ECMT

Friday, September 11

Monday, August 3

Friday, July 17

Friday, July 3

Monday, June 29

Staging

Version number on the login screen

Results per page

Renaming "Diesel only" field to "No catenary"

Bulk editing notice, if an object is selected where the editing cannot be performed on

Long dwell time on a station (midnight crossing)

Show running days instead of validity period in TCR tooltip

Select objects on this page or on all pages

Extend traffic measures matrix of TCRs with commuter trains

Presentation of objects on overlapping lines

Multiple day object presentation

Demo

Personal filters

Detection of overlapping objects

Description

Please find here the summary of the ECMT developments, including the latest news from the Sprint Review meetings and also the candidate of features available currently in the Staging or in the Demo site.

ECMT Production                                                                           

https://ecmt-online.rne.eu/

ECMT Staging

https://ecmt.staging.i0.hu/

Both environments are accessible without a user account, using the public access as an anonymous profile.

Documentation

The documentation is publicly available on the following page: https://ecmt-online.rne.eu/documentation

Recent Activity

Filter 

September 11, 2020

  • 3:11pm
    [ECMT] [Discussions] Máté Bak updated 2020.09.11. Sprint review Friday, September 11Máté Bak updated 2020.09.11. Sprint review  Too many differences to list. See Revisions for details.Máté Bak updated 2020.09.11. Sprint review

    [ECMT] [Discussions] Máté Bak updated 2020.09.11. Sprint review

    Friday, September 11

    Máté Bak updated 2020.09.11. Sprint review

     

    Too many differences to list. See Revisions for details.

    Too many differences to list. See Revisions for details.
    Read more
  • 3:08pm
    [ECMT] [Discussions] Máté Bak created 2020.09.11. Sprint review Friday, September 11Máté Bak created 2020.09.11. Sprint review  ReviewFurther development of the overlapping object detectionThere was an issue with the parallel objects (paths) and those that are crossing each other with an offset (same offset).There are two problems to solve to avoid false-positive conflict detection:If there are two objects, both of them offsets, but the offset values are not the same (+2 and +1)If there are two objects with an already detected false positive conflict and a real conflict, we don't show them now.Overlapping objects on lines where location places are differentWe had a presentation mistake for those objects that are on lines that cross each other more than one time, e.g. the Breda lines. Now, the objects are properly shown where they enter and where they leave the line.Before the bug fix:After the bug fix: Entity type filter doesn't work consequentlyOn the conflict pages (still available only in the demo), if the user added the TCR to the objects, the formerly shown other objects (e.g. paths) disappeared.Circulation days are still deleted during importThere was a mistake in the exported Excel file. Even though the user could see there 0 and 1 for the calendar days, they were functions. During a re-import, it would have caused trouble. Now, it's fixed and really 0 and 1 are exported as numbers.Next sprintsWe will spend time on the analysis of the false positive conflict detection. The implementation will follow only after the analysis.Máté Bak created 2020.09.11. Sprint review

    [ECMT] [Discussions] Máté Bak created 2020.09.11. Sprint review

    Friday, September 11

    Máté Bak created 2020.09.11. Sprint review

     

    Review

    Further development of the overlapping object detection

    There was an issue with the parallel objects (paths) and those that are crossing each other with an offset (same offset).

    There are two problems to solve to avoid false-positive conflict detection:

    • If there are two objects, both of them offsets, but the offset values are not the same (+2 and +1)
    • If there are two objects with an already detected false positive conflict and a real conflict, we don't show them now.

    Overlapping objects on lines where location places are different

    We had a presentation mistake for those objects that are on lines that cross each other more than one time, e.g. the Breda lines. Now, the objects are properly shown where they enter and where they leave the line.

    Before the bug fix:

    After the bug fix:

     

    Entity type filter doesn't work consequently

    On the conflict pages (still available only in the demo), if the user added the TCR to the objects, the formerly shown other objects (e.g. paths) disappeared.

    Circulation days are still deleted during import

    There was a mistake in the exported Excel file. Even though the user could see there 0 and 1 for the calendar days, they were functions. During a re-import, it would have caused trouble. Now, it's fixed and really 0 and 1 are exported as numbers.

    Next sprints

    We will spend time on the analysis of the false positive conflict detection. The implementation will follow only after the analysis.

    Review

    Further development of the overlapping object detection

    There was an issue with the parallel objects (paths) and those that are crossing each other with an offset (same offset).

    There are two problems to solve to avoid false-positive conflict detection:

    • If there are two objects, both of them offsets, but the offset values are not the same (+2 and +1)
    • If there are two objects with an already detected false positive conflict and a real conflict, we don't show them now.

    Overlapping objects on lines where location places are different

    We had a presentation...

