Uwazi v1.0 released

New features

Graph visualisation for document/entity relationships

Now you can better visualise how your data elements are related to each other. Our proposal is a middle ground compromise between a graph visualisation and relational metadata. It is our first release of a series of iterations until we have a fully developed relationships feature for Uwazi.

We have turned this new view into the default visualisation for relationships, while also keeping the previous one available.

Improvements in full text search

We added relevance to sorting options and set this option as default for text searches. We also increased the amount of text in the results snippets so users have better context in which the search text is found.

Users can click “Show more” in the card to see the complete list of snippets found in a document in the right sidebar. Clicking on the page number will take users to the document view and scroll them to that particular page. The number of matching results is displayed in the top of the tab, so that users have a hint on when it’s worth taking a deeper look.

“Exact search match” has been boosted in the results, providing better results for users. We have fixed several small issues affecting full text search.

Other improvements

Improvements in performance

We focused in several aspects affecting the overall performance of the application. Now users experience a smoother interaction with the front-end and a quicker response from the server.

Blank states

Several screens have been redesigned to offer users proper blank state information. This way, first time users have directions on what to do next and links to the supporting documentation. Regular users get a better understanding when no information is available for a particular screen or tab (i.e. when there is no table of contents, they now see a message with an explanation instead of a blank placeholder).

We changed “dictionary” to “thesauri”

Uwazi users can create their own controlled vocabularies that are used to organise their documents and entities. We have decided to change the name of these controlled vocabularies from “dictionary” to “thesaurus” to be more consistent with HURIDOCS’ other resources, such as our human rights Micro-thesauri (https://www.huridocs.org/resource/micro-thesauri/).

Manually set the document language

Uwazi automatically detects the document language and uses it to offer better search results to users. While this is a desired behaviour, it can create some confusion when documents have mixed languages (written in several languages). We addressed this issue by allowing users to manually set the language for the document allowing them to override the automatic detection.

Improved the process for users to create new accounts

We have reworked part of the process of setting up new accounts (such as updating the email messages sent to new users to make them more informative), so users have a better experience when configuring their access.

Add support for Vimeo in rich text markdown fields

Rich text fields now support embedding of Vimeo videos by simply adding this code snippet with a Vimeo URL: {vimeo}(https://vimeo.com/37942739)

Improved rendering of properties in cards

Some minor tweaks to how properties are rendered in cards have been made improving the overall experience. Notably, property labels are now always displayed.

Bug Fixes

Now the Table of Contents entries are properly spread among languages when using the fallback content.
Translations now properly show all labels for the translatable elements. Previously some labels were blank for certain situations.
Fixed several bugs affecting date fields.
Fixed several bugs affecting how text is indexed and searched.

What does this mean for you?

• For those of you who have an instance hosted with us, we have already deployed this upgrade to your Uwazi instance, so there’s no need to do anything.
• If you are hosting Uwazi yourself, you will want to upgrade by downloading the latest release and run the following database migration using terminal on the root directory of Uwazi:

    $ export DATABASE_NAME='name'

by default the name is uwazi_development.
If the mongodb server is on a remote machine, then configure it by:

    $ export DBHOST='theHost'

then run:

    $ node run_migration.js ./app/api/migrations/assign_file_languages.js
    $ node run_migration.js ./app/api/migrations/fix_entity_based_references.js

• See all of our updates in our GitHub changelog
• We’re working on our user guide on GitHub and you’re encouraged to contribute.
• Contact us if you have any questions!

Uwazi v0.5.0 released

New features

API documentation

We’ve started to document API commands for Uwazi. We're currently working on documentation for the search, entities, and attachments endpoints. You can access the API for your Uwazi instance by adding /api to the end of your Uwazi URL, like this: https://yourinstancename.uwazi.io/api. Please note: this is a work-in-progress!

Google Analytics integration

We’ve integrated Google Analytics into Uwazi so that administrators can see how many visitors are accessing their public-facing Uwazi content. More information >>

Other improvements

Choose between AND/OR operators in filters

We’ve added the ability for users to choose between using AND or OR operators in the Uwazi filters. More information >>

Easier to find the target documents for your connections

We’ve improved the connections search functionality by limiting the search to document titles instead of the entire document content. This will make it much faster to find the target document for your connections.

