We are aware of the following issues and limitations in v33.0.0. We expect to address them in future versions of Hyperscience.
User Experience
Drop-down menus’ text in IE 11 – The text in drop-down menus is not centered in IE 11.
[Fixed in 33.0.1] Making selections in dialog boxes – When a user attempts to select text in dialog boxes, the selection is automatically cleared.
Using calendars in date filters – Using the drop-down menus in calendars to select different months sometimes causes the calendars to automatically close.
Workaround: To select different months, use the Previous (
) and Next (
) buttons that are located next to the drop-down menus.
Layout Editor
[Working as designed] “Multi-select contiguous items” keyboard shortcut in the Layout Editor – Using the Shift + Click shortcut does not multi-select contiguous items if a field is only selected but not expanded.
Workaround: Expand a field, and then use the Multi-select contiguous items keyboard shortcut.
“Duplicate” and “Delete” buttons in the Layout Editor – The Duplicate and Delete buttons are disabled if a field is only selected but not expanded.
Workaround: Expand a field, and then use the Duplicate and Delete buttons.
Pasting fields from another instance to a layout – If you copy fields from a layout in one instance and paste the same fields to a layout in another instance, these fields won’t be created in the destination instance.
Identification Supervision setting for table columns – The Identification Supervision setting is non-editable and enabled by default for table columns in Semi-structured layouts. Instead, the Identification Supervision setting should be configurable.
Empty field in a shared fields list – If there is an empty field in a shared field list, you are unable to click the Commit changes button in the Layout Editor.
Layouts
Page matching and layouts with the same name – If two Semi-structured layouts exist with the same name in different flows, and a submission is sent to both flows via the application UI with that layout name selected, one submission is matched to the layout and the other is marked as No Layout Found.
Potential Layout Variations
“Run” and “X” buttons on the Run Potential Layout Variation Job page – The Run and X buttons on the Run Potential Layout Variation Job page are displayed very close to each other.
Icon for canceling a Potential Layouts job – The icon for canceling a Potential Layouts job is a trash can, which can be confusing for users.
Extra scrollbar on the Potential Layout Variations Viewer page in IE 11 – An unnecessary scrollbar appears on the Merged tab of the Potential Layout Variations Viewer page.
Two scrollbars on the Run Potential Layout Variation Job page in Chrome and IE 11 – Two scrollbars appear on the Run Potential Layout Variation Job page in Chrome and IE 11.
Models
Deploying version-incompatible Table ID models – Deploying version-incompatible Table ID models results in receiving an “Unexpected error” message. Instead, you should receive a descriptive error message.
Training
Starting a model training job without a trainer – If you start a model training job without a deployed trainer, you will receive a message that training has started, even though the training is only queued and cannot be completed without a trainer.
Training and deploying a Field ID model for the first time – When you train a Field ID model for the first time, you cannot deploy it.
Layout fields with the same name and model training – If at least two layout fields share the same name, model training fails.
Releases
[Fixed in 33.0.1] “Current Release” column and “Assign to Flow” button in the “Assign to Flow” dialog box – If you have upgraded your Hyperscience application consecutively from v31 to v32 to v33 and have assigned a release to a flow in v31, your Assign to Flow dialog box in v33 has an active Assign to Flow button, even though your release is assigned to all flows. The Current Release column’s value is “N/A” for all flows.
Creating releases with no layouts – The Create button for creating a release is active even if you haven’t added any layouts. Clicking the Create button in this case causes an empty locked release to be created.
Flows
Importing Flows with invalid JSON file UUIDs – Currently, if an imported flow contains a JSON file with an invalid file UUID, an error message appears, which says that a valid JSON file must be present. However, there is no indication that the issue relates to the UUID specifically.
Line numbers when showing code differences – When importing a new version of a flow, the line numbers in the JSON code comparison stop before the end of the file if the JSON file is longer than 600 lines.
Error message when importing flows – If the first bracket is missing from a flow's JSON file, a 500 error ("internal server error") occurs, but that is not the correct error code for the error.