    Read more

August 3, 2020

  • 2:00pm
    [ECMT] [Discussions] Máté Bak created 2020.08.03. Sprint review Monday, August 3Máté Bak created 2020.08.03. Sprint review  ReviewConflict detection offset casesThe whole detecting algorithm needed adjustment because the offset related objects were not detected correctly.The other issue was related to TCRs. If a TCR is between A and B, and a path arrives at B, we don't show it as a conflict, because the TCR is on the section. We had to solve this special use case.The algorithm has two parts:Queries and presents the overlapping objects on a gridLooks for the type, place and time of the overlappingWe found a new complex issue: we cannot handle the cases, where there are two lines with more common stations. It works now only when there is one common station. We have lines like that in the Austrian pilot, so we need to prepare the query for cases like that too.Testing conflictsWe are going through all the test cases prepared by the following scenariosObject vs. objectDirectionsTotal closureEtc.Import/export bugAn empty cell is exported as FALSE or 00:00. This had to be fixed. This caused the error on the Brenner pilot when they re-imported the exported catalogues.Currently, we write True(), False() functions for the boolean values, not strings. Then, it's the matter of the application (Libre Office, MS Office, WPS Office, etc.) how it is interpreted and presented on the screen.Brenner pilot issueIt was related to the export issue, and the objects were shown wrong on the Chart. Now, it's fixed and it will be released after the demo meeting (3rd August 2020), including the fix of the import/export issue.Calculate offset problemThe problem was found during the cloning of the original object. Times were cloned, but the offsets were not taken into account. Plus, we showed the calculated times too.Next sprintsBug fixing on the test cases already detected issuesFurther testing of the conflict detectionPreparation of the automated testingThe next review is expected on 4th September due to the vacation period. The conflict (overlapping) detection feature is also expected to be released only in the Autumn.Máté Bak created 2020.08.03. Sprint review

    [ECMT] [Discussions] Máté Bak created 2020.08.03. Sprint review

    Monday, August 3

    Máté Bak created 2020.08.03. Sprint review

     

    Review

    Conflict detection offset cases

    The whole detecting algorithm needed adjustment because the offset related objects were not detected correctly.

    The other issue was related to TCRs. If a TCR is between A and B, and a path arrives at B, we don't show it as a conflict, because the TCR is on the section. We had to solve this special use case.

    The algorithm has two parts:

    1. Queries and presents the overlapping objects on a grid

    2. Looks for the type, place and time of the overlapping

    We found a new complex issue: we cannot handle the cases, where there are two lines with more common stations. It works now only when there is one common station. We have lines like that in the Austrian pilot, so we need to prepare the query for cases like that too.

    Testing conflicts

    We are going through all the test cases prepared by the following scenarios

    • Object vs. object
    • Directions
    • Total closure
    • Etc.

    Import/export bug

    An empty cell is exported as FALSE or 00:00. This had to be fixed. This caused the error on the Brenner pilot when they re-imported the exported catalogues.

    Currently, we write True(), False() functions for the boolean values, not strings. Then, it's the matter of the application (Libre Office, MS Office, WPS Office, etc.) how it is interpreted and presented on the screen.

    Brenner pilot issue

    It was related to the export issue, and the objects were shown wrong on the Chart. Now, it's fixed and it will be released after the demo meeting (3rd August 2020), including the fix of the import/export issue.

    Calculate offset problem

    The problem was found during the cloning of the original object. Times were cloned, but the offsets were not taken into account. Plus, we showed the calculated times too.

    Next sprints

    Bug fixing on the test cases already detected issues

    Further testing of the conflict detection

    Preparation of the automated testing

    The next review is expected on 4th September due to the vacation period. The conflict (overlapping) detection feature is also expected to be released only in the Autumn.

    Review

    Conflict detection offset cases

    The whole detecting algorithm needed adjustment because the offset related objects were not detected correctly.

    The other issue was related to TCRs. If a TCR is between A and B, and a path arrives at B, we don't show it as a conflict, because the TCR is on the section. We had to solve this special...

    Read more

July 17, 2020

  • 11:30am
    [ECMT] [Discussions] Máté Bak created 2020.07.17. Sprint review Friday, July 17Máté Bak created 2020.07.17. Sprint review  ReviewAdd new location country bugLast used country is presented.Maximize lines tooltipIf there are too many (more than 10 lines) of periods in the overlapping, then we cut it and indicate that with "..."Path/CP crossing Band on stationThis was the missing option from the combinationsDetecting conflicts when offset for one object existsWe extended the search criteria with the offset value. Now, it may create false positive results, but that shall be fixed in the next Sprint.Base entity (the one that arrives at a station with an offset) has the list entity as a result.List entity (the one that runs on that day where the base entity has the offset) doesn't have currently the base entity. That will be fixed with the next sprint.Also, the offset related conflicts doesn't have yet the type and place of the overlapping. That will be solved later too.Test casesExcel with the possible test cases. Including also the false positive cases that should not appear. We will need two additional options:Overlap with TCR on stationOverlap with TCR on station - total closureNext SprintFixing those missing offset related issues.Testing the not offset related overlapping.Re-designing the filter options on the overlapping result view. Goes to production:Multiple day objectsOverlapping sectionsMáté Bak created 2020.07.17. Sprint review

    [ECMT] [Discussions] Máté Bak created 2020.07.17. Sprint review

    Friday, July 17

    Máté Bak created 2020.07.17. Sprint review

     

    Review

    Add new location country bug

    Last used country is presented.