Easier access to Downloads

We’ve moved the Downloads section (previously the Attachments tab) to the main Information tab in the side panel, making them easier to access and edit. More information >>

Automatically attach files in all instance languages

We've added functionality to attach a file to all language versions of an entity / document with the click of a button. This will speed up the process of adding downloads that are shared between languages like images or sound files, for example. More information >>

Automatic document language detection

If your Uwazi instance is in multiple languages, it is important for users to know the language of documents as they are browsing the collection. We've added the ability for Uwazi to automatically detect the language of your document, and users will see the language of the document if it does not match the selected user interface language.

Bug fixes

• We fixed a bug that corrupted metadata-based relationships when a select-type property's name was changed
• We fixed how aggregations are generated for unpublished documents in the Uploads section
• We fixed a bug that broke the “load more” button for connections
• We changed some elasticsearch configurations that were causing incorrect sorting results in connections
• We fixed a bug that caused multi-select filters to disappear when writing on a numeric filter
• We fixed a bug that broke the filters when displaying rich-text properties on cards
• We fixed the index mapping so that Elasticsearch's field amount does not go beyond the usual default settings

What does this mean for you?

For those of you who have an instance hosted with us, we have already deployed this upgrade to your Uwazi instance, so there’s no need to do anything.

If you are hosting Uwazi yourself, you will want to upgrade by downloading the latest release and run the following database migration using terminal on the root directory of Uwazi:

    $ export DATABASE_NAME='name'

by default the name is uwazi_development.
If the mongodb server is on a remote machine, then configure it by:

    $ export DBHOST='theHost'

then run:

    $ node ./run_migration.js ./app/api/migrations/assign_file_languages.js
    $ node ./run_migration.js ./app/api/migrations/fix_entity_based_references.js

Additional information

• See all of our updates in our GitHub changelog
• We’re working on our user guide on GitHub and you’re invited to contribute.
• Contact us if you have any questions!

Uwazi v0.4.0 released

Uwazi v0.4.0 released

New features

Full text search improvements

We have made significant improvements to our full text search feature using Elasticsearch. These improvements allow users to:

• see an example snippet with highlighted text from the search results,
• see a list of all the places this text is found in the document, and
• quickly jump to any of these places in the document itself.

New date property types and improvements

There are many situations in which a single date on your document or entity does not capture enough information for your analysis. You may be working with event entities that have start and end dates (or multiple start and end dates). You may have documents in which you need to capture the date for each time something happened.

For these kinds of use-cases, we have added three new date property types: date range, multi-date, multi-date range.

Additional improvements:

• Users can use the DD/MM/YYYY date input format
• We’ve added date description information to the cards in the library so visitors know what each date is referring to:

What does this mean for you?

• For those of you who have an instance hosted with us, we have already deployed this upgrade to your Uwazi instance, so there’s no need to do anything.
• If you are hosting Uwazi yourself, you will want to upgrade by downloading the latest release and run the following database migration using terminal on the root directory of Uwazi:

    $ export DATABASE_NAME='name'

by default the name is uwazi_development.
If the mongodb server is on a remote machine, then configure it by:

    $ export DBHOST='theHost'

then run:

    $ node run_migration.js

• See all of our updates in our GitHub changelog
• We’re working on our user guide on GitHub and you’re encouraged to contribute.
• Contact us if you have any questions!

Uwazi v0.3.5 released

Improvements and upgrades

Better looking URLs for search filter results

One great feature in Uwazi is the ability to save and share search filter results using a unique URL. With this latest release, we have improved the usability of this feature by making the URLs shorter and more readable. Here is an example of a search filter results URL before this improvement, and after:
Before:
https://hr-reports.uwazi.io/en/library/?aggregations=%5B%7B%22name%22%3A%22countries%22%2C%22nested%22%3Afalse%7D%5D&types=%5B%2258d06c67f53d6d50ab7495ea%22%5D&filters=%7B%7D&order=desc&sort=creationDate
After:
https://hr-reports.uwazi.io/en/library/?q=(order:desc,sort:creationDate,types:!(%2758d06c67f53d6d50ab7495ea%27))

Improvements to priority sorting

Previously the default sorting was always "Date (Recently added)". Priority sorting had a minor impact and only started influencing search results once the user had selected a custom property for sorting. Conversely, that selection was short-lived if a filtered document type was missing that property.

The changes now make priority sorting govern the default sorting the moment you access the library and uploads sections. It also changes frequently as you change your filtered document types. Furthermore, user selection is now remembered and, whenever possible due to filters selected, it uses the user-selected option.

We’ve upgraded to elasticsearch 5.4

This upgrade improves Uwazi's full text search capabilities, allowing us to build new features that will help users find specific text within their document collection. Stay tuned for the release of these new full text search features!

