knox revisou este gist . Ir para a revisão
1 file changed, 21 insertions
Chroma-2.py(arquivo criado)
| @@ -0,0 +1,21 @@ | |||
| 1 | + | import chromadb | |
| 2 | + | # setup Chroma in-memory, for easy prototyping. Can add persistence easily! | |
| 3 | + | client = chromadb.Client() | |
| 4 | + | ||
| 5 | + | # Create collection. get_collection, get_or_create_collection, delete_collection also available! | |
| 6 | + | collection = client.create_collection("all-my-documents") | |
| 7 | + | ||
| 8 | + | # Add docs to the collection. Can also update and delete. Row-based API coming soon! | |
| 9 | + | collection.add( | |
| 10 | + | documents=["This is document1", "This is document2"], # we handle tokenization, embedding, and indexing automatically. You can skip that and add your own embeddings as well | |
| 11 | + | metadatas=[{"source": "notion"}, {"source": "google-docs"}], # filter on these! | |
| 12 | + | ids=["doc1", "doc2"], # unique for each doc | |
| 13 | + | ) | |
| 14 | + | ||
| 15 | + | # Query/search 2 most similar results. You can also .get by id | |
| 16 | + | results = collection.query( | |
| 17 | + | query_texts=["This is a query document"], | |
| 18 | + | n_results=2, | |
| 19 | + | # where={"metadata_field": "is_equal_to_this"}, # optional filter | |
| 20 | + | # where_document={"$contains":"search_string"} # optional filter | |
| 21 | + | ) | |
Próximo
Anterior