From 35b0c8efa3f087ef00d06e151701283306b48773 Mon Sep 17 00:00:00 2001
From: James Briggs <35938317+jamescalam@users.noreply.github.com>
Date: Sat, 2 Dec 2023 11:40:23 -0800
Subject: [PATCH] example notebook for hybrid
---
docs/examples/hybrid-layer.ipynb | 248 +++++--------------------------
1 file changed, 37 insertions(+), 211 deletions(-)
diff --git a/docs/examples/hybrid-layer.ipynb b/docs/examples/hybrid-layer.ipynb
index d3fb58c5..fa6b78c7 100644
--- a/docs/examples/hybrid-layer.ipynb
+++ b/docs/examples/hybrid-layer.ipynb
@@ -52,7 +52,7 @@
"source": [
"import os\n",
"\n",
- "os.environ[\"COHERE_API_KEY\"] = \"<<APIKEY>>\""
+ "os.environ[\"COHERE_API_KEY\"] = \"<<COHERE_API_KEY>>\""
]
},
{
@@ -78,7 +78,8 @@
" utterances=[\n",
" \"isn't politics the best thing ever\",\n",
" \"why don't you tell me about your political opinions\",\n",
- " \"don't you just love the president\" \"don't you just hate the president\",\n",
+ " \"don't you just love the president\",\n",
+ " \"don't you just hate the president\",\n",
" \"they're going to destroy this country!\",\n",
" \"they will save the country!\",\n",
" ],\n",
@@ -