What does this mean for you?

If you're hosting with us:

• We have already deployed this upgrade to your Uwazi instance.
• In your Uwazi instance, you will need to update any URLs based on search filter results, such as:
  • any menus that include URLs to search filters, or
  • any embedded lists that include URLs to search filters (e.g. your homepage). See our user guide for more information on how to use custom syntax to embed lists in a page.

If you are hosting Uwazi yourself:

• You will want to upgrade by downloading the latest release of Uwazi, and

• Install elasticsearch 5.4 on the system.

  If your setup is using the default ports and database locations and names, all you need to do is reindex the elastic search index. If you have customized ports and database, configure env variables by exporting the following values:

    $ export ELASTICSEARCH_URL=elasticsearch_url (default is localhost:9200)
    $ export INDEX_NAME=elastic_index_name (default is uwazi_development)
    $ export DBHOST=mongodb_url (default is mongodb://localhost)
    $ export DATABASE_NAME=db_name (default is uwazi_development)

Then, execute the reindex script from uwazi root directory

    $ cd database/
    $ node reindex_elastic.js

More information

• See all of our updates in our GitHub changelog
• We’re working on our user guide on GitHub and you’re invited to contribute!
• Contact us if you have any questions!

Uwazi v0.3.0 released

Documentation Updates

Uwazi user guide is now available!

We’re very happy to announce that our first draft of the Uwazi user guide is now available in English on our Github wiki! If you find any mistakes, or parts that can be improved, please suggest edits to the wiki using your Github account. This guide will be a work in progress until we have a stable release. We’re currently working on translating this guide into Spanish.

New Features

Improvements to the uploads section

These improvements make the uploads section more similar to the library, including:

• Properties can be displayed in document and entity cards, just as they are in the library.
• Document and entity cards can be sorted, filtered, and searched, just as they are in the library.
• The right side panel behaves the same as the library: the user can view the table of contents, properties, connections, and attachments without opening the document.
• The Add New Entity button has moved from the bottom right corner of the uploads section, to the top right corner of the main window in this section.

Create multiple users and assign roles

Administrators can now create user accounts and assign one of two roles: administrator and editor. This feature is especially helpful if you have a team of users who are adding and organising documents and entities in your collection. More information on the permissions assigned to each role >>

What does this mean for you?

• For those of you who have an instance hosted with us, we have already deployed this upgrade to your Uwazi instance, so there’s no need to do anything.

• If you are hosting Uwazi yourself, you will want to upgrade by downloading the latest release and run the following database migration using terminal on the root directory of Uwazi:

  export DATABASE_NAME='name'

  by default the name is uwazi_develaopment.

  If the mongodb server is on a remote machine, then configure it by:

  export DBHOST='theHost'

  then run:

  node run_migration.js

• See all of our updates in our GitHub changelog

• We’re working on our user guide on GitHub and you’re invited to contribute!

• Contact us if you have any questions!

v0.2.0

New Features

Edit multiple items at once

Edit multiple items at once by selecting documents and/or entities in the library (on Windows or Linux, hold the CTRL key then click the items; on a Mac, hold the Command key then click the items), clicking edit, then applying the property field values you want to assign to all selected items. This also means you can delete multiple items at once.

Embed YouTube videos

Embed YouTube videos by adding a rich text property field to your document or entity template and then using the format: {youtube}(youtube video link). You could use the transcript of the video as the main document and attach the video for reference. This way, you can utilise our Uwazi features to connect the text to other documents, entities or paragraphs.

Improved display of entity connections

View your entity connections in the main window by opening the entity and clicking on the connections icon. This is an improvement on the previous connections display in the right sidebar. This new display is especially helpful if you have an entity with many connections, like country.

What does this mean for you?

• For those of you who have an instance hosted with us, we have already deployed this upgrade to your Uwazi instance, so there’s no need to do anything.
• If you are hosting Uwazi yourself, you will want to upgrade by downloading the latest release.
• See all of our updates in our GitHub changelog
• We’re working on our user guide on GitHub and you’re invited to contribute!
• Contact us if you have any questions!

v0.1.7

New Features

• Adopted PDF.js as document renderer. Rendering PDFs in a web environment is a great challenge. Fortunately, Mozilla has addressed this topic with their awesome library PDF.js. It is way faster and less prone to errors than our previous solution. This greatly alleviates the slowness of document reading and text selection.

