jo-fu / timelinecurator

A visual timeline authoring tool that extracts temporal information from freeform text

TimeLineCurator

A visual timeline authoring tool that extracts temporal information from freeform text

TimeLineCurator was created as part of a Master's thesis at the University of British Columbia, Canada and University of Munich (LMU), Germany.

The running application is hosted on the cloud platform Heroku.

More information and instructions can be found on the project's page

Here's a short video explaining TLC's purpose and functionality.

The project's folder structure:

• nltk_data: Natural Language Toolkit for Python, necessary for splitting freeform text into sentences, tokenizing words etc.
• static: contains all static front-end files, like JavaScript files, stylesheets, images and example data
• templates: the HTML templates to which the flask app refers
• ternip: The "Temporal Expression Recognition and Normalisation in Python" library used to find temporal information inside the freeform text (project on GitHub)
• Procfile: declares how to run app dynos on Heroku platform
• app.py: Flask app that contains the Python commands and connects the front-end with the back-end
• requirements.txt: tells Heroku (where the app is hosted) what dependencies we need

timelinecurator's Issues

TERNIP not catching "early", "middle", "late" decade modifiers

For example, should tag "late 1990s" or "early 1980s".

Right now, only the decade itself is recognized, so early 1990s is always represented as 1990-2000.

Export as Google Spreadsheet

Potential extension to export options: Create Google Spreadsheet as used for TimelineJS, that can then easily be saved into the user's Google Drive folder.
But it seems rather complicated to create a spreadsheet that is not initially connected to an account, see this Stackoverflow question.
The Google App Script documentation does not include such a case as far as I can see. Are there alternatives?

"1860s" are understood as "1960s"

Hi, kudos for the very interesting approach.
I am trying the website for an historical research on New York architects in the Eighteen Hundreds.
There is a bug when a decennial date is provided, eg. 1860s. TLC understands it as "1960-1969) instead of "1860-1869".

Multiple event selection

In the timeline view, I would like to drag- or lasso- select multiple events.

In the list view, the current shift-click multi-select behaviour is ctrl-click multi-select in most other applications. I would like to be able to shift-click select multiple consecutive events in the list view; the current shift-click multi-select should be retained as ctrl-click select.

Text Encoding

Great tool overall - thanks for the effort.

We're having some issues with non-ASCII text. If a URL has non-ASCII characters (eg something from Japanese Wikipedia - https://ja.wikipedia.org/wiki/%E3%83%8F%E3%83%AF%E3%82%A4%E5%B7%9E) it won't scrape correctly. If you paste in the text instead, TimeLineCurator still grabs dates correctly, but the text is mangled rather than accurately representing the characters - see below.

Self Hosted Installation

Is there a way I can install the whole application locally, for example using a docker container?
I have a lot of biographies to treat (approx. 2200) and it would be more efficient to do it locally.

permalinks for documents and UI state for easy sharing

just discovered this project from a @tamaramunzner talk. nice work!

I'd really like to be able to share a permalink for a document that I am looking at so that I can show my friend this document with the exact selection (highlighted node and text) that I am looking at.

an example TimeLineCurator state that I would like to share with a friend: