20230902 - Bug Fixes and New Features

Version summary

Notable new features

Some major database changes have been made to support the features below

Click to expand

1. To the asset_cycle_data table:
   • new columns
     • cycleCount - how many cycles each entry represents. Defaults to 1 for all existing records.If two cycles are logged in succession, for the same asset, X, with a cycle count of V, and then Y, then X + V = Y.
     • sourceType - Where did the cycle entry originate from.
       • NORMAL
       • AGGREGATED : stamping
       • GAP : describes a set of cycles that were artificially generated
   • columns that get renamed:
     • totalCycleNumber -> startCycleNumber : Cycle number of the asset at the start of the cycle entry.
     • cycleInterval -> cyclesDuration : How long this entry took, in ms. To retrieve the cycle interval, divide this value to the cycleCount
       • For NORMAL entries, equal to the cycleInterval .
       • For AGGREGATED entries, equal to the sum aggregationPeriod of the cycle events
       • For GAP entries, it's value is equal to the amount of time between when the gap started, and when and estimation of when it finished. GAP entries always have the artificial flag set to true
2. To the asset_kpi TABLE:
   • new columns
     • trustedCycleCount - how many "trusted" cycles occurred in this interval. A trusted cycle is defined as being a "NORMAL or "AGGREGATED" cycle source type, and not part of a "GAP". trustedCycleCount <= "endCycleNumber" - "startCycleNumber"
   • generated columns that previously used "endCycleNumber" - "startCycleNumber" to determine the number of cycles in an interval will now use trustedCycleCount . This number will always be `<=
     • the performance formula
     • all calculations for the avgCycleTime

Major changes to how cycle gaps are processed

Previously, all gaps > 100 cycles would be ignored, and they would not lead to any changes on the tool attached to the monitor. The behaviour is now different.

Gaps are of three kinds:

• SMALL gaps (<100 cycles) will be filled using artificially generated cycles. They will affect KPIs.
• MEDIUM gaps (between 100 and 100_000) will be logged as gaps, they won't affect kpis but will lead to an increase of the cycle count on the tool. The duration of the gap is determined as being the time between when the gap started and the time the message with the cycle that ended the gap was received, minus the duration of all other cycles that were received on the same timestamp. This gives us an estimation that is as close to reality as possible, given the available data. The cycles in the gap do not affect KPIs. Gaps of this type to not affect any downstream machines or inserts because we do not know where they occured.
• LARGE gaps (>100_000) will be completely ignored.

Additionally, decreased the max size of the unprocessed cycles from 100 to 15. This should make the application receive cycles a bit faster.

Lastly, the maximum possible duration of a cycle used to be restricted to 24 days. It now can be over 2 million years.

Initial stamping support

Initial support for stamping, allowing the server to process stamping messages from v4 gateways. Note that this support still requires more testing as there are major issues that need to be clarified on the format of the data coming from the monitors themselves.

Improvements

• improve the favicon of the application and make it look better when using it as a desktop shortcut (#426)

  

  

Cycle time report

• Sort the data by timestamp, to work around #461

• Show the total cycle count in that report interval. This is useful both for normal tools, but also for stamping dies, that receive multiple aggregated cycles

  

Molds dashboard

• Always show both the tool and press name

  

• show when mold last processed cycle on overview and dashboard also clarify some translations. Note that this is also available on the asset overview page, and as a column in the asset list view.

  

• tool name and press names are shown on different rows
• icons are used to tell the two apart
• sorting is done by the press display name
• when "out of tolerance first" is enabled, sorting sorting inside the assets with the same tolerance is done by the press name

Other changes

• Replace reverse geocoding api provider to MapTiler and replace the use of state to region , for a more non-US centric view.
• Remove the @logged flag from properties since it wasn't really used, and were providing useless load on the database
• Add @indexed flag on properties that are usually searched using QueryImplementingThingsOptimized

No longer broken

• PM warnings are now generated when currentCycleCount > (cycleCountGoal - 5% *pmCycleInterval) . Previously they were generated at currentCycleCount > cycleCountGoal * 95% . Same logic applies for hour active conditions. This is also reflected on the PM countdown on the dashboard where yellow is used correctlyThe dashboard PM indicator bar now uses the same limits to show green/yellow/red colors.
• Resolve issue where logging downtime may fail with a database error
• Don't show the default temperature in UI on the Asset Overview page (don't show -273.15)