to managed data on DB and query by ma-point-query:

But I have to take into account an ID that identifies a single production process process for both immediate consultation and future data consultation.

I would probably have a meta data point which generates a new ID whenever the start/stop binary data point goes from 0->1. The data point could be an alphanumeric data point or just a plain numeric data point containing an incrementing integer.

Another alternative would be to generate the new product ID on the frontend (dashboard) and allow the operator to override it via a text box. It seems like you need the operator to also select the product type e.g. "type: Yellow glass". You could provide a drop down select for this. Then when the operator clicks the start button, it would set/save these two values to two separate data points as well as setting the start/stop binary data point to true.

pierjsap

@Jared-Wiltshire
Its pretty easy to enable/disable data points from a dashboard using a point's enable() method. And yes you can also enable/disable data points or data sources from a script. However I'm not sure this is necessary, I would probably just leave the data points logging continuously.

What's wrong?
Would you give me an example of how to recall classes and methods for working with datasoruce and datapoint and extracting the fields?
I do not know the Java architecture you have implemented ...

Tag each data point with a set of tags to uniquely identify the machine.

It is possible to set "tags{}" field ?
Eg. tags:{"nameProduct" : "productAlpha" , "id": "164", xid": "DP_RomeElettricoScrPotenza","name":"RomeElettricoScrPotenza"}

Thanks in advance

phildunlap

You should consult the Mango JavaScript help, linked to by the blue (?) in the scripting data source properties and here: https://github.com/infiniteautomation/ma-core-public/blob/main/Core/web/WEB-INF/dox/mangoJavaScript.htm

The answer would be to use the RuntimeManager object, like RuntimeManager.enableDataPoint(ScrAmperaggio.getDataPointWrapper().getXid());

To the tags modification question, it sounds like you may be seeking this thread: https://forum.infiniteautomation.com/topic/3439/accessing-a-points-tag-value-via-script

Jared Wiltshire

@pierjsap said in How to managed data on DB and query by ma-point-query:

What's wrong?

I specified "from a dashboard", you are trying to use the function in a server side script. These are two different environments.

@pierjsap said in How to managed data on DB and query by ma-point-query:

Would you give me an example of how to recall classes and methods for working with datasoruce and datapoint and extracting the fields?

I was not suggesting to tag data points from within a script, in fact I would advise against this. I was suggesting to setup all your data sources for the sensors and possibly meta data sources as required then tagging these through the bulk data point editor in the UI. You can these use these tags from your dashboards to locate the applicable data points for the machine you are viewing.