diff --git a/pyproject.toml b/pyproject.toml
index 76ac7539cb2b21a902572467ed188ed871a34573..9abd9c7cdb8b648b4b0d3ff2bf496d8e233500f6 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -1,6 +1,6 @@
[tool.poetry]
name = "semantic-router"
-version = "0.0.5"
+version = "0.0.7"
description = "Super fast semantic router for AI decision making"
authors = [
"James Briggs <james@aurelio.ai>",
diff --git a/semantic_router/__init__.py b/semantic_router/__init__.py
index 734906f8554e11898fbefe8c95ef92a555dc5c8d..b10cc687b0924faf516432e89f87bc1f95bb0c9e 100644
--- a/semantic_router/__init__.py
+++ b/semantic_router/__init__.py
@@ -1,3 +1,3 @@
-from .layer import DecisionLayer, HybridDecisionLayer
+from .layer import RouteLayer, HybridRouteLayer
-__all__ = ["DecisionLayer", "HybridDecisionLayer"]
+__all__ = ["RouteLayer", "HybridRouteLayer"]
diff --git a/semantic_router/layer.py b/semantic_router/layer.py
index adff961c16cb87965f06975afed8b185d4932f8c..c29f190f915aea61630eee2c26943a19b951db59 100644
--- a/semantic_router/layer.py
+++ b/semantic_router/layer.py
@@ -9,15 +9,15 @@ from semantic_router.encoders import (
BM25Encoder,
)
from semantic_router.linear import similarity_matrix, top_scores
-from semantic_router.schema import Decision
+from semantic_router.schema import Route
-class DecisionLayer:
+class RouteLayer:
index = None
categories = None
score_threshold = 0.82
- def __init__(self, encoder: BaseEncoder, decisions: list[Decision] = []):
+ def __init__(self, encoder: BaseEncoder, routes: list[Route] = []):
self.encoder = encoder
# decide on default threshold based on encoder
if isinstance(encoder, OpenAIEncoder):
@@ -26,11 +26,11 @@ class DecisionLayer:
self.score_threshold = 0.3
else:
self.score_threshold = 0.82
- # if decisions list has been passed, we initialize index now
- if decisions:
+ # if routes list has been passed, we initialize index now
+ if routes:
# initialize index now
- for decision in tqdm(decisions):
- self._add_decision(decision=decision)
+ for route in tqdm(routes):
+ self._add_route(route=route)
def __call__(self, text: str) -> str | None:
results = self._query(text)
@@ -41,18 +41,18 @@ class DecisionLayer:
else:
return None
- def add(self, decision: Decision):
- self._add_decision(decision=decision)
+ def add(self, route: Route):
+ self._add_route(route=route)
- def _add_decision(self, decision: Decision):
+ def _add_route(self, route: Route):
# create embeddings
- embeds = self.encoder(decision.utterances)
+ embeds = self.encoder(route.utterances)
- # create decision array
+ # create route array
if self.categories is None:
- self.categories = np.array([decision.name] * len(embeds))
+ self.categories = np.array([route.name] * len(embeds))
else:
- str_arr = np.array([decision.name] * len(embeds))
+ str_arr = np.array([route.name] * len(embeds))
self.categories = np.concatenate([self.categories, str_arr])
# create utterance array (the index)
if self.index is None:
@@ -73,10 +73,10 @@ class DecisionLayer:
# calculate similarity matrix
sim = similarity_matrix(xq, self.index)
scores, idx = top_scores(sim, top_k)
- # get the utterance categories (decision names)
- decisions = self.categories[idx] if self.categories is not None else []
+ # get the utterance categories (route names)
+ routes = self.categories[idx] if self.categories is not None else []
return [
- {"decision": d, "score": s.item()} for d, s in zip(decisions, scores)
+ {"route": d, "score": s.item()} for d, s in zip(routes, scores)
]
else:
return []
@@ -85,15 +85,15 @@ class DecisionLayer:
scores_by_class = {}
for result in query_results:
score = result["score"]
- decision = result["decision"]
- if decision in scores_by_class:
- scores_by_class[decision].append(score)
+ route = result["route"]
+ if route in scores_by_class:
+ scores_by_class[route].append(score)
else:
- scores_by_class[decision] = [score]
+ scores_by_class[route] = [score]
# Calculate total score for each class
total_scores = {
- decision: sum(scores) for decision, scores in scores_by_class.items()
+ route: sum(scores) for route, scores in scores_by_class.items()
}
top_class = max(total_scores, key=lambda x: total_scores[x], default=None)
@@ -107,14 +107,14 @@ class DecisionLayer:
return False
-class HybridDecisionLayer:
+class HybridRouteLayer:
index = None
sparse_index = None
categories = None
score_threshold = 0.82
def __init__(
- self, encoder: BaseEncoder, decisions: list[Decision] = [], alpha: float = 0.3
+ self, encoder: BaseEncoder, routes: list[Route] = [], alpha: float = 0.3
