From 1d777b0f5933762dc55b500f7bb52a8e8806cb29 Mon Sep 17 00:00:00 2001
From: James Briggs <35938317+jamescalam@users.noreply.github.com>
Date: Sun, 12 Jan 2025 11:17:15 +0000
Subject: [PATCH] fix: init index logic
---
semantic_router/index/pinecone.py | 14 ++++++++++++--
semantic_router/routers/base.py | 4 ++++
semantic_router/routers/hybrid.py | 3 ---
semantic_router/routers/semantic.py | 3 ---
4 files changed, 16 insertions(+), 8 deletions(-)
diff --git a/semantic_router/index/pinecone.py b/semantic_router/index/pinecone.py
index f1ec2a7c..c0176992 100644
--- a/semantic_router/index/pinecone.py
+++ b/semantic_router/index/pinecone.py
@@ -236,7 +236,12 @@ class PineconeIndex(BaseIndex):
else:
# if the index doesn't exist and we don't have the dimensions
# we return None
- logger.warning("Index could not be initialized.")
+ logger.warning(
+ "Index could not be initialized. Init parameters: "
+ f"{self.index_name=}, {self.dimensions=}, {self.metric=}, "
+ f"{self.cloud=}, {self.region=}, {self.host=}, {self.namespace=}, "
+ f"{force_create=}"
+ )
index = None
if index is not None:
self.host = self.client.describe_index(self.index_name)["host"]
@@ -272,7 +277,12 @@ class PineconeIndex(BaseIndex):
else:
# if the index doesn't exist and we don't have the dimensions
# we raise warning
- logger.warning("Index could not be initialized.")
+ logger.warning(
+ "Index could not be initialized. Init parameters: "
+ f"{self.index_name=}, {self.dimensions=}, {self.metric=}, "
+ f"{self.cloud=}, {self.region=}, {self.host=}, {self.namespace=}, "
+ f"{force_create=}"
+ )
self.host = index_stats["host"] if index_stats else ""
def _batch_upsert(self, batch: List[Dict]):
diff --git a/semantic_router/routers/base.py b/semantic_router/routers/base.py
index ae2108f2..25cb8b6b 100644
--- a/semantic_router/routers/base.py
+++ b/semantic_router/routers/base.py
@@ -351,6 +351,8 @@ class BaseRouter(BaseModel):
for route in self.routes:
if route.score_threshold is None:
route.score_threshold = self.score_threshold
+ # initialize index
+ self._init_index_state()
def _get_index(self, index: Optional[BaseIndex]) -> BaseIndex:
if index is None:
@@ -370,6 +372,7 @@ class BaseRouter(BaseModel):
def _init_index_state(self):
"""Initializes an index (where required) and runs auto_sync if active."""
+ print("JBTEMP _init_index_state")
# initialize index now, check if we need dimensions
if self.index.dimensions is None:
dims = len(self.encoder(["test"])[0])
@@ -862,6 +865,7 @@ class BaseRouter(BaseModel):
The name must exist within the local SemanticRouter, if not a
KeyError will be raised.
"""
+ # TODO JB: should modify update to take a Route object
current_local_hash = self._get_hash()
current_remote_hash = self.index._read_hash()
if current_remote_hash.value == "":
diff --git a/semantic_router/routers/hybrid.py b/semantic_router/routers/hybrid.py
index 241fb57c..70e00440 100644
--- a/semantic_router/routers/hybrid.py
+++ b/semantic_router/routers/hybrid.py
@@ -62,9 +62,6 @@ class HybridRouter(BaseRouter):
and self.routes
):
self.sparse_encoder.fit(self.routes)
- # run initialize index now if auto sync is active
- if self.auto_sync:
- self._init_index_state()
def _set_score_threshold(self):
"""Set the score threshold for the HybridRouter. Unlike the base router the
diff --git a/semantic_router/routers/semantic.py b/semantic_router/routers/semantic.py
index 41c92d53..5efa2e48 100644
--- a/semantic_router/routers/semantic.py
+++ b/semantic_router/routers/semantic.py
@@ -32,9 +32,6 @@ class SemanticRouter(BaseRouter):
aggregation=aggregation,
auto_sync=auto_sync,
)
- # run initialize index now if auto sync is active
- if self.auto_sync:
- self._init_index_state()
def _encode(self, text: list[str]) -> Any:
"""Given some text, encode it."""
--
GitLab