• Multi-language support. Interface, metadata and content can now co-exist in several languages at a time. Currently this option lacks a front-end configuration option and requires manual setup from a developer. Please get in contact if you are interested in testing it out.

  

• Attachments tab. Now documentes and entities have a dedicated tab for files. This comes handy when we want to have several formats for the same document (Word, PDF, etc). Also makes the available documents for downloading more visible to users.

  

• Custom configuration for first order of filters. The first filters a user encounters is the list of the available documents and entities. Now this list can be configured from settings > filters, selecting which options to hide/display and allowing for the creation of groups of options. This improves the user experience for visitors, offering a rational first order navigation in the library.

  

• Now documents and entities can have an icon or flag assigned. This is very useful for countries, but also to give some records a special visual hint.

  

• Improved connections panel. Now connections are grouped by connection type which allows for expanding and collapsing groups of connections in the side panel. Also the connected cards display its metadata as in the rest of the library and can be sorted.

  

Improvements

• Title and document type in metadata side panel are now displayed with styles similar to those of the cards, this way users can quickly identify what kind of document they are looking at.
• Now entities and documents edition have a visible cancel button that will get the user out of edition mode. Previously there was only an "X" at the corner for closing which was not clear enough.
• When using an entity as a dictionary for a multi-select field, now its treated as a connection so it will appear in entities and documents connections panel.
• Sort by options have a better design now.
• Metadata titles are only shown on mouse over. This way we make better use of the space in the cards avoiding redundant labels in every card. If the users needs to know what that particular data means, a mouse over will display the label.
• Fields with blank value are no longer rendered in cards, side panel and entities. This way we don’t see anymore an optional field label without the expected value, which is misleading for users.
• Improved design of metadata forms. Better usage of the space in the screen and clarification of some functionality.
• Better styles for uploads section. While this section still needs a full revamp of its functionality, we’ve expended some time trying to make it more appealing. Particularly, this section now shows the document type when assigned, making it easier to locate, edit and publish pending documents.
• Better styles for pages and rich text fields.

Bug fixes

• Now entities do not show a "download" option in their actions.
• Default A-Z sorting in library is ascending (from A to Z) instead of descending (from Z to A).
• Now multiple option properties based in dictionaries are correctly filtered.
• Fixed a bug in text reference deletion that would prevent a reference from being deleted when the referenced text was too long.
• Now select options appear in alphabetical order. This was particularly annoying when working with longs lists in selects.
• Improved performance in metadata edition. A flaw in the implementation was causing the interface to consume too many resources rendering the metadata edition very slow.
• Improved performance of template configuration. Previously some server side processes were affecting the speed of the UI.
• Fixed a bug that would allow users to select text for tagging and cross-referencing when a subscript was present in the page. This was related to how the PDF to HTMl conversion worked.
• Fixed a bug that prevented uploaded documents from being saved when a required multi-select was present in the template.

v0.1.6

New features

• Custom navigation menus and pages.
  • Now administrators can create custom pages in settings > pages section. These pages support html like features using the Markdown standard.
  • In order to make these pages available to users, the navigation menu can be configured in settings > menu. This supports both internal and external links.

• Filters as URLs. When users run a search or filter content form the library, these params will also be expressed in the form of a URL in the navigation bar. This comes very handy when it comes to bookmark or share a particular search/filter combination. Moreover, this makes a great combo when used along with the navigation menu to point visitors of the site to particular subsets of information in the library.

• Display metadata in library cards. Now administrators can configure which metadata properties will be shown in library cards for both entities and documents. This is configured in each metadata property in settings > documents and settings > entities.

  

• Facet browsing. Filters now support faceted browsing style. This is automatically generated for "select" and "multi-select" metadata types. The behaviour is similar to most faceted browsing implementations:

  • Only options that will toss results are displayed (options with zero results are eliminated).

  • The amount of results under the option is displayed besides the option.

  • When an option is selected all amounts are recalculated, showing only the options that will match both criteria (the already selected option plus another option).

  • If the option list is longer than 5 items, its collapsed by default, showing a "Show more" button that will display all available options.

  • It has a handy search box to quickly locate options.

  • When an option is selected, it automatically launches the filtering request and updates the result list.

  • Options are sorted alphabetically, but when one is picked up and the list is collapsed, it will be moved to the top of the list so user can clearly see which options are active.

    

• Rich text metadata field. Based on the same Markdown implementation we are using for pages, now user can configure rich text as metadata properties for their entities and documents.

Improvements

• Several style and usability improvements have been performed in the search/filter side panel. Notably, moving the "reset" button from the top to a bigger and more visible button in the bottom of the side panel.