    Maximize lines tooltip

    If there are too many (more than 10 lines) of periods in the overlapping, then we cut it and indicate that with "..."

    Path/CP crossing Band on station

    This was the missing option from the combinations

    Detecting conflicts when offset for one object exists

    We extended the search criteria with the offset value. Now, it may create false positive results, but that shall be fixed in the next Sprint.

    Base entity (the one that arrives at a station with an offset) has the list entity as a result.

    List entity (the one that runs on that day where the base entity has the offset) doesn't have currently the base entity. That will be fixed with the next sprint.

    Also, the offset related conflicts doesn't have yet the type and place of the overlapping. That will be solved later too.

    Test cases

    Excel with the possible test cases. Including also the false positive cases that should not appear. We will need two additional options:

    • Overlap with TCR on station

    • Overlap with TCR on station - total closure

    Next Sprint

    Fixing those missing offset related issues.

    Testing the not offset related overlapping.

    Re-designing the filter options on the overlapping result view.

     

    Goes to production:

    • Multiple day objects

    • Overlapping sections

    Review

    Add new location country bug

    Last used country is presented.

    Maximize lines tooltip

    If there are too many (more than 10 lines) of periods in the overlapping, then we cut it and indicate that with "..."

    Path/CP crossing Band on station

    This was the missing option from the combinations

    Detecting conflicts when offset for one object exists

    We extended the search criteria with the offset value. Now, it may create false positive results, but that shall be fixed in the next Sprint.

    Base entity (the one that arrives at a station with an offset) has the list entity as a result.

    List entity (the one that runs on that day where the base entity has the offset) doesn't have currently the base entity. That will be fixed with the next sprint.

    Also, the offset related conflicts doesn't have yet the type and place of the overlapping. That will be solved later too.

    Test cases

    Excel with the possible...

    Read more

July 3, 2020

  • 11:03am
    [ECMT] [Discussions] Máté Bak created 2020.07.03. Sprint review Friday, July 3Máté Bak created 2020.07.03. Sprint review  ReviewValidate band timetableThe aim of the task was to avoid situations when the user enters only one time for a station for a band. Now, it's not possible anymore, and even the GUI gives an error. It works the same way in the import too, the tool gives back an error that a time field is empty.TCR Traffic Measures - Estimated delay minutes storage and displayUntil now, we stored the delay minutes in the wrong format, that is why they were presented wrong on the GUI after creation, e.g. 12:00 instead of 12 because it's a minute field. Now, it's fixed and the presentation also works fine.Band add problemDuring RNE testing, Lukas found an issue. This has been fixed.When adding band "from time" and moving to "to time" of the band the system automatically selects offset +1:After entering "to time" of the band and clicking save the system opens error message:Conflict detectionAvailable on demo only.From the navigation point of view, each object type has now the "Detect overlapping objects" option. The form with the search results is identical with the Chart search options.The harmonization status filter is visible only for Admin and Pilot users. Read-only and Anonymous users can see only published objects.In the tooltip, we show also the time frame of the overlapping.Currently, the following overlapping scenarios are covered for overlapping, crossing and overtaking:Path/CP - Path/CPBand - BandBand -TCRTCR - TCRPath/CP - TCRPath/CP - BandThere is one missing, the Path/CP - Band crossing between two stations. Also, when the objects have different offsets on a station, we cannot detect the overlapping yet.Next SprintPreparation of the test casesImplementation of the missing Path/CP - Band crossing between two stations detectionAutomated testingAs the number of possible scenarios and the complexity of the system is growing, we will prepare a plan for the implementation of automated testing. From the effort point of view, the implementation might be 2-3 times more than a human end to end test, but that is only once. In the long-term it would be definitely beneficial. RNE's priority on automated testing:Successful object creationValid object presentationValid overlapping detectionVersionsStaging features with the multiple-day objects (offsets) and the overlapping sections will be released by the end of the next sprint (after 17th July).On the demo site, RNE will start testing the personal filters and the potential conflict detection.Máté Bak created 2020.07.03. Sprint review

    [ECMT] [Discussions] Máté Bak created 2020.07.03. Sprint review

    Friday, July 3

    Máté Bak created 2020.07.03. Sprint review

     

    Review

    Validate band timetable

    The aim of the task was to avoid situations when the user enters only one time for a station for a band. Now, it's not possible anymore, and even the GUI gives an error. It works the same way in the import too, the tool gives back an error that a time field is empty.

    TCR Traffic Measures - Estimated delay minutes storage and display

    Until now, we stored the delay minutes in the wrong format, that is why they were presented wrong on the GUI after creation, e.g. 12:00 instead of 12 because it's a minute field. Now, it's fixed and the presentation also works fine.

    Band add problem

    During RNE testing, Lukas found an issue. This has been fixed.

    When adding band "from time" and moving to "to time" of the band the system automatically selects offset +1:

    ...

    Read more