Riffyn Release Log

v1.b434 (Version 1, Build 434) - Released October 5, 2017

Riffyn has just been upgraded in 4 very big ways!

 

1) You can now filter your Process and Experiment Libraries with shareable Tags

  • To apply your first tag, try opening the "Add/Remove Tag" modal by right-clicking on any one of the entities you have selected in the Process Library. You can create your first tags from within this modal simply by typing their names and hitting ENTER after each.
  • Tags are private until shared via the Tag Library with other users or teams. Access controls for tags are independent of the access controls on the entities that are tagged. Tag Viewers will be able to use your tag for filtering their libraries. Tag Operators will be able to apply your tag to additional entities. Tag Admins will be able to delete or rename your tag.  

2) There are new ways to export data for individual columns and individual steps:

  • Copy Selected Values: right-click on any header to copy its selected values in the order displayed
  • Export Single-Valued Data: in Execute mode, first click on the name of the run group displayed to the right of the step name in the bottom panel
  • Export Planned Values: first click on the name of the run group displayed to the right of the step name in the bottom panel

3) The bottom panel in Measure mode has been re-designed to facilitate data entry for Resources

  • Input Resources can now be assigned by typing their names or scanning their barcodes directly into the bottom panel. As always, a resource's Fixed Property values will appear automatically when it is assigned successfully. If the input resource name provided is not recognized or is not sufficiently unique, you can still access the Resource Search and Filter modes in the top panel to resolve each assignment by clicking on the corresponding gray cell. If one or more resource assignments on a step was not made successfully (e.g., an uploaded file contained a typo), the gray cell in the header region will display an alert. Clicking this alert will automatically scroll the next unresolved assignment into view. A right-click on this alert in the header will fix all unresolved assignments by creating new resources using the names provided.  
  • The names of Output Resources can now be edited in the bottom panel on the runs where those resources are generated.

4) You can now search and read our Help articles without leaving Riffyn.

  • Our help menu now offers options for "Quick Help" and "Browse Help". When you select "Quick Help" a search bar will appear in the lower right corner of your window that accesses the same content you can still review in a separate tab by selecting "Browse Help."

  

The new release contains these additional features and bug fixes.

 

New Features 

Design & Measure:

  • Fixed Properties are now indicated by a square icon so that the tag icon can be used exclusively for actions related to the new Tags feature.

Measure:

  • The "Rename Experiment" interface has been itself renamed and updated with new features for communicating why an experiment was performed, what was observed and what decisions were made. You can now attach files directly to the "Experiment Summary" interface to enhance the text you provide as either "Purpose" or "Summary and Next Steps" within its "Overview" panel. There is also a new "Observations" panel that integrates into a single location every Observation written to every Step of an Experiment.
  • The navigation bar in the bottom panel has been updated to facilitate navigation among Riffyn's interfaces for designing experiments (Plan), executing experiments (Execute), enhancing experiments with additional resource types or properties (Modify), quickly navigating across steps (Step Order), and accessing files uploaded to Procedure Tasks and Observations (Attachments).
  • Interfaces for defining Data Cleaning Procedures and viewing Cleaned Data are now accessed via the wrench icon displayed in the bottom right corner of Execute mode.
  • You can now "Apply saved aggregation settings" by right-clicking on any header in Execute mode. If you want to quickly view the runs inside a run aggregate by toggling the run aggregations off and on again, first "Show individual runs" and then "Apply saved aggregation settings".
  • "Run Description" is now displayed as "Run Name", to align with the way the same data are accessed via API.

Parsely:

  • Sometimes the files generated by your instruments are missing a key piece of information that characterizes the data, such as a Plate ID. You can now upload the data in these files without having to modify them by using Parsely's new features for Manual Data Entry.
  • Queries from Riffyn to your databases can already be parameterized using data that has been written to runs, but the values of those parameters have been calculated using all rows of all runs on the step. There is now an option to Execute Database Queries Run by Run, which executes one query for each selected run, with independent parameter values estimated from the data on each run.

Analyze:

  • The Data Status Indicator has been updated such that progress is now indicated by the fraction of the data status indicator with a green background.

API:

 

Issues Fixed 

Design & Measure:

  • Since the last release hitting the ENTER key in some modals would close them without saving changes or performing the expected action. This has been fixed according to the following scheme: if your cursor is in a single-lined text field, ENTER will usually save the value, close the modal and perform the expected action; if your cursor is in a multi-lined text field, ENTER will add a new line to the field.
  • Zero can now be entered as a target value for numeric properties.
  • Users can now successfully share a Process or Experiment while viewing the steps within one of its Groups.

Measure:

  • The Resource Search feature in the top panel will now return resources with exactly matching names at the top of the list even if there are spaces or special characters in their names.
  • If a Resource with no value of a particular Fixed Property is assigned to a run with its own value of the same Fixed Property, the Resource will now "pick up" the Run Value as if it had been written to the Resource itself. When the same resource is later assigned to other runs, it will bring this First Assigned Value with it.
  • Property attributes are now saved in the expected way when new properties are created in the Measure mode interface for adding properties to Steps.
  • The Resource Viewer in the top panel will once again display values for "Not Fixed" properties associated with a resource. Each value displayed is the one that has been most recently written to a run where the Resource has been assigned.

Parsely:

  • More intuitive mouse selection behavior when selecting Configurations in a Collection.
  • When re-naming a saved Configuration in the File Setup modal, hitting ENTER the first time will now save the name and exit the edit state; hitting ENTER a second time will trigger the 'Edit' button and open the Configuration editor.

Analyze:

  • Some exported files were missing all or nearly all data. This has been fixed.
  • Some requests for data export returned an unrelated error message. This has been fixed.
  • We have fixed one more edge case where combinations of single- and multi-valued data were not exported in the same form as they were being rendered in the multi-value data editor.
  • Plot Result Data again works as expected
  • Error message will be displayed if user attempts to Plot Result Data when there is no data.

Library:

  • The Welcome Video would sometimes appear upon switching between two Riffyn accounts within the same browser window. This has been fixed.

API:

  • Documentation for listPropertyTypes has been corrected (no ID parameter is required for this end point).
  • Removed "unique" field from Documentation for createPropertyType and updatePropertyType.
  • Users can now access Public processes and experiments via API.

v1.b400 (Version 1, Build 400) - Released September 11, 2017

Riffyn has just been upgraded!

