py-entity-linking-el 0.1.7

A lightweight library for entity linking in Greek.

Lightweight entity linking solution for Greek language.

Please consider citing our works if you use code from this repository. Also, we recommend using a Colab T4 GPU for faster results.

Main dependencies

• python>=3.10
• numpy==1.26.4
• SPARQLWrapper==2.0.0
• sentence_transformers==3.1.1
• aiohttp==3.9.5
• openai==1.55.3
• httpx==0.28.1
• beautifulsoup4==4.12.2
• nest_asyncio==1.5.8

Example & Usage

from linking import main

# Your API token which can be found here (https://github.com/marketplace/models/azure-openai/gpt-4o)
api_token = "YOUR_API_TOKEN"

main.EL(api_token=api_token,
	sentence="Χάρη στην ΕΦΓ, μια μηχανή μπορεί να 'καταλάβει' το περιεχόμενο των εγγράφων, συμπεριλαμβανομένων των αποχρώσεων του πλαισίου των της γλώσσας σε αυτά.",
	mention="ΕΦΓ",
	single="No",
	combination="No",
	embedding_model="intfloat/multilingual-e5-large-instruct")

The correct entity for 'ΕΦΓ' is:

Wikipedia: https://el.wikipedia.org/wiki/Επεξεργασία_φυσικής_γλώσσας

Wikidata: https://www.wikidata.org/wiki/Q30642

DBpedia: http://dbpedia.org/resource/Natural_language_processing


Execution Time: 00:00:29

Parameters

• api_token: Your API token from here. (Required)
• sentence: A Greek text. (Required)
• mention: The mention you want to perform the linking, the mention should be from inside the provided sentence. (Required)
• single: Usually used for difficult mentions, it searches each word of the mention individually, (deafult="No"), (Values: "Yes", "No"). (Optional)
• combination: Usually used for difficult mentions, it makes combinations for each word of the mention, (deafult="No"), (Values: "Yes", "No"). (Optional)
• embedding_model: A sentence-transformers model to perform text similarity, (deafault="intfloat/multilingual-e5-large-instruct"), (Values: str of the name of any sentence-transformers model that supports Greek). (Optional)

Licence

This library is licensed under the CC-BY-NC 4.0 license.