):
self.encoder = encoder
self.sparse_encoder = BM25Encoder()
@@ -126,11 +126,11 @@ class HybridDecisionLayer:
self.score_threshold = 0.3
else:
self.score_threshold = 0.82
- # if decisions list has been passed, we initialize index now
- if decisions:
+ # if routes list has been passed, we initialize index now
+ if routes:
# initialize index now
- for decision in tqdm(decisions):
- self._add_decision(decision=decision)
+ for route in tqdm(routes):
+ self._add_route(route=route)
def __call__(self, text: str) -> str | None:
results = self._query(text)
@@ -141,25 +141,25 @@ class HybridDecisionLayer:
else:
return None
- def add(self, decision: Decision):
- self._add_decision(decision=decision)
+ def add(self, route: Route):
+ self._add_route(route=route)
- def _add_decision(self, decision: Decision):
+ def _add_route(self, route: Route):
# create embeddings
- dense_embeds = np.array(self.encoder(decision.utterances)) # * self.alpha
+ dense_embeds = np.array(self.encoder(route.utterances)) # * self.alpha
sparse_embeds = np.array(
- self.sparse_encoder(decision.utterances)
+ self.sparse_encoder(route.utterances)
) # * (1 - self.alpha)
- # create decision array
+ # create route array
if self.categories is None:
- self.categories = np.array([decision.name] * len(decision.utterances))
- self.utterances = np.array(decision.utterances)
+ self.categories = np.array([route.name] * len(route.utterances))
+ self.utterances = np.array(route.utterances)
else:
- str_arr = np.array([decision.name] * len(decision.utterances))
+ str_arr = np.array([route.name] * len(route.utterances))
self.categories = np.concatenate([self.categories, str_arr])
self.utterances = np.concatenate(
- [self.utterances, np.array(decision.utterances)]
+ [self.utterances, np.array(route.utterances)]
)
# create utterance array (the dense index)
if self.index is None:
@@ -199,10 +199,10 @@ class HybridDecisionLayer:
top_k = min(top_k, total_sim.shape[0])
idx = np.argpartition(total_sim, -top_k)[-top_k:]
scores = total_sim[idx]
- # get the utterance categories (decision names)
- decisions = self.categories[idx] if self.categories is not None else []
+ # get the utterance categories (route names)
+ routes = self.categories[idx] if self.categories is not None else []
return [
- {"decision": d, "score": s.item()} for d, s in zip(decisions, scores)
+ {"route": d, "score": s.item()} for d, s in zip(routes, scores)
]
else:
return []
@@ -217,15 +217,15 @@ class HybridDecisionLayer:
scores_by_class = {}
for result in query_results:
score = result["score"]
- decision = result["decision"]
- if decision in scores_by_class:
- scores_by_class[decision].append(score)
+ route = result["route"]
+ if route in scores_by_class:
+ scores_by_class[route].append(score)
else:
- scores_by_class[decision] = [score]
+ scores_by_class[route] = [score]
# Calculate total score for each class
total_scores = {
- decision: sum(scores) for decision, scores in scores_by_class.items()
+ route: sum(scores) for route, scores in scores_by_class.items()
}
top_class = max(total_scores, key=lambda x: total_scores[x], default=None)
diff --git a/semantic_router/schema.py b/semantic_router/schema.py
index 37a43dd41007634706caedc1b5c5dadd6d734bba..3763db0351b8939b611133f7485c3061c3dab53d 100644
--- a/semantic_router/schema.py
+++ b/semantic_router/schema.py
@@ -10,7 +10,7 @@ from semantic_router.encoders import (
)
-class Decision(BaseModel):
+class Route(BaseModel):
name: str
utterances: list[str]
description: str | None = None
@@ -45,12 +45,12 @@ class Encoder:
@dataclass
class SemanticSpace:
id: str
- decisions: list[Decision]
+ routes: list[Route]
encoder: str = ""
- def __init__(self, decisions: list[Decision] = []):
+ def __init__(self, routes: list[Route] = []):
self.id = ""
- self.decisions = decisions
+ self.routes = routes
- def add(self, decision: Decision):
- self.decisions.append(decision)
+ def add(self, route: Route):
+ self.routes.append(route)
diff --git a/walkthrough.ipynb b/walkthrough.ipynb
index 81bb3ec270a090b699c633d272a40d20bf25c19e..6731ee0a9a2fe72e6ae1b0507e21445ed18207a6 100644
--- a/walkthrough.ipynb
+++ b/walkthrough.ipynb
@@ -11,7 +11,7 @@
"cell_type": "markdown",
"metadata": {},
"source": [
- "The Semantic Router library can be used as a super fast decision making layer on top of LLMs. That means rather than waiting on a slow agent to decide what to do, we can use the magic of semantic vector space to make decisions. Cutting decision making time down from seconds to milliseconds."
+ "The Semantic Router library can be used as a super fast route making layer on top of LLMs. That means rather than waiting on a slow agent to decide what to do, we can use the magic of semantic vector space to make routes. Cutting route making time down from seconds to milliseconds."
]
},
{
@@ -41,18 +41,28 @@
"cell_type": "markdown",
"metadata": {},
"source": [
- "We start by defining a dictionary mapping decisions to example phrases that should trigger those decisions."
+ "We start by defining a dictionary mapping routes to example phrases that should trigger those routes."
]
},
{
"cell_type": "code",
- "execution_count": null,
- "metadata": {},
- "outputs": [],
- "source": [
- "from semantic_router.schema import Decision\n",
+ "execution_count": 1,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stderr",
+ "output_type": "stream",
+ "text": [
+ "/Users/jamesbriggs/opt/anaconda3/envs/decision-layer/lib/python3.11/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
+ " from .autonotebook import tqdm as notebook_tqdm\n",
+ "None of PyTorch, TensorFlow >= 2.0, or Flax have been found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.\n"
+ ]
+ }
+ ],
+ "source": [
+ "from semantic_router.schema import Route\n",
"\n",
- "politics = Decision(\n",
+ "politics = Route(\n",
" name=\"politics\",\n",
" utterances=[\n",
" \"isn't politics the best thing ever\",\n",
@@ -74,11 +84,11 @@
},
{
"cell_type": "code",
- "execution_count": null,
+ "execution_count": 2,
"metadata": {},
"outputs": [],
"source": [
- "chitchat = Decision(\n",
+ "chitchat = Route(\n",
" name=\"chitchat\",\n",
" utterances=[\n",
" \"how's the weather today?\",\n",
@@ -89,7 +99,7 @@
" ]\n",
")\n",
"\n",
- "decisions = [politics, chitchat]"
+ "routes = [politics, chitchat]"
]
},
{
@@ -101,7 +111,7 @@
},
{
"cell_type": "code",
- "execution_count": null,
+ "execution_count": 5,
"metadata": {},
"outputs": [],
"source": [
@@ -109,7 +119,7 @@
"from getpass import getpass\n",
"import os\n",
"\n",
- "os.environ[\"COHERE_API_KEY\"] = os.environ[\"COHERE_API_KEY\"] or \\\n",
+ "os.environ[\"COHERE_API_KEY\"] = os.getenv(\"COHERE_API_KEY\") or \\\n",
" getpass(\"Enter Cohere API Key: \")\n",
"\n",
"encoder = CohereEncoder()"
@@ -119,18 +129,26 @@
"cell_type": "markdown",
"metadata": {},
"source": [
- "Now we define the `DecisionLayer`. When called, the decision layer will consume text (a query) and output the category (`Decision`) it belongs to — to initialize a `DecisionLayer` we need our `encoder` model and a list of `decisions`."
+ "Now we define the `RouteLayer`. When called, the route layer will consume text (a query) and output the category (`Route`) it belongs to — to initialize a `RouteLayer` we need our `encoder` model and a list of `routes`."
]
},
{
"cell_type": "code",
- "execution_count": null,
- "metadata": {},
- "outputs": [],
- "source": [
- "from semantic_router import DecisionLayer\n",
+ "execution_count": 6,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stderr",
+ "output_type": "stream",
+ "text": [
+ "100%|██████████| 2/2 [00:01<00:00, 1.04it/s]\n"
+ ]
+ }
+ ],
+ "source": [
+ "from semantic_router.layer import RouteLayer\n",
"\n",
- "dl = DecisionLayer(encoder=encoder, decisions=decisions)"
+ "dl = RouteLayer(encoder=encoder, routes=routes)"
]
},
{
@@ -142,18 +160,40 @@
},
{
"cell_type": "code",
- "execution_count": null,
- "metadata": {},
- "outputs": [],
+ "execution_count": 7,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "'politics'"
+ ]
+ },
+ "execution_count": 7,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
"source": [
"dl(\"don't you love politics?\")"
]
},
{
"cell_type": "code",
- "execution_count": null,
- "metadata": {},
- "outputs": [],
+ "execution_count": 8,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "'chitchat'"
+ ]
+ },
+ "execution_count": 8,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
"source": [
"dl(\"how's the weather today?\")"
]
@@ -162,12 +202,12 @@
"cell_type": "markdown",
"metadata": {},
"source": [
- "Both are classified accurately, what if we send a query that is unrelated to our existing `Decision` objects?"
+ "Both are classified accurately, what if we send a query that is unrelated to our existing `Route` objects?"
]
},
{
"cell_type": "code",
- "execution_count": null,
+ "execution_count": 9,
"metadata": {},
"outputs": [],
"source": [