 

You can now aggregate runs in Execute by the unique values in any column. You can turn on this feature by clicking on any header and selecting "Aggregate by..." When runs have been aggregated the bottom panel will update in a number of ways:

1) The header of the column you are aggregating by will appear dark gray.

2) The number of objects rendered in the bottom panel will be reduced from the number of runs on the step to the number of unique values in the "Aggregate by..." column. In the example above, 384 runs representing the distinct cell culture resources in each of 384 wells collapse to just 4 objects, one for each 96-well plate.

3) If more than one distinct value of a property has been written to the individual runs within each aggregation, the corresponding cell will display an aggregation of the individual values. By default Riffyn displays a sorted list of the unique values, but the aggregation method for each property can be customized. For example, numeric properties can be aggregated as the Mean, Range, Min, Max or Median of the individual values. Process Admins and Designers can save aggregation settings for any step to the process itself, so that future experiments will display data according to the stored aggregation methods every time that step is selected.

4) Empty check boxes at left are replaced with aggregation icons that reveal a check box on mouse-over. Selected run aggregations can be flowed to other steps.

For the time being users must turn off the run aggregations before they can manually edit any data on the runs. If you select one or more run groups and then select "Show individual runs" from the context menu on any header, the bottom panel will automatically snap to the first of the selected runs.  

 

The new release contains these additional enhancements and bug fixes.

Design & Measure:

  • ENTER key now adds line breaks when editing Process or Experiment description.
  • You can now download a file from a Procedure Task in an edit state.
  • Text wrap has been re-enabled when editing Procedure Tasks.
  • Various UI tweaks have been made to the Attachments panel.
  • Default units for "time (duration)" have been set to seconds.

Design:

  • Designers can once again rename Processes.
  • Pressing ENTER in the "Target" field for a Property of the "Character" data type is interpreted as a request to add a new Property.
  • Focus now stays in the Units drop-down menu when using a keyboard to select Units.
  • Exception Loops within the full-process task order can now be removed from the last task in the loop.

 Measure:

  • Resource Assignment Manager (top panel) now opens more quickly. This performance improvement was achieved primarily by switching the default mode for the Resource Finder (top left) from Filter to Search. Also, the Filter mode now loads data for only the top 200 results (it was previously 1600).
  • The Resource Viewer (top middle) is now refreshed to display any values picked up by the Resource upon assignment to a run.
  • If user has changed the displayed Resource Type in the Resource Assignment Manager, "Create New" will now create a resource with the displayed Resource Type (instead of the Resource Type designed on the step).
  • When creating a new Experiment its first step is now selected automatically.
  • Experiments in the Experiments Library modal can now be sorted by version in the expected way.

Parsely:

  • Individual Configurations can be more easily selected within the Collections panel of the File Setup modal.
  • In some cases where the Current Step was selected in the Configurations panel of the File Setup modal, configurations not on the current step were displayed. This has been fixed.
  • In some cases where the Current Step was selected in the Configurations panel of the File Setup modal, configurations on the current step were not displayed. This has been fixed.
  • If user uploads data to multiple steps while in Flow Mode, both screens are now refreshed automatically after all data has been uploaded from the file.
  • Some legacy data files were not appearing in the FILTER drop-down. This has been fixed.
  • Viewers are no longer able to open the file upload modal.
  • Correct units are now displayed for numeric Properties associated with Components.
  • Calculated Properties are once again excluded from the Property Rules interface.

Analyze:

  • Run paths have been regenerated for all experiments to ensure consistency with designed export behavior.
  • There were edge cases where combinations of single- and multi-valued data were not exported in the same form as they were being rendered in the multi-value data editor. This has been fixed.
    • If export is requested at the process level, Riffyn will utilize its previously calculated files for experiments that have not been modified since the last export request.
    • If export is requested from any one experiment, its data will be re-calculated using the latest version of our export service.
    • If you think your data files could benefit from these updates to our export service, you can align your exported data files with this latest version using the exportExperiment API with refresh = True. If you want to achieve the same outcome without the API, you must first edit your experiment in some way (e.g., add-then-delete a run on Step 1) before triggering calculation or export in the normal ways.
  • Requests to export all process data will be processed using only the experiments the user has access to.
  • Performance has been improved for Data Cleaning procedures utilizing either sliding window or exact match methods.
  • If user requests export for a file that Riffyn estimates to be larger than 2 GB, it will prompt user to Clean Data before export. 
  • Data Cleaning back-end has been re-factored to facilitate processing individual runs with >500 MB of multi-valued data. 
  • Property columns in data exported from an experiment are now in the same order as they appear in our UI.
  • Property columns in data exported from a process are now grouped together across all experiments and process versions according to their shared Permanent Step ID (S#).
  • The file_id column has been temporarily removed from exported data.
  • The event_id column for each step is now exported ahead of its first input column (as opposed to its first output column). 
  • Experiments without any runs are no longer included in export. 

Library:

  • Public entities cannot be shared except by those who have been shared directly.
  • Public Resource Types and properties are viewable.

API:

  • The updateRun endpoint has been created for editing run name and type.
  • A collection of shareX and unshareX endpoints have been created to control access to Processes, Experiments and more.
  • The listResources endpoint has been updated to allow filtering requested resources by Resource Type

v1.b360 (Version 1, Build 360) - Released August 7, 2017

Riffyn has just been upgraded in three big ways!

 

1) You can now upload data to many steps from a single file by applying Collections of Parsely Configurations:

2) You can now customize the order that properties appear in Design and Measure modes by re-ordering resource types, components and properties in Design mode. E.g., when you click an input resource type you can now drag it into a new position among the other input resource types on that step.

3) You can now use one or more single-valued properties as run identifiers when writing data from a file or database query. For example, you can use this new feature to upload plate-formatted data by using the “row” and “column” fields of a matrix data block as run identifiers when configuring your Run Rules.

 

The new release contains these additional features and bug fixes.

