From 8c0ba48000dda79be7d8c48da18ff997752902d1 Mon Sep 17 00:00:00 2001
From: tolgadevAI <164843802+tolgadevAI@users.noreply.github.com>
Date: Mon, 19 Aug 2024 23:41:41 +0300
Subject: [PATCH] develop the process of storing the function_schemas field for
the route_layer add function on Pinecone
---
semantic_router/index/base.py | 1 +
semantic_router/index/pinecone.py | 19 ++++++++++++++++---
semantic_router/layer.py | 7 ++++++-
3 files changed, 23 insertions(+), 4 deletions(-)
diff --git a/semantic_router/index/base.py b/semantic_router/index/base.py
index d0f12ac6..d25d41dc 100644
--- a/semantic_router/index/base.py
+++ b/semantic_router/index/base.py
@@ -26,6 +26,7 @@ class BaseIndex(BaseModel):
embeddings: List[List[float]],
routes: List[str],
utterances: List[Any],
+ function_schemas: List[Dict[str, Any]],
):
"""
Add embeddings to the index.
diff --git a/semantic_router/index/pinecone.py b/semantic_router/index/pinecone.py
index b4f6033f..0b84ad4c 100644
--- a/semantic_router/index/pinecone.py
+++ b/semantic_router/index/pinecone.py
@@ -22,6 +22,7 @@ class PineconeRecord(BaseModel):
values: List[float]
route: str
utterance: str
+ function_schema: str
def __init__(self, **data):
super().__init__(**data)
@@ -34,7 +35,11 @@ class PineconeRecord(BaseModel):
return {
"id": self.id,
"values": self.values,
- "metadata": {"sr_route": self.route, "sr_utterance": self.utterance},
+ "metadata": {
+ "sr_route": self.route,
+ "sr_utterance": self.utterance,
+ "sr_function_schemas": self.function_schema,
+ },
}
@@ -305,6 +310,7 @@ class PineconeIndex(BaseIndex):
embeddings: List[List[float]],
routes: List[str],
utterances: List[str],
+ function_schemas: List[Dict[str, Any]] = "",
batch_size: int = 100,
):
"""Add vectors to Pinecone in batches."""
@@ -313,8 +319,15 @@ class PineconeIndex(BaseIndex):
self.index = self._init_index(force_create=True)
vectors_to_upsert = [
- PineconeRecord(values=vector, route=route, utterance=utterance).to_dict()
- for vector, route, utterance in zip(embeddings, routes, utterances)
+ PineconeRecord(
+ values=vector,
+ route=route,
+ utterance=utterance,
+ function_schema=str(function_schema),
+ ).to_dict()
+ for vector, route, utterance, function_schema in zip(
+ embeddings, routes, utterances, function_schemas
+ )
]
for i in range(0, len(vectors_to_upsert), batch_size):
diff --git a/semantic_router/layer.py b/semantic_router/layer.py
index 6b548fc0..2285a7d9 100644
--- a/semantic_router/layer.py
+++ b/semantic_router/layer.py
@@ -430,12 +430,17 @@ class RouteLayer:
if route.score_threshold is None:
route.score_threshold = self.score_threshold
- # add routes to the index
self.index.add(
embeddings=embeds,
routes=[route.name] * len(route.utterances),
utterances=route.utterances,
+ function_schemas=(
+ route.function_schemas * len(route.utterances)
+ if route.function_schemas
+ else [""] * len(route.utterances)
+ ),
)
+
self.routes.append(route)
def list_route_names(self) -> List[str]:
--
GitLab