language

[2260383858]

Can Chinese be supported?

[lairgiyassir]

Hello,

To the best of my knowledge, I think yes! it is possible :)

Here is the output

kg = itext2kg_openai.build_graph(sections=['iText2KG 是一个 Python 包，旨在利用大型语言模型（LLM）从文本文档中提取实体和关系，逐步构建一致的知识图谱。'], ent_threshold=0.7, rel_threshold=0.7)
----
[INFO] ------- Extracting Entities from the Document 1
{'entities': [{'name': 'iText2KG', 'label': 'Python Package'}, {'name': 'Large Language Model', 'label': 'Technology'}, {'name': 'Text Documents', 'label': 'Data Source'}, {'name': 'Entities', 'label': 'Concept'}, {'name': 'Relationships', 'label': 'Concept'}, {'name': 'Knowledge Graph', 'label': 'Data Structure'}]}
[INFO] ------- Extracting Relations from the Document 1
{'relationships': [{'startNode': {'name': 'itext2kg', 'label': 'Python_Package'}, 'endNode': {'name': 'text documents', 'label': 'Data_Source'}, 'name': 'extracts from'}, {'startNode': {'name': 'itext2kg', 'label': 'Python_Package'}, 'endNode': {'name': 'entities', 'label': 'Concept'}, 'name': 'extracts'}, {'startNode': {'name': 'itext2kg', 'label': 'Python_Package'}, 'endNode': {'name': 'relationships', 'label': 'Concept'}, 'name': 'extracts'}, {'startNode': {'name': 'itext2kg', 'label': 'Python_Package'}, 'endNode': {'name': 'knowledge graph', 'label': 'Data_Structure'}, 'name': 'builds'}, {'startNode': {'name': 'itext2kg', 'label': 'Python_Package'}, 'endNode': {'name': 'large language model', 'label': 'Technology'}, 'name': 'utilizes'}]}
[INFO] Verification of invented entities

If you want the output in Chinese, you need to prompt the LLM in Chinese (fork this repo ;))

[2260383858]

Thanks