 

New Features 

General:

  • Browser tab names now mirror page content. Also, the fav icon is either Library green, Design blue or Measure blue depending on what kind of page is being viewed.
  • For multi-line text boxes the ENTER key will now always be interpreted as a line break.
  • New and current users are now asked to read and accept a disclaimer regarding the content they might upload to or download from Riffyn.

Design & Measure:

  • Exported data files can now be deleted. First open the “Data Export Library” modal by selecting “View All” from the “Data” menu in the upper right hand corner of either Design or Measure mode. Then click on the trash can icon that appears when you roll over the file you want to delete.

Design:

  • You can now re-order the tasks of your process no matter what you have selected in the Design Canvas. Previously you could not re-order tasks if a group was selected.

Measure:

  • You can now use the ESC key to close the multi-value editor and resource viewer panels that take over the Design Canvas.

Library:

  • New filters in the sidebar allow you to view only the Processes and Experiments that have been shared with you via your membership on various teams.
  • Resources are now shared with your Org as Designer so that other members of your Org can edit their fixed property values. Previously resources had been shared with Org as Viewer.

 

 

Issues Fixed

Design & Measure:

  • When you are editing a task you can now highlight its text without activating the task re-ordering feature.
  • On the procedural task panel, several changes to the keys and interactions were made, including:
    • Edit is now enabled with Double click. Single click will no longer enable editing.
    • Current task indicator interaction area has been increased.
    • Edits are automatically saved if user changes which step is viewed.
  • Some issues related to media upload via drag and drop have been fixed.
  • The "Fixed Properties" tag is no longer missing from fixed system properties displayed in autocomplete lists.
  • Experiments in trash no longer show on the “Experiments” menu or the “View All Experiments” modal.

Design:

  • Clicking the process name once again opens the Rename Process modal.
  • The “reorder tasks across steps” modal has been updated to improve clarity.
  • Using the cross-step task list to re-order tasks on the same step will no longer create exception loops.
  • The Version menu in the upper right corner is now just an indicator. To view other versions of your process or to create a new version you can instead use the Versions panel in the sidebar.

 Measure:

  • Performance of experiments with thousands of runs has been improved by disabling the logic that had allowed resources to “pick up” fixed property values written to runs prior to assignment.
  • Once again you can push task changes to the working version of your process.
  • Viewers are now able to view multi-valued data in the top panel.
  • There is now a limit to the number of rows shown in the multi-value data editor. A message is shown whenever this limit has been reached.
  • Character properties can now be created when adding properties to experiments.
  • When the resource assignment panel is operated in filter mode, once again the list of resources that is automatically loaded into memory can be filtered by string matching.

Parsely:

  • An alert is now provided if users attempt to write data to fixed properties of pass-through resource types. This action is not prohibited, but it is discouraged.
  • Run identifiers can now be deleted from Run Rules.
  • Run Rules for stored configurations are once again displaying properly.
  • Data blocks can once again be deleted.
  • After data upload the bottom panel should more consistently display the updated values.

Analyze:

  • Export algorithm has been updated to improve interpretability and accuracy of exported data for processes exhibiting split-merge step graph topologies.

Library:

  • The share button on the share modal is now disabled until a first collaborator has been specified.
  • Admins can once again delete their resource types.
  • Admins on a team are now allowed to change their role on that team.
  • Data displayed as Resource Use Details was updated to make more room for process, experiment and step names.
  • Highest sharing access right should always prevail when an entity is shared via more than one team
  • Tooltips for processes and experiments will now display new lines contained in the corresponding descriptions.
  • Resouce Inventory can now be searched for resources with special characters in their names.

 

v1.b325 (Version 1, Build 325) - Released July 4, 2017

Riffyn has just been upgraded in three big ways!


1) Riffyn no longer requires you to assign resources before you write data to runs. Resource assignment is now something you do only when you choose to do so -- e.g., because assigning a resource will automatically display the values of its fixed properties on your experiment. Data entered on non-fixed properties will survive subsequent resource assignment or removal unchanged, but data written to fixed properties will no longer be displayed if a subsequently assigned resource carries a previously specified value of the same fixed property. If you find yourself writing data to fixed properties prior to resource assignment, consider using Plan mode instead. You can read more detail about fixed properties here.

 

2) When you view resources in the inventory you can now request to “Load Resource Use Details”:

Pressing this button returns a list of the experiments where the resource has been used:

Double clicking on any row of this table will open the corresponding experiment in a new tab.

 

3) We have streamlined the UX for deleting one or more entities from the libraries. E.g., after you have used shift-click to select multiple processes you can then select “Move to Trash” from any of the selected menus on the right. The deleted entities can be rescued from the Trash up to 7 days after they have been moved there.

 

The new release contains these additional features and bug fixes.

 

New Features 

Design & Measure:

Reminder: if you find yourself writing fixed property values via Parsely to input or pass-through resources, consider saving yourself the trouble. It is usually the case that those fixed property values have already been written to those resources and do not need to be overwritten.

  • All image and video files that have been uploaded to Riffyn can now be expanded to full screen.
  • There are no longer any restrictions on the kinds of files you can upload as attachments to procedure tasks, design notes or observations.
  • The “Make Samples Of…” feature has been enhanced with a complimentary and functionally equivalent option to “Make New…” For example, if your step transforms your cell culture in a flask into a set of new cell cultures in sample vials, you might ask Riffyn to issue new RifIDs for each sample by selecting “Make Samples Of…” On the other hand, if your step transforms a cell culture in a flask into a new set of cell cultures in a new set of flasks, you might ask Riffyn to issue new RifIDs for each new flask by selecting “Make New…”

Analyze:

  • Input values are no longer cleaned automatically. Exported column headers that used to read “Step | input | Resource Type | Property | value | mean” by default will now read “Step | input | Resource Type | Property | value” by default.
  • Data Cleaning interfaces have been updated to accommodate input properties as well as output properties.

 

Issues Fixed

General:

  • Once again users can view their Profiles from the User menu in the top right corner of the application

 Design & Measure:

  • Some issues encountered when reordering tasks using the cross-step task viewer have been fixed.
  • In the Attachments panel the Current Step column now shows data for all 3 kinds of attachments (Procedure, Design Notes and Observations).
  • Users with Viewer access to Processes could not upload, download or delete data files on the Experiments that were shared with them as Operator. This is now fixed.
  • All images uploaded to procedure tasks should now be displayed correctly.
  • Sharing an Experiment with a user as Operator, Designer or Admin will now grant that user Operator privileges on the corresponding Process if he did not previously have access to the Process.

Design:

  • You can no longer enter the edit state for tasks when viewing previous versions of a process. (Those edits were being saved to the working version of the process.)

 Measure:

  • Viewers are no longer presented the interface to create new runs.
  • Viewers are once again able to view run connections.
  • Parsely Configuration Preview now shows a black alert instead of a red warning if a user chooses not to write data to any properties on the step.
  • Clean Data Tab no longer shows sort carets since the rows are not sortable.
  • Units were not displayed for some properties when viewing a resource in the top panel. This has been fixed.
  • Parsely will now show an error if a run identifier has not been specified in a parse configuration.
  • All worksheets in a workbook are now accessible when defining data blocks in Parsely.

 Analyze:

  • Experiments with a split-merge run graph topology would occasionally export rows where the output resource on an upstream step was not equal to the input resource on the corresponding downstream step. This bug has been fixed in a way that has introduced gaps in the exported table where previously there were no gaps. Both representations of the data are accurate, and each has distinct advantages for data analysis. We intend soon to give users control over which form is used for data export.
  • Single-valued data are now automatically exported on all rows associated with a run, so what you see in the multi-value editor (top panel) is now always the same as what you get upon export. As part of this implementation we decided to stop exporting the “File id” columns for the time being.
  • Viewers are once again able to request calculation and export of Riffyn data files.

 Library:

  • Designers can now remove members from a team.
  • For Mac users the key mapping for selecting multiple entities in the libraries has been updated from the “control” key to the “command” key.

 

Known Issues

We intend to resolve the following issues with our next release in ~4 weeks:

  • If a sufficiently large amount of data has been written to an individual run, Riffyn will ask users to Clean Data before export. The current issue is the confusion that results when sometimes -- but not all the time -- a trivial Data Cleaning Procedure (which doesn’t actually clean the data at all) proves sufficient to export the data.
  • If a user switches the order of two tasks on the same step using the cross-step viewer he is subsequently unable to reverse the switch when viewing the same tasks on the step itself.
  • Users are unable to create new character properties while adjusting the process design in Measure Mode

v1.b292 (Version 1, Build 292) - Released June 2, 2017

Riffyn has just been upgraded in two big ways!


1) You can now download the files you have uploaded to Riffyn. For example, you can now download the very same files that your team has uploaded to Parsely for run creation, resource assignment and/or data entry. The kinds of file types that can be uploaded to Procedure Tasks, Observations and Design Notes have also been expanded to include PDFs, Excel templates, PowerPoint presentations, JMP scripts and more. Please let us know if there are additional file formats we should add to this list.

2) Experiments can now be shared independently of one another and of the access controls on the corresponding Process. If you want to continue sharing each new Experiment with everyone who has access to its Process, you can continue to do so without any extra clicks. On the other hand, if your organization desires more flexible controls regarding who has access to which Experiments for a particular Process, this feature is for you. If you want you can now invite a collaborator to view just one of your Experiments, or you can invite your team to see your latest Experiment only after you are ready to share it with them. You can read more about this feature here.

 
The new release contains these additional features and bug fixes.

 

New Features 

Measure:

  • Output resources can now be viewed in the top panel when you click on their names in the bottom panel; output resource names can also be edited in this top panel
  • Resource search mode (indicated by magnifying glass icon) will now return resources of matching name even if they are not the matching resource type
  • Parsely UI has been updated to facilitate specifying Run Rules and Resource Rules

Analyze:

  • New experiments are now part of the API tasklist.

 

Issues Fixed

Measure:

  • After adding a run, generated resources would not show on the bottom panel until it was refreshed. This has been fixed.
  • Performance issues that arose when handling thousands of runs have been addressed:
    • Loading guard screen should no longer hang after propagating thousands of runs
    • You can now delete thousands of runs much faster
    • You can now more quickly adjust which steps or groups are viewed in the split-panel viewer
  • When opening the multi-value data editor for large data tables the system did not display the loading screen. Now regardless of data table size the screen will display the loading indicator while it loads the data.
  • In the Plan tab there was no loading indicator while loading runs. Now a loading indicator is used while the run list loads.
  • Fill Down will now save pending edit to focused cell.
  • First click on any cell in bottom panel now highlights entire string.
  • In Parsely, when trying to add a new Output Resource the application would instead add a new Input Resource. This has been fixed.

 

Known Issues

We intend to resolve the following issues with our next release in ~4 weeks:

  • Although fixed properties are intended to be single-valued, users are in some cases allowed to write more than one value to a particular run. When this happens the value displayed in the bottom panel is not always the same as the value exported in the CSV.
  • If a sufficiently large amount of data has been written to an individual run, Riffyn will ask users to Clean Data before export. The current issue is the confusion that results when sometimes -- but not all the time -- a trivial Data Cleaning Procedure (which doesn’t actually clean the data at all) proves sufficient to export the data.

 

v1.b272 (Version 1, Build 272) - Released May 5, 2017

Riffyn has just been upgraded in three very big ways!

1) You can now create, delete and edit runs within the “Record Data” interface. You can even create connected runs within the “Record Data” interface, or simply view the data for two steps at one time, side by side. Also, this split-screen interface can now be entered (or its step selection changed) by clicking on the anchor icons that have been added to the resource type headers in the bottom panel:



2) You can now sort the runs in the bottom panel by the values of any column, even Run Type or Resource Name. If you wish to sort on both a primary column and a secondary column, first sort the runs by the values of the secondary column and then again by the values of the primary column.

 

3) You can now view tasks on multiple steps simultaneously and even create detailed workflows that deviate from the higher level material flows and step order emphasized by the Design Canvas. Please click on any of these hyperlinks to learn more about these new features:



Additional New Features

General:

  • User profile information is now available by clicking on the user icon in the top right corner of the application and selecting “Profile”

Measure:

Library:

  • We have improved the data entry for Fixed properties in the Resource Details modal, allowing for faster actions when adding several properties.

Analyze:

  • Columns are now exported in Step Order. If you are using JMP for data analysis, be sure to take advantage of our RiffynTools AddIn. If you are using Excel to explore your data table, check out our Process Data Navigator.
  • Calculated inputs are now calculated for export. In upcoming releases we will allow viewing calculated inputs in the “Clean Data” panel and stop automatically summarizing input properties to a single mean value.
  • Timeout when preparing data for export has been increased from 15 minutes to 60 minutes. If your experiment is somehow so big and so in need of refreshing that you hit the 60 minute timeout, click on the red refresh button to clear the error state and then click again on the gray refresh icon to try again -- Riffyn will pick up pretty much where it left off and hopefully get your data to you before it times out a second time. Please write us at support@riffyn.com if ever you think something has gone wrong.

 

Issues Fixed

General:

  • Under certain circumstances the screen resolution warning would pop up despite checking the Don’t show this warning again option. This is now fixed.
  • In certain cases with a bad internet connection the app would retry operations until successful. To prevent undesired side effects the application will now send an error message if long running operations are unable to finish due to connectivity issues.

Design:

  • Whatever step you have selected is now retained when changing process version using the sidebar.

Measure:

  • Performance has been substantially improved when assigning resources via the top panel, when propagating resources in the bottom panel and when using the Fill Down button or keyboard shortcut (ctrl-d on PC; cmd-d on Mac) in the bottom panel. These enhancements are especially noticeable when operations are performed on >2000 runs at one time.
  • The bottom panel will now refresh automatically after uploading data from a file or database, even when there are thousands of runs.
  • Keyboard navigation in Plan mode is now working the same as in other tables.
  • Calculated Properties no longer appear in the “Property Mapping” section of Parsely.
  • Users attempting to add properties to their experiments can now create new properties.
  • Null dates are no longer displayed as “invalid date”.
  • Users can once again use the resource finder in the top panel to assign resources of non-matching resource types.
  • Measure mode was preventing data entry for properties that were once, but are no longer, calculated properties. This has been fixed.
  • Each time you create a new configuration in Parsely, the default configuration name is now being updated correctly.
  • In Parsely you can now update the dimensions of a data block without losing edits you have previously made to its field data types.

Library:

  • When duplicating an experiment on the library the screen will show a guard screen while the experiment is cloned.
  • On environments configuration, when requesting a metadata update the loading screen is now dismissed automatically when the update is done.

 

Known Issues

We aim to resolve the following issues with our next release in ~4 weeks:

  • Multi-valued input data should not always be averaged on export
  • Calculated inputs are not yet available in our UI as Clean Data
  • Under some circumstances collaborators are unable to assign resources using a Parsely configuration that someone else had made
  • Although fixed properties are intended to be single-valued, users are in some cases allowed to write more than one value to a particular run. When this happens the value displayed in the bottom panel is not always the same as the value exported in the CSV.
  • Process Designs that exhibit a split-merge topology sometimes export extra rows containing non-existent run connections
  • If a sufficiently large amount of data has been written to an individual run, Riffyn will ask users to Clean Data before export. The current issue is the confusion that results when sometimes -- but not all the time -- a trivial Data Cleaning Procedure (which doesn’t actually clean the data at all) proves sufficient to export the data.

 

v1.b242 (Version 1, Build 242) - Released April 2, 2017

Riffyn has just been upgraded in three very big ways!


1) The interface for uploading data from files has been completely reworked and enhanced and is now officially named Parsely. With Parsely you can now...

  • write file data to run attributes like run description, run type and run start/stop
  • write file data to the correct runs using your custom run descriptions (you can still use Riffyn's run numbers if you want to, but you have more choices now!)
  • assign resources from your resource inventory using file data
  • create connected runs between steps using file data
  • write resource names for output resources generated on the current step
  • write the same set of values from your file to multiple properties in Riffyn
  • write file data to multiple run groups in a single transaction
  • do all of these things with just one click when you APPLY a stored configuration to a new file

 

We have also created a collection of detailed help documents for the updated Parsely (Please note that if you click on the hyperlinks in this e-mail, Zendesk will ask you to log in before reading them. If you have not done so already, please "sign up" for your own Zendesk account -- which is distinct from your Riffyn account.):

If you have stored configurations that you have been using to upload file data you may need to update them briefly before you can use them. If you don't change anything else about your existing data pipeline, you may still need to visit Parsely's Run Rules section to map the file data field containing your Riffyn Run IDs to "Run | Number" as a Run Identifier. That should do it...

 

2) We have finally combined "Record Inputs" and "Record Outputs" into a single "Record Data" interface where you can do single-valued data entry in its bottom panel and multi-valued data entry in its top panel. We have also made "Record Data", instead of "Plan Experiment", the panel that you see each time you open Measure mode in your browser.

You can access the multi-valued data editor in the top panel in either of two ways:
right click on any cell in the bottom panel and select "Edit Data", or

click on the yellow indicator that appears on any multi-valued cell in your experiment

As part of this consolidation you can now edit multiple output values using the same Fill Down button that we recently introduced for Recording Inputs. “Fill Down” can also be performed using ctrl-d (PC) or cmd-d (Mac). Please note that Plan mode will continue to use its distinct Edit Multiple feature for a little while. That feature can still be accessed using the pencil button that appears along the bottom of your browser when one or more runs have been selected in Plan mode.

In addition to these 2 big upgrades, the new release also contains the following features and bug fixes.

 

3) We have re-worked the UX for data export to give you more control over the time it takes to export experiment and process data to a CSV file.

The architecture that allows Riffyn to broadcast your latest data instantly to all of your collaborators’ web browsers requires that same data to be substantially re-shaped before we can export it to you as a big, beautiful statistical data table. Already Riffyn has been engineered to accelerate that re-shaping by focusing its calculations towards the data that has been edited since the last time these data tables were calculated. Now you can refresh that re-shaping calculation any time and as many times as you want, without the extra work of of naming and creating a file. When you finally need that data table, Riffyn can export it to you more quickly than if it needed to re-shape all of your data from scratch.