Validating flows and blocks with identical identifiers – If a flow's identifier is the same as a block's identifier in that flow, validation may produce incorrect results.
Trimming whitespace in settings – The trimming of whitespace in flow and block settings is not consistent and not apparent to the user upon saving the flow.
Submission Processing
High-volume instances with MSSQL databases – We're working to resolve an issue that prevents the Submissions and Tasks pages from loading when a large number of submissions are present. This issue affects instances with MSSQL databases.
Submissions
Selecting the first submission on the Submissions page – If you select the checkbox for the first submission on the Submissions page, the checkbox for the entire table is also automatically selected.
"View Flow Execution" for submissions without flows – The "View Flow Execution" option on the Submission Output page should not be available if a submission was processed outside of a flow (e.g., it was processed in v28 or earlier).
Documents
Missing pages in the document viewer – If pages are missing from a document, any missing pages that come after the first present page in the document should be marked as "Not Present" in the document viewer, but they are not.
Cases
"Per Page" drop-down list in IE 11 – On the Case Details page, when the Per Page drop-down list in the Data card is clicked, white space is added at the bottom of the page to accommodate the length of the list.
Document Classification
“Unsorted Documents” header’s location during foldering – The Unsorted Documents header’s location is inconsistent when you scroll the page during the foldering task.
Table Identification
Table ID Supervision tasks that take more than 4 minutes to complete – Table ID Supervision tasks that take more than 4 minutes to complete are abandoned and re-assigned to another keyer.
Template-row identification
Template-row identification for non-gridded tables in multi-page documents – Template-row identification for non-gridded tables is only possible on the first page of a document.
Table-identification review
[Fixed] “Next row” keyboard shortcut during table-identification review in IE 11 – Pressing the ⬇ key during template-row identification does not select the next row in IE 11.
“Redo” keyboard shortcut during table-identification review in IE 11 – The Redo keyboard shortcut starts working only after you press Command + Shift + Z or Control + Shift + Z a few times.
Table Transcription
Copying and pasting a cell’s transcription value – If you copy a cell’s transcription value and press Enter on another cell before pasting the copied text, you will actually paste a different text from the one you’ve copied.
Workaround: Paste the copied text without pressing Enter.
Zooming in with a trackpad during Transcription tasks in IE 11 – Zooming in with a trackpad during Transcription tasks in IE 11 redirects you to the upper-left corner of the page you’re viewing.
“Override Validation” keyboard shortcut during Table Transcription – If a keyer adds at least one symbol to a cell’s transcription value and then presses F9 to override validation, the symbols added to the transcription value will be lost.
"Loading" indicator during Table Transcription tasks – Depending on your browser window’s size, the “loading” indicator’s location during Table Transcription tasks might be inconsistent.
Transcription of table cells with lower confidence than Table Minimum Legibility Threshold’s value – Table cells with transcription confidence that is lower than the Table Minimum Legibility Threshold’s value do not have any exceptions. Instead, such table cells should have Illegible Field exceptions.
Entering table cell transcription and scrolling the page – If you scroll the page while entering a transcription value for a table cell, the value you’ve entered will be cleared.
Deleting a row during Table Transcription when no columns are selected – When no columns are selected during Table Transcription, the Delete a Row link is still active instead of being inactive. Clicking the Delete a Row link results in an “Unexpected error” message.
Custom Supervision
[Fixed in 33.0.1] Custom Supervision in “Status” filters – On the Submissions and Documents pages, there is no filter option in the Status filters for Custom Supervision.
Focusing on pages in tall browser resolutions – When completing Custom Supervision tasks in browsers whose resolutions are taller than they are wide, clicking on a crop in a different page does not always shift the focus to that page.
User experience in IE 11 – We're working to resolve the following issues that occur when completing Custom Supervision tasks in IE 11:
The Complete task button may overlap text in the right-hand sidebar.
Clicking a drop-down list at the bottom of the page causes extra white space and an extra vertical scrollbar to appear.
