scribe-data 3.3.0

Wikidata and Wikipedia language data extraction

Wikidata and Wikipedia language data extraction

Scribe-Data contains the scripts for extracting and formatting language data from Wikidata and Wikipedia. Updates to the data can be done using scribe_data/wikidata/update_data.py and the notebooks within the scribe_data/load directory.

[!NOTE]
The contributing section has information for those interested, with the articles and presentations in featured by also being good resources for learning more about Scribe.

Scribe applications are available on iOS, Android (WIP) and Desktop (planned).

Check out Scribe's architecture diagrams for an overview of the organization including our applications, services and processes. It depicts the projects that Scribe is developing as well as the relationships between them and the external systems with which they interact. Also check out the Wikidata and Scribe Guide for an overview of Wikidata and querying language data from it.

Contents

• Process
• Contributing
• Environment Setup
• Supported Languages
• Featured By

Process ⇧

scribe_data/wikidata/update_data.py and the notebooks within the various scribe_data directories are used to update all data for Scribe-iOS, with this functionality later being expanded to update Scribe-Android and Scribe-Desktop when they're active.

The main data update process in update_data.py triggers SPARQL queries to query language data from Wikidata using SPARQLWrapper as a URI. The autosuggestion process derives popular words from Wikipedia as well as those words that normally follow them for an effective baseline feature until natural language processing methods are employed. Functions to generate autosuggestions are ran in gen_autosuggestions.ipynb. Emojis are further sourced from Unicode CLDR, with this process being ran in gen_emoji_lexicon.ipynb.

Running update_data.py is done via the following CLI command:

python3 src/scribe_data/wikidata/update_data.py

The ultimate goal is that this repository will house language packs that are periodically updated with new Wikidata lexicographical data and data from other sources. These packs would then be available to download by users of Scribe applications.

Contributing ⇧

Scribe uses Matrix for communications. You're more than welcome to join us in our public chat rooms to share ideas, ask questions or just say hi :)

Please see the contribution guidelines and Wikidata and Scribe Guide if you are interested in contributing to Scribe-Data. Work that is in progress or could be implemented is tracked in the issues and projects.

[!NOTE]
Just because an issue is assigned on GitHub doesn't mean that the team isn't interested in your contribution! Feel free to write in the issues and we can potentially reassign it to you.

Those interested can further check the -next release- and -priority- labels in the issues for those that are most important, as well as those marked good first issue that are tailored for first time contributors.

After your first few pull requests organization members would be happy to discuss granting you further rights as a contributor, with a maintainer role then being possible after continued interest in the project. Scribe seeks to be an inclusive and supportive organization. We'd love to have you on the team!

Ways to Help ⇧

• Reporting bugs as they're found 🐞
• Working on new features ✨
• Documentation for onboarding and project cohesion 📝
• Adding language data to Scribe-Data via Wikidata! 🗃️

Road Map ⇧

The Scribe road map can be followed in the organization's project board where we list the most important issues along with their priority, status and an indication of which sub projects they're included in (if applicable).

[!NOTE]
Consider joining our bi-weekly developer syncs!

Data Edits ⇧

[!NOTE]
Please see the Wikidata and Scribe Guide for an overview of Wikidata and how Scribe uses it.

Scribe does not accept direct edits to the grammar JSON files as they are sourced from Wikidata. Edits can be discussed and the queries themselves will be changed and ran before an update. If there is a problem with one of the files, then the fix should be made on Wikidata and not on Scribe. Feel free to let us know that edits have been made by opening a data issue and we'll be happy to integrate them!

Environment Setup ⇧

[!IMPORTANT]

Suggested IDE extensions

VS Code

• blokhinnv.wikidataqidlabels
• charliermarsh.ruff
• streetsidesoftware.code-spell-checker

The development environment for Scribe-Data can be installed via the following steps:

1. Fork the Scribe-Data repo, clone your fork, and configure the remotes:

[!NOTE]

Consider using SSH

Alternatively to using HTTPS as in the instructions below, consider SSH to interact with GitHub from the terminal. SSH allows you to connect without a user-pass authentication flow.