The upper right hand corner of each experiment and process now has a data status indicator to the right of the help button. A gray refresh button indicates that some data needs to be re-shaped before the whole data table can be exported:

Clicking directly on this refresh button will initiate the re-shaping without requesting that data be exported to a CSV file. The refresh button will turn orange and spin as your data is re-shaped:

If no data has been edited since the last time the re-shaping calculation was completed successfully, the data status indicator will now appear as a white download button:

Clicking on the white download button will then initiate the familiar data export process. After you specify the name of your exported file, the white download button will turn orange as your CSV file is prepared, and then green when it is ready for download:

Your CSV file can then be downloaded with a final click on the green download button or by selecting its name from the Data menu. The download button will revert to a gray refresh button as soon as the next edit is made to your experiment or process.

We have also updated the experiment and process libraries so that you can quickly review the state of each experiment and process at the moment when you opened or refreshed these libraries:

If an experiment has a refresh indicator, its data will need to be re-shaped before it can be exported as a CSV. If a process has a refresh indicator, you can now figure out which of its experiments needs to be refreshed: first filter or sort the experiment library by process name, then review the data status indicators of its experiments.

 

Additional New Features

Design & Measure:

  • You can now initiate data export from the data menus as well
  • Experiments are now ordered reverse chronologically when you select "view all experiments" from the data menus
  • Design notes and observations are now listed from top to bottom in chronological order

Measure:

  • Fixed properties can now be edited in the bottom panel
  • A resource’s name or Rif ID can now be copied to the clipboard with a right click on the bottom panel

  • Procedure task order is now indexed across an entire process, in accord with its step order
  • You can now navigate between steps and across group boundaries using the new backward and forward buttons at the top and bottom of each step’s Procedure task list

  • Run group size limit has been increased from 1600 to 4000 runs
  • Selected runs are now highlighted in yellow
  • Riffyn will no longer append text to the Run Description when creating connected runs; this change will make it easier to use the Run Descriptions you upload to one step as the Run Identifiers you use to upload file data to a second step
  • When you use the "+" to create multiple runs in Plan mode, you can now choose to index each of the generated run descriptions with an appended number

Library:

  • The Units modal has new help hints for its fields
  • Org Admins can now create "governance teams" that have access to all entities within their orgs.

 

Issues Fixed

Measure:

  • Occasionally stored override data in Plan mode would not be visible on the bottom panel when an experiment was first opened. Also, if an override had been stored and was later removed, the process target value was not always visible in the “Record Outputs” panel. Both issues have been fixed.
  • Data flow is now working when the same fixed property is associated with multiple components on the same resource
  • All interfaces for renaming experiments and naming new experiments are again working properly
  • View Step Order and Show Step IDs can now be requested from the Measure menu

Analyze:

  • Experiments with runs that branch-then-merge no longer have gaps in their exported data tables

Library:

  • Search is now working with special characters
  • Sharing is now working with e-mails that contain special characters

 

Known Issues

  • The bottom panel does not always refresh automatically after writing a large amount of data from a file. Toggling to/from Plan or to/from another step should resolve the issue.
  • Multi-valued input data is automatically averaged on export
  • Output resources cannot yet be viewed or their names edited in the top panel
  • Calculated properties that were visible in the “Record Outputs” panel are not yet visible in either the bottom or top panels of “Record Data”, but they can still be viewed in the “Clean Data” panel and in exported data
  • Calculated inputs are not yet available in our UI or in exported data tables

 

v1.b202 (Version 1, Build 202) - Released March 2, 2017

New Features

Design:

  • Permanent step IDs are now hidden from the upper left-hand corner of each step by default. When needed these permanent step IDs can still be accessed by (a) right-clicking on a step and selecting “Show Step ID”, or (b) viewing the step order in the bottom panel. Other parts of our UI where you can still find references to these permanent step IDs will soon be updated accordingly.

Measure:

  • Procedure tasks can now be edited while Experiments are executed. Admins, Designers and Operators can unlock the task editor by clicking on the new settings button found at the right side of the sidebar:

         

  • Admins and Designers can even push edited task lists to the working version of their processes:

         

  • Parse configurations stored on steps of linked processes will still be available after updating to latest version.
  • Opening the original process from a linked process will now take you to the working/editable version of your process.
  • Riffyn's Edit Multiple feature has been replaced with a Fill Down button for Recording Inputs. “Fill Down” can also be performed using ctrl-d (PC) or cmd-d (Mac). Please note that Plan mode will continue to use the Edit Multiple feature for a little while. That feature can still be accessed using the pencil button that appears along the bottom of your browser when any runs have been selected in Plan mode.
  • Record Inputs, like Record Outputs, now supports a multi-valued data editor in the top panel. You can access this interface in either of two ways:

    1) right click on any cell in the bottom panel and select "edit data", or

         

         2) click on the yellow indicator that appears on any multi-valued cell in your experiment

       

  • Enter, tab and arrow keys have been mapped to more familiar behaviors in Record Inputs

... going forward, think of Execute's bottom panel as your single-valued data editor and its top panel as your multi-valued data editor. We have posted more detail about single- and multi-value editing here. Please note that we are planning soon to unify “Record Inputs” and “Record Outputs” into a single “Execute” panel with the same single- and multi-value editing interfaces of the updated “Record Inputs” panel. Let us know what you think!

Library:

  • You can now customize the message that is sent to your collaborators when sharing.
  • "Share" and other actions can now be applied to multiple entities in the library all at once. Use a shift-click to select contiguous entities and ctrl-click (PC) or cmd-click (Mac) to select separated entities. "Delete" is unavailable for multiple Process and Experiment entities. "Remove from my library" is unavailable for multiple Experiments. Organization admins can also Enable or Disable multiple users at the same time.

 

Issues Fixed

Design:

  • Share process from design mode and experiments is working again.
  • Units can no longer be created with empty values.
  • Using the Add unit for a property will autofocus on Add unit field instead of the Name field. This prevents accidental unit rename
  • If a user with view-only permissions opens a process they will no longer see the edit buttons.
  • From a duplicate process it is now possible to open a past version of the process.
  • Pictures and videos can once again be removed from procedure tasks.

Measure:

  • Operators will no longer see the Push to Working Version button in its active state. Their permissions have been updated as well to allow most actions..
  • Exported data file default name will once again have a valid extension.
  • Data entered when the bottom panel is in the split-screen “flow” mode will now be saved.

