[ENH] generate IDs during .add() if not provided

chromadb/api/models/AsyncCollection.py

@@ -7,6 +7,7 @@
 
 from chromadb.api.types import (
     URI,
+    AddResult,
     CollectionMetadata,
     Embedding,
     Include,
@@ -31,7 +32,7 @@
 class AsyncCollection(CollectionCommon["AsyncServerAPI"]):
     async def add(
         self,
-        ids: OneOrMany[ID],
+        ids: Optional[OneOrMany[ID]] = None,
         embeddings: Optional[
             Union[
                 OneOrMany[Embedding],
@@ -42,10 +43,10 @@ async def add(
         documents: Optional[OneOrMany[Document]] = None,
         images: Optional[OneOrMany[Image]] = None,
         uris: Optional[OneOrMany[URI]] = None,
-    ) -> None:
+    ) -> AddResult:
         """Add embeddings to the data store.
         Args:
-            ids: The ids of the embeddings you wish to add
+            ids: The ids of the embeddings you wish to add. If None, ids will be generated for you. Optional.
             embeddings: The embeddings to add. If None, embeddings will be computed based on the documents or images using the embedding_function set for the Collection. Optional.
             metadatas: The metadata to associate with the embeddings. When querying, you can filter on this metadata. Optional.
             documents: The documents to associate with the embeddings. Optional.
@@ -74,6 +75,7 @@ async def add(
         )
 
         await self._client._add(ids, self.id, embeddings, metadatas, documents, uris)
+        return {"ids": ids}
 
     async def count(self) -> int:
         """The total number of embeddings added to the database

chromadb/api/models/Collection.py

@@ -4,6 +4,7 @@
 from chromadb.api.models.CollectionCommon import CollectionCommon
 from chromadb.api.types import (
     URI,
+    AddResult,
     CollectionMetadata,
     Embedding,
     Include,
@@ -39,7 +40,7 @@ def count(self) -> int:
 
     def add(
         self,
-        ids: OneOrMany[ID],
+        ids: Optional[OneOrMany[ID]] = None,
         embeddings: Optional[  # type: ignore[type-arg]
             Union[
                 OneOrMany[Embedding],
@@ -50,10 +51,10 @@ def add(
         documents: Optional[OneOrMany[Document]] = None,
         images: Optional[OneOrMany[Image]] = None,
         uris: Optional[OneOrMany[URI]] = None,
-    ) -> None:
+    ) -> AddResult:
         """Add embeddings to the data store.
         Args:
-            ids: The ids of the embeddings you wish to add
+            ids: The ids of the embeddings you wish to add. If None, ids will be generated for you. Optional.
             embeddings: The embeddings to add. If None, embeddings will be computed based on the documents or images using the embedding_function set for the Collection. Optional.
             metadatas: The metadata to associate with the embeddings. When querying, you can filter on this metadata. Optional.
             documents: The documents to associate with the embeddings. Optional.
@@ -82,6 +83,7 @@ def add(
         )
 
         self._client._add(ids, self.id, embeddings, metadatas, documents, uris)
+        return {"ids": ids}
 
     def get(
         self,

chromadb/api/models/CollectionCommon.py

@@ -10,7 +10,7 @@
     cast,
 )
 import numpy as np
-from uuid import UUID
+from uuid import UUID, uuid4
 
 import chromadb.utils.embedding_functions as ef
 from chromadb.api.types import (
@@ -243,7 +243,7 @@ def _validate_embedding_set(
 
     def _validate_and_prepare_embedding_set(
         self,
-        ids: OneOrMany[ID],
+        ids: Optional[OneOrMany[ID]],
         embeddings: Optional[  # type: ignore[type-arg]
             Union[
                 OneOrMany[Embedding],
@@ -261,6 +261,30 @@ def _validate_and_prepare_embedding_set(
         Optional[Documents],
         Optional[URIs],
     ]:
+        def count_one_or_many(value: OneOrMany[Any]) -> int:
+            if isinstance(value, list):
+                return len(value)
+            return 1
+
+        if ids is None:
+            if embeddings:
+                if isinstance(embeddings[0], list) and len(embeddings[0]) > 1:
+                    set_size = len(embeddings)
+                else:
+                    set_size = 1
+            elif documents:
+                set_size = count_one_or_many(documents)
+            elif images:
+                set_size = count_one_or_many(images)
+            elif uris:
+                set_size = count_one_or_many(uris)
+            else:
+                raise ValueError(
+                    "You must provide either ids, embeddings, documents, images, or uris."
+                )
+
+            ids = [str(uuid4()) for _ in range(set_size)]
+
         (
             ids,
             embeddings,
@@ -269,7 +293,7 @@ def _validate_and_prepare_embedding_set(
             images,
             uris,
         ) = self._validate_embedding_set(
-            ids, embeddings, metadatas, documents, images, uris
+            cast(OneOrMany[ID], ids), embeddings, metadatas, documents, images, uris
         )
 
         # We need to compute the embeddings if they're not provided

chromadb/api/types.py

@@ -165,6 +165,10 @@ class IncludeEnum(str, Enum):
 L = TypeVar("L", covariant=True, bound=Loadable)
 
 
+class AddResult(TypedDict):
+    ids: List[ID]
+
+
 class GetResult(TypedDict):
     ids: List[ID]
     embeddings: Optional[List[Embedding]]

chromadb/test/property/test_embeddings.py

@@ -134,12 +134,14 @@ def add_embeddings(self, record_set: strategies.RecordSet) -> MultipleResults[ID
                 if normalized_record_set["embeddings"]
                 else None,
             }
-            self.collection.add(**normalized_record_set)  # type: ignore[arg-type]
+            result = self.collection.add(**normalized_record_set)  # type: ignore[arg-type]
+            assert result["ids"] == normalized_record_set["ids"]
             self._upsert_embeddings(cast(strategies.RecordSet, filtered_record_set))
             return multiple(*filtered_record_set["ids"])
 
         else:
-            self.collection.add(**normalized_record_set)  # type: ignore[arg-type]
+            result = self.collection.add(**normalized_record_set)  # type: ignore[arg-type]
+            assert result["ids"] == normalized_record_set["ids"]
             self._upsert_embeddings(cast(strategies.RecordSet, normalized_record_set))
             return multiple(*normalized_record_set["ids"])
 

chromadb/test/test_api.py

@@ -1,13 +1,16 @@
 # type: ignore
 import traceback
 import httpx
+from hypothesis import given
+import hypothesis.strategies as st
 
 import chromadb
 from chromadb.errors import ChromaError
 from chromadb.api.fastapi import FastAPI
 from chromadb.api.types import QueryResult, EmbeddingFunction, Document
 from chromadb.config import Settings
 from chromadb.errors import InvalidCollectionException
+import chromadb.test.property.strategies as strategies
 import chromadb.server.fastapi
 import pytest
 import tempfile
@@ -225,6 +228,19 @@ def test_add(client):
     assert collection.count() == 2
 
 
+collection_st = st.shared(strategies.collections(), key="coll")
+
+
+@given(
+    record_set=strategies.recordsets(collection_st, min_size=1, max_size=5),
+)
+def test_add_without_ids(client, record_set):
+    client.reset()
+    collection = client.create_collection("testspace")
+    result = collection.add(**{k: v for k, v in record_set.items() if k != "ids"})
+    assert collection.count() == len(result["ids"])
+
+
 def test_collection_add_with_invalid_collection_throws(client):
     client.reset()
     collection = client.create_collection("test")