When the right-hand sidebar contains three tabs, a horizontal scrollbar appears, which may hide some page elements. The background color of the tabs may change part-way across the right-hand sidebar.
Long external case IDs on the Cases page are cut off instead of wrapped when a small resolution is used (e.g., 1366x768 pixels).
"Continue & Don't show this again" button and saving choices – When a user clicks "Continue & Don't show this again" on the instructions page, some decision choices made by the user during Custom Supervision may not be saved.
Saving case decision choices – If a case contains multiple submissions, the case's decision array contains only the case choice made for the first submission added to the case.
Loading times for tasks with a large number of crops – Users may experience delays when loading pages and interacting with crops if there are a large number of crops in a task.
"Validation Override" exceptions – If a transcription field's validation is overridden in a Custom Supervision task, the "Validation Override" exception is not recorded for the submission.
[Fixed in 33.1.9] Values of fields with overridden formats – If a transcription field's format is overridden before it appears in a Custom Supervision task (e.g., during Machine Transcription), the field is empty when it is loaded in Custom Supervision.
Document Output
“Output Value Correct” column for table cells in the Document Output page CSV file – Currently, the Document Output page CSV file does not include an “Output Value Correct” column for table cells.
Flow Execution
Routing Blocks in Flow Run diagrams – We plan on changing the way Routing Blocks and their connections are displayed in Flow Run diagrams. For example, some connections between Routing Blocks and other blocks may be shown as a broken line. Also, unnecessary connection lines may appear if there are multiple nested Routing Blocks.
Resizing the top and bottom Flow Execution panes – When resizing the top and bottom panes of the Flow Execution page, the Flow Run diagram pane is automatically scrolled back to its beginning.
Actionable area of resizing tool – The actionable area of the resizing tool for the top and bottom panes of the Flow Execution page is misaligned and extends above the tool.
Scrolling the Flow Diagram pane right and left – When scrolling the Flow Diagram pane of the Flow Execution page to the left after scrolling it to the right, some blocks may not be rendered properly or may be loaded after a delay.
Reporting
Time zones ahead of UTC time – When the system time zone is a time zone ahead of UTC time, some tooltips on the Reporting pages show dates that are one day ahead of the data's actual dates.
[Working as designed] "Pages Processed" count and Full Page Transcription Blocks – When processing submissions in custom flows with Full Page Transcription Blocks, the Pages Processed metric only counts pages that have been matched to layouts.
Labels on the "Performance Distribution" chart – The values on the Y axis of the Performance Distribution chart may overlap with the axis's label.
Refresh button on the Accuracy page – Clicking the Refresh button on the Accuracy page (Reporting > Accuracy) does not refresh all data.
Workaround: Refresh the browser page instead.
Dashboard
[Fixed in 33.1.0] Active Workers card and OpenID Connect user's names – When an OpenID Connect authentication provider is used, and the system cannot retrieve the first and last names of the users, the Active Workers card may be blank.
Keyer Data Management
"Change Label Location" keyboard shortcut – We're working to resolve an issue that prevents the "Change Label Location" keyboard shortcut (F4) from functioning.
“View Annotations” for a table column without any annotations in a document – Clicking the View Annotations link for a particular table column and then opening a document with annotations from the documents list causes the table column to be preselected in the Review Table Column Annotations section. If the selected table column does not have any annotations in the opened document, the table column is still preselected.
Review Table Column Annotations page’s header text in IE – The Review Table Column Annotations page’s header text is not centered in IE.
Upgrades
[Fixed in 33.1.0] "Document Processing Notifications (V33)" not automatically deployed – When upgrading from v32 to v33, the "Document Processing Notifications (V33)" flow is not automatically deployed.
Workaround: Ask your Hyperscience representative to make this flow visible in the application. Then, deploy the flow.
Security
Parsing of tag contents in HTML submissions – We're working to resolve an issue that allows GET requests within HTML
tags to be executed when HTML files are processed as submissions. Note that the responses to these requests are not retrieved by the system.