Library:

  • Sharing a Resource entity on the library with a collaborator was sometimes not refreshing properly when sharing and removing collaborators.

 

Known Issues

  • Multi-valued input data is automatically averaged to a single-value on export.
  • Calculated inputs are not yet available in our UI or in exported data tables.
  • Context menus are not available in Design mode for steps belonging to linked processes.
  • There is no way to show permanent step IDs in Measure mode if a process has been constructed completely from linked processes.

 

v1.b189 (Version 1, Build 189) - Released February 1, 2017

New Features:

  • It is now possible to re-number the steps in the order that your team executes them. The customizable step order tags appear within circles on the top left corner of each step. The permanent step IDs that have been written to the top left corner of each step can now be hidden by right-clicking on any step that is not part of a linked process. (Soon they will also be hidden by default.) The same context menus can also be used to access a new “Step Order” interface. In Design mode, the drag handles in the Step Order interface can be used to re-number the steps however you like. This interface can also be used to navigate directly to any of the steps of your process, no matter how extensively you have grouped them. Please note that experiments created prior to this release, as well as duplicates of these experiments, should have either missing or empty step order tags at various locations within Measure mode.
  • Repeated data export has been accelerated by focusing re-calculation on only the steps that have been modified since the last data export event. If data export times-out at 15 minutes, you might try submitting a second request as everything that was calculated during that first request will not need to be recalculated.

 

Issues Fixed:

Design:

  • Selecting a character property using the Enter key will now move cursor to its Target field.
  • Formula of calculated property can now be viewed on old versions of a process.
  • Procedure tasks are now autosaved when user clicks away from the task being edited.
  • When selecting different versions of a process the bottom panel will update accordingly.
  • Change component functionality is now working as intended

Measure:

  • Data upload indicators added to steps of linked process can now be pushed to the parent version of the linked process itself.
  • Restored file configuration wizard step names.
  • Can now select run path even when connected runs have been selected.
  • Corrected error message that appears when user attempts to edit calculated property in Record Inputs
  • Complete Run Plan for any step -- both run labels and planned property values -- can now be exported by clicking “export runs” button, then “Download Run Data”

Analyze:

  • Data export algorithm has been updated to address edge cases associated with unassigned resources and passed-through resources. Users need no longer use the “Make Samples Of” feature in many cases where we have recommended it to them previously. There are still known issues associated with process designs exhibiting split/merge topology.

Library:

  • Team members can now open processes shared with a team they belong to.
  • Processes duplicated from a public process now show the full version history
  • Duplicating a process no longer takes you to the first page of the process list.
  • When a new team is created it will automatically show on the list of teams. No other action needs to be taken.
  • Edit resource modal has been shortened to fit most screens.
  • Edit property modal now allows changing the property type.

 

Known issues:

  • Experiments that were created prior to this release and duplicates of these experiments will have empty or missing step order tags at various locations in Measure mode.
  • The Step Order interface cannot be opened by right clicking on the steps of a linked process. If you want to re-number the steps of a linked process you must start by right-clicking on any non-linked step of the process.
  • Permanent step IDs cannot be shown (or hidden) by right-clicking on the steps of a linked process. If you want to view the permanent step IDs of a linked process you must start by right-clicking on any non-linked step of the process.
  • Exported data tables can be improved for process designs the exhibit a split-merge topology (e.g., when a single “media preparation” step is used to prepare individual batches of media for each of 2 downstream steps)

 

v1.b163 (Version 1, Build 163) - Released January 12, 2017

Issues fixed:

  • Minor improvements to media file upload, adding captions and error reporting
  • Minor style improvement to file counter

Known issues:

  • Record outputs tab guard screen sometimes hangs after database query
  • Run join incorrect for process with split output resources

 

v1.b157 (Version 1, Build 157) - Released December 23, 2016

New features:

Design & Measure:

Data upload indicators

  • Steps will now report how many times data has been written to them via files or database queries
  • These indicators can also be used to tag steps where the process designer anticipates uploading data via files or database queries. These indicators are available from each step’s context menu as:
    • Require Data Upload from a File
    • Require Data Upload from a DB Query
  • Clicking on these indicators in Measure mode will open the file upload modal or database query modal, as appropriate
  • Various improvements to media uploading and display

Library:

  • Units can now be created and edited through the library

 

Issues fixed:

Measure:

  • Issues with resizing the lower pane in Plan Experiment are fixed. Now all runs should show up when the resizing happens.
  • Users with the role Viewer ca no longer see the panel where Operators can add resource types, components and properties to their experiments.
  • Viewers can once again see the Record Inputs tab
  • Fixed Properties are now working properly on Record Inputs.
  • Guards were not disappearing throughout  the app under certain conditions. This is now fixed except for Known Issue noted below.
  • Scientific notation with 4 decimal places is now employed for any numbers entered >999999 or <0.001. Note that some values will be zero padded (e.g 0.000123 as 1.2300e-4) until a future update that addresses numeric precision.

Design:

  • Calculated properties will now work as expected when a process is shared with someone outside its designer’s org
  • Custom properties and units will display as expected when a process is shared with someone outside its designer’s org
  • Remove Export option on the context menu is working again.
  • Remove Resource option on the context menu is working again.
  • Add quick step is working again (Alt+click)
  • Component propagation from input to output is now working as expected. If a component is propagated and both input and output matching resources exist the app will do the propagation.

Measure and Design:

  • There were several inconsistencies with dates on the version and experiment dropdowns. These are now fixed.
  • Minor UI fixes on the Delete Process dialog. In some particular cases the text would not be aligned properly.

Library:

  • Users outside of your org can be added to a team if you already know their e-mail addresses
  • Team members can now be added or removed from a team by a Designer
  • Team roles can now be changed after the team is shared.

 

Known issues:

  • Record outputs tab guard screen sometimes hangs after database query. 
  • Run join is incorrect for process with split output resources
  • Responsiveness of the UI can sometimes become sluggish on large process after using for a period of time.  Page refresh will clear the issue.

 

v1.b144 (Version 1, Build 144) - Released December 9, 2016

Features

Design & Measure:

  • Pictures and video up to 100MB can now be added to procedures, notes and observations in both Design and Measure mode
  • Formulas that are "broken" due to deleted precedent properties are now highlighted in red and ignored in calculations.

