impl text splitter with tiktoken

This commit is contained in:
tsukasa
2023-09-26 14:43:56 +08:00
parent 8b862c0075
commit 411ab61880
3 changed files with 147 additions and 32 deletions
+1 -5
View File
@@ -50,11 +50,7 @@ class DocumentObjectBuilder:
return self
def build(self) -> DocumentObject:
chunk_list = KnowledgeStore().get_chunk_list_writer().create_chunk_list_from_text(
self.text,
1024 * 4,
".?!\n"
)
chunk_list = KnowledgeStore().get_chunk_list_writer().create_chunk_list_from_text(self.text)
doc = DocumentObject(self.meta, self.tags, chunk_list)
doc_id = doc.calculate_id()