To run git commands with SSH, remember then to substitute the HTTPS URL, https://github.com/..., with the SSH one, [e-mail]:....

• e.g. Cloning now becomes git clone [e-mail]:<your-username>/Scribe-Data.git

GitHub also has their documentation on how to Generate a new SSH key 🔑

# Clone your fork of the repo into the current directory.
git clone https://github.com/<your-username>/Scribe-Data.git
# Navigate to the newly cloned directory.
cd Scribe-Data
# Assign the original repo to a remote called "upstream".
git remote add upstream https://github.com/scribe-org/Scribe-Data.git

• Now, if you run git remote -v you should see two remote repositories named:
  • origin (forked repository)
  • upstream (Scribe-Data repository)

2. Use Python venv to create the local development environment within your Scribe-Data directory:

• On Unix or MacOS, run:

  python3 -m venv venv  # make an environment named venv
  source venv/bin/activate # activate the environment
  

• On Windows (using Command Prompt), run:

  python -m venv venv
  venv\Scripts\activate.bat
  

After activating the virtual environment, install the required dependencies and set up pre-commit by running:

pip install --upgrade pip  # make sure that pip is at the latest version
pip install -r requirements.txt  # install dependencies
pip install -e .  # install the local version of Scribe-Data
pre-commit install  # install pre-commit hooks
# pre-commit run --all-files  # lint and fix common problems in the codebase

[!NOTE] Feel free to contact the team in the Data room on Matrix if you're having problems getting your environment setup!

Supported Languages ⇧

Scribe's goal is functional, feature-rich keyboards and interfaces for all languages. Check the language_data_extraction directory for queries for currently supported languages and those that have substantial data on Wikidata.

The following table shows the supported languages and the amount of data available for each on Wikidata and via Unicode CLDR for emojis:

Languages	Nouns	Verbs	Translations*	Prepositions†	Emoji Keywords
French	18,044	6,574	67,652	-	2,488
German	194,687	3,634	67,652	215	2,898
Italian	59,191	7,649	67,652	-	2,457
Portuguese	5,268	538	67,652	-	2,327
Russian	194,567	15	67,652	15	3,827
Spanish	61,650	7,912	67,652	-	3,134
Swedish	47,007	4,678	67,652	-	2,913

* Given the current beta status where words are machine translated.

† Only for languages for which preposition annotation is needed.

Featured By ⇧

Articles and Presentations on Scribe

2024

• April: Blog post on Medium about Scribe-Data and its functionalities
• February: Presentation slides for Scribe's participation at the Wikimedia Tech Safari Program

2023

• August: Scribe-iOS final submission report for Google Summer of Code 2023
• June: Scribe-iOS development blog post on Nested UITableViews & Apple's built-in ViewControllers in app menu for GSoC '23
• March: Presentation slides for a talk at Berlin Hack and Tell (Hack of the month winner 🏆)

2022

• August: Presentation slides for a session at the 2022 Wikimania Hackathon
• July: Presentation slides for a talk at CocoaHeads Berlin
• July: Video on Scribe for Wikimedia Celtic Knot 2022
• June: Presentation slides for a talk with the LD4 Wikidata Affinity Group
• June: Scribe featured for new developers on MediaWiki
• May: Presentation slides for Wikimedia Hackathon 2022
• March: Blog post on Scribe-iOS for Wikimedia Tech News (DE / Tweet)
• March: Presentation slides for Wikidata Data Reuse Days 2022

                             

Powered By ⇧

Contributors

Many thanks to all the Scribe-Data contributors! 🚀

Blog posts

List of referenced posts

• Building a Recommendation System Using Neural Network Embeddings by WillKoehrsen

• Wikipedia Data Science: Working with the World’s Largest Encyclopedia by WillKoehrsen

Wikimedia Communities