Measure:

  • Edit resources and properties on active experiment.  If you started and experiment, and forgot to add a property or resource during the Design phase - that's no problem anymore.  No you can simply add it on-the-fly and start collecting data without having to return to a process design.  You can also promote those on-the-fly changes to the working copy of the process Design so they are there for every future experiment.
  • Security upgrades on the data agent. Agent key is now shown only once on agent creation.
  • Data export is now 5X faster. 

Library:

  • Experiments list can now be filtered by experiment ID

API:

  • An API is now available at https://api.riffyn.com.  Documentation and SDK available there.

 

Bug Fixes

Design:

  • Text entered into Procedure tasks is now autosaved when clicking away from the task.
  • Image upload to tasks now places those images on the correct step.

Library:

  • Fixed inconsistencies in style and key input handling to improve the user experience.
  • Fixed glitches that prevented experiments from deleting on the first try.

Measure:

  • Fixed date/time offset issues.  Now what you what you enter is what you get.
  • "Guard" screens (which show will Riffyn is "thinking") are now properly removed in certain cases where that wasn't happening before.
  • Scrolling in the Measure mode data table no longer clears the data from view.

Sharing:

  • Teams shared with you will now be available to choose from on the share modal.

 

Known Issues (to be fixed in next release)

  • Agent status indicator does not changes from green to gray when agent goes offline
  • Run join incorrect for process with split output resources
  • The user cannot propagate a Component to the upstream in a flow of steps with resources type of the same type.

 

v1.b138 (Version 1, Build 138) - Released November 14, 2016

Features

Measure:

  • Restructured run data tables store 70% less data. This will allow a tripling in the number of data points that can be viewed/exported as Raw Data on a single run before requiring the application of a Cleaning procedure. (Eventually, the view/export data limit will be removed entirely.)
  • Fixed dangling tooltips
  • Null values are shown as blanks, zeros shown as zeros.
  • Run counter on the "Record Inputs" tab is fixed.

Bug Fixes

Design:

  • Group descriptions now use the full height of the sidebar.
  • Some issues with grouping steps have been fixed.
  • Repaired context menus that failed to appear on inputs/outputs.
  • It is now possible to create new resources when changing components.
  • When duplicating a process, the currently-viewed version will be duplicated instead of the working version.
  • Units and properties creation is repaired.
  • Adding more than three variables to a property formula auto-scrolls the variable list

Library:

  • Full URL path to library pages is now shown in the browser, allowing bookmarking
  • Browser back button now works in the Library
  • Many small UI/UX fixes to the modals for editing entities
  • Made deletion of processes and experiments a little bit more user-friendly
  • Various fixes to public processes

Measure:

  • The components name is now shown on the header row of the run data editor on Record Outputs
  • Guard screens on bottom panel are now working (they are closed when they should be)
  • Standardized date input / display format (YYYY-MM-DD HH:MM:SS) throughout

Known issue:

  • In Measure->Plan, entered dates are offset by (Current Timezone - GMT) hours.  This will be fixed in the next release.

 

v1.b130 (Version 1, Build 130) - Released October 26, 2016

Measure > Plan & Execute:

  • “High-level” view has been removed from application; in a future release we will introduce new interfaces for summarizing plans and specifying default values for created runs.
  • “Constant” and “Variable” designations that had previously restricted data entry have been removed from application; in order to assign a single value to all runs, use the existing “Edit Multiple” feature: first select runs using check boxes, then click on pencil icon in ACTIONS bar at bottom

  • “2A. Execute: Record Inputs & Deviations” has been renamed as “2. Execute: Record Inputs”
  • “2B. Execute: Collect Data” has been renamed as “2. Execute: Record Outputs”

Sidebar > Variables:

  • “Constant” and “Variable” designations that had previously restricted data entry have been removed from application

Measure > Record Inputs:

  • Clicking on the source button in the Resource header will now navigate to that upstream step, even if the steps are located in different groups.

Library

  • Resources can now be created from the Library
  • Experiments are now be viewable on public processes

Bug Fixes

Design > Resource/Property Editing

  • Fixed alignment, truncation and spacing issues
  • Empty procedure tasks can no longer be added. Empty tasks are both imageless and textless tasks.
  • A formula with errors will no longer be saved.
  • Circular references in formulas are now considered errors.
  • Units can now be added to newly created properties.
  • Under certain circumstances the calculated properties modal did not close. This is now fixed.

Measure mode > Resource assignment

  • Newly created resources now appear in the search list immediately
  • Deleted resources will not show up on the global search

Measure Mode > Plan & Execute

  • Edit buttons are now to the left and right of the edit-multiple strip.
  • The Raw Data tab (Now called "Record Outputs") remains accessible, even when the data table is too large to view. This allows uploading additional data.
  • Fixed issues with the display of source buttons on some input resources
  • Fixed date-time formatting issues & incorrect offsets on Record Outputs and Clean Data tabs
  • Various fixes to the display of guard screens in the bottom panel

Library:

  • Fixed close and cancel buttons on modals.
  • Resources: Resource search button now correctly triggers a new search.
  • Resources: Moved “New Search” option to the right.
  • Teams: Fixed overlapping text between name and email of team members in team modal

Ontologies:

  • Clean up of some default entities in ontologies

 

 

v1.b115 (Version 1, Build 115) - Released October 9, 2016

1. Partial fix to issues with date/time properties of resources

  • Fixed: Date/times upload now assumes timezone of the browser.  Date/times display in browser in time zone of the browser.
  • Not yet fixed: editing date/time in Tab 1 (Plan) still assumes GMT, inconsistent display of date/time (should be YYYY-MM-DD HH:MM:SS), handling of times without dates

2. Resource editing in the Library is restored
3. Restored missing “guard” screens after executing long-running operations
4. Support for exponential number format (both ‘e’ and ‘E’) in data upload
5. Fixes to some unitless properties

 

v1.0 - Released August 23, 2016

This is Riffyn's first commercial release.  Yahoooo!  Two-years of blood, sweat and tears to bring you computer-aided design and analytics for the scientific method.  We hope it helps you discover the next big thing with less effort, in less time.

Was this article helpful?
0 out of 0 found this helpful
Have more questions? Submit a request