Revert to EXTENSION_TO_DOCUMENT_CLASS implementation of classify_documents

This commit is contained in:
Leon Luithlen 2024-11-13 14:32:10 +01:00
parent aa1480ca2c
commit cd80525420

View file

@ -1,12 +1,16 @@
from cognee.modules.data.models import Data
from cognee.modules.data.processing.document_types import Document, PdfDocument, AudioDocument, ImageDocument, TextDocument
EXTENSION_TO_DOCUMENT_CLASS = {
"pdf": PdfDocument,
"audio": AudioDocument,
"image": ImageDocument,
"txt": TextDocument
}
def classify_documents(data_documents: list[Data]) -> list[Document]:
documents = [
PdfDocument(id = data_item.id, name=f"{data_item.name}.{data_item.extension}", raw_data_location=data_item.raw_data_location) if data_item.extension == "pdf" else
AudioDocument(id = data_item.id, name=f"{data_item.name}.{data_item.extension}", raw_data_location=data_item.raw_data_location) if data_item.extension == "audio" else
ImageDocument(id = data_item.id, name=f"{data_item.name}.{data_item.extension}", raw_data_location=data_item.raw_data_location) if data_item.extension == "image" else
TextDocument(id = data_item.id, name=f"{data_item.name}.{data_item.extension}", raw_data_location=data_item.raw_data_location)
EXTENSION_TO_DOCUMENT_CLASS[data_item.extension](id = data_item.id, title=f"{data_item.name}.{data_item.extension}", raw_data_location=data_item.raw_data_location, name=data_item.name)
for data_item in data_documents
]
return documents