@nestbox-ai/cli 1.0.62 → 1.0.63
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
|
@@ -873,6 +873,29 @@ graphrag:
|
|
|
873
873
|
|
|
874
874
|
---
|
|
875
875
|
|
|
876
|
+
### Embeddings
|
|
877
|
+
|
|
878
|
+
The `embeddings` section controls how entity and community texts are vectorised for similarity search. It does **not** have an `enabled` field — to disable embeddings, omit the section entirely.
|
|
879
|
+
|
|
880
|
+
```yaml
|
|
881
|
+
graphrag:
|
|
882
|
+
embeddings:
|
|
883
|
+
model: text-embedding-3-large # embedding model (matches graphrag.models.embeddingModel)
|
|
884
|
+
dimensions: 3072 # vector dimensions (3072 for text-embedding-3-large)
|
|
885
|
+
batchSize: 100 # texts per API batch
|
|
886
|
+
```
|
|
887
|
+
|
|
888
|
+
Fields:
|
|
889
|
+
- `model` — the embedding model name. Must match what `graphrag.models.embeddingModel` uses.
|
|
890
|
+
- `text-embedding-3-large` → set `dimensions: 3072` (default)
|
|
891
|
+
- `text-embedding-3-small` → set `dimensions: 1536`
|
|
892
|
+
- `dimensions` — must match the model's output dimensions.
|
|
893
|
+
- `batchSize` — number of text chunks sent per embedding request. Default `100`.
|
|
894
|
+
|
|
895
|
+
> **Do not write `enabled: true` or any other key here.** The schema only allows `model`, `dimensions`, and `batchSize`.
|
|
896
|
+
|
|
897
|
+
---
|
|
898
|
+
|
|
876
899
|
### Local Search
|
|
877
900
|
|
|
878
901
|
Local search answers specific, entity-focused questions ("What is the rent?", "Who are the parties?").
|