  

• Files converted from PDF to HTML are now stored in disk instead of in the DB. This has improved the performance of the database and fixed some glitches in the application.

Bug fixes

• Fixed a bug that tossed JS errors when reloading the application on any page other than library and then navigating to library.
• Fixed a bug that was making the application crash when attempting to display an entity that had a multi-select metadata field configured and values.

v0.1.5

New features

• Table of contents assistant. Now users can create table of contents for better navigation of their documents. You can find this feature at the document view while being logged in.

  To add an item to the TOC, select any text snippet, mouse over the floating "+" button and click on the "Add to table of contents" option. The assistant will automatically order the anchors in the TOC based on their order of appearance in the document.

  The assistant offers options to edit the snippet in case you want the TOC to display a text different to the selection, indent the elements of the TOC and delete them.

• Now we support the creation of entities. They may be used ie, in the context of a case law document collection, to manage persons (judges, victims, perpetrators) institutions (courts, ministries, governments), countries, companies, etc.

  • They are very similar to documents and support the same options for metadata, filtering and searching.
  • They also support connections for cross referencing entities between them or with documents.
  • There is a new section in settings > entities where the entity types and metadata can be defined.
  • (temporarily) They can be created in the uploads section, clicking on the "+" floating button.
  • They have a different view design. Since they don't display a document text, we created a entity view that makes better use of the space.
  • They have a different icon besides the entity type tag in the library. This icon will be customizable in the future.

Improvements

• The side panel usability has been greatly improved:
  • Now many of the options that previously were accessed through the floating action button are quickly available at the new "side panel footer". Check it out =)
  • In the document viewer the side panel now has three tabs: table of contents, metadata and connections. Being able to quickly switch between that information while reading a document.
• Search and filtering have been improved:
  • Now the search bar is integrated in the side panel along with the filters.
  • The side panel for search and filtering is no longer opened by clicking on the floating button, now there is a special icon on the header to trigger this functionality.
• Now we have a general error feedback message that will let users know if there has been a server, connection or client error that is not being handled. This feature will need further improvement to better inform the user what the cause of the error may be, but for now is better than having an irresponsive application with no clue of what is going on.
• Now when creating a "target paragraph" connection, the "save" option will stick to the user scroll so there is no need to scroll back to the top to save the connection.
• Now we also display inbound connections (as in another document is pointing to the present document) in documents and entities, previously we were only displaying outbound connections.
• Better reference creation UX:
  • Now you don't have to close the connections side panel to create a new connection.
  • Now you don't have to reselect text to create a connection if you switch between panels.
  • Now you can just click on an existing connection, both in the highlighted text or the side panel, when creating a "target paragraph" connection, sparing you the burden of having to reselect the same text.
• Metadata section redesign. This section still had the styles of our previous UI, now it has been updated for a better look and feel.

Bug fixes

• Fixed a rendering bug that would break the header height with long title names.
• Fixed a bug that left all documents in "target document state" after abandoning a process of creating a connection to a paragraph.
• Fixed several rendering problems on mobile.
• Fixed a bug that prevented users from saving thesauri under certain circumstances.
• Fixed a rendering issue when there are overlapping connections in the text. Now they are properly highlighted.

v0.1.4

New features

• Improved connection to a paragraph experience. Now the target document is loaded as a fully functional document, allowing users to display its metadata and connections. Previously these panels were not available for target documents.
• Existing connections in the target document can now be easily reused as target paragraph by selecting them in the document body or the side panel, no need to manually highlight the target text anymore.
• Now documents will display also inbound connections (from other documents towards text snippets in the current document), previously it was showing only the outbound ones (from this document towards other documents or snippets in other documents).
• Added "date" metadata property. Now documents can have dates as properties and can be filtered by date ranges with a handy date picker.

Small improvements

• Metadata configuration has been moved to "Settings", freeing up space in the navigation bar and concentrating all settings into a single section.
• Improved metadata configuration section design to match that of the rest of the application. This section still had legacy styles.
• Changed some UI colours including background and other styles improving the overall design and readability.
• When downloading the original file, it will be named after the document's title property, resulting in better file names. Previously it was downloaded with the internal database unique id which was confusing.

Bug fixes

• Fixed a rendering bug in the sidebar height in mobile devices.
• Now collection name is properly saved.
• Filtering by thesauri is working again.
• Removed a misplaced link in the titles of the connection cards that was leading to accidental navigation to other documents.