Forum (Archive)

This forum is kept largely for historic reasons and for our latest changes announcements. (It was focused around the older EPPI Reviewer version 4.)

There are many informative posts and answers to common questions, but you may find our videos and other resources more informative if you are an EPPI Reviewer WEB user.

Click here to search the forum. If you do have questions or require support, please email

<< Back to main Help page

HomeHomeEPPI-Reviewer 4...EPPI-Reviewer 4...Forum announcem...Forum announcem...Latest Changes (22/09/2022 - V and 16/10/2022 - Changes (22/09/2022 - V and 16/10/2022 -
New Post
22/09/2022 16:30

Version is a minor release comprising of bug fixes and one new feature. The important bug fixes apply to the "Update review" area, resolving problems with combining searches and importing/browsing search results. The new feature is a new report called "Duplicate report 2" added in EPPI-Reviewer Web, references tab, in the "Export to RIS" sub-menu.

[Update: 16/10/2022] Version includes a critical update and some associated bugfixes, all related to applying classifiers. See also notes about the crashes that affected EPPI-Reviewer Web on 06/10/2022 and on 13/10/2022. Details on this are at the bottom.]

EPPI-Reviewer Web, New feature: Duplicate report 2

This report is accessible via the sub-menu of the Export (to RIS) button, in the References tab. It lists selected items in rows and will report about the source of their duplicates, producing one column per relevant source. In the source columns, the presence of the letter "M" indicates which source the master item belongs to; the letter "X" indicates that one or more duplicates belong to the source in question.

Bug fix, both versions: importing and browsing search results.

Some rare OpenAlex references had an unexpected data structure, which made their conversion into EPPI-Reviewer native data structures fail. As a consequence, search results that contained such references could not be imported, and browsing (previewing) search results could also fail.
The data "oddity" in question is: some references have just one author, and the author has no value in the name field. In such cases, EPPI-Reviewer was failing to interpret the authors list correctly. This problem is now solved.

Bug fix, both versions: combining OpenAlex searches.

Version, being "substantial" and involving most of the underlying technologies, was always somewhat risky. In fact, it "broke" the "combine searches" functionality available in "Update review", "Search" tab. This problem is now solved.

Bug fix, EPPI-Reviewer Web: logging errors.

EPPI-Reviewer Web includes a facility to log all errors that happen in the server-side (which usually produce an error dialog on the client). In version, logging was (apparently) working as expected in the "Development/testing" environments, but was mostly failing to work in production. This problem is obviously not visible to end-users, but naturally, it limits our ability to detect and inspect real-world issues, so it needed to be resolved quickly.

Bug fix, EPPI-Reviewer Web: searching OpenAlex topics.

In "Update review", the "match records" tab includes a little known functionality which allows to search for OpenAlex topics (or "concepts"), and then browse the relative references directly. Version included a number of relatively small "look and feel" changes, which had the side-effect of making the list of "matching topics" hard to read. This problem is now solved. Additionally, the list of topics shown is now "paged" to aid exploring it, when it contains more than 50 elements.

EPPI-Reviewer Web: layout improvements.

In "import coding tools", the columns showing the coding tools structure were not adjusting their height in the best possible way. This page has been re-designed (slightly) to make it react better to different window sizes. We also have re-adjusted the visual structure of "confirmation" dialogs, which became a little awkward in version

Out of Schedule changes

The following changes were published on short notice after version, but before the present release.

EPPI-Reviewer Web: selected code in "Item details" page

Version included a new system to render the code trees; as a consequence, it "lost" the ability to provide a visual cue about what code was "active" (/selected) in the tree. This cue is important especially when coding text selections in PDFs. An out-of-schedule, ad-hoc update resolved this issue.

EPPI-Reviewer Web: uploading EPPI-Visualisation images

Version also introduced a new bug which made it impossible to upload brand images / logos in the setup area of EPPI-Visualisations, this problem was also resolved in between "official" releases.

Update: V.

Critical bugfix: apply classifier.

On Thursday 6 October, EPPI-Reviewer Web crashed and was unexpectedly offline for about 20 minutes. This was caused by a previously unknown bug, which can produce a series of crashes (bad enough to put the whole application offline) if/when the same person tries to apply classifiers from within the same review, in an extremely quick succession (we didn't think people could be that quick), and when this fails, repeats the same attempt a few more times. This problem is now fixed, via the addition of a number of measures.
In both versions, the server-side is now protected from this kind of problem, preventing the crashes at the root. In ER-Web, we have corrected the mechanisms in the user interface, which now awaits for "please apply classifiers" requests to report back, ensuring that the user interface will not allow to produce another "apply classifier" request too quickly.

EPPI-Reviewer Web: applying pre-trained COVID models.

While investigating the previous bug/crashes, we discovered that the two pre-trained COVID models were not working in EPPI-Reviewer Web. This problem is now solved and they are now working correctly in both versions.

EPPI-Reviewer Web: crash on 13/10/2022.

On 13 Oct. at around 5AM (UK time) EPPI-Reviewer Web crashed again, and was offline until around 8AM. Unlike the crash that happened on Oct. 6, this crash is still under investigation, and currently lacks a full explanation. Between 5 and 9:40AM, EPPI-Reviewer Web was unstable, and produced errors we never saw before or after. Traffic analysis did not yeild any sign of unusual activity, while the errors stopped appearing before we could identify what triggered them. From the data that was possible to collect, it seems that the problem was not connected to any specific user action, and might be caused by the component that extracts text from PDF documents. We will continue to monitor the situation closely, while we have put in place some measures designed to reduce the impact of such unexplained errors.

HomeHomeEPPI-Reviewer 4...EPPI-Reviewer 4...Forum announcem...Forum announcem...Latest Changes (22/09/2022 - V and 16/10/2022 - Changes (22/09/2022 - V and 16/10/2022 -

Copyright 2021 by EPPI-Centre :: Privacy Statement :: Terms Of Use :: Site Map :: Login
Home::Help::EPPI-Mapper::RIS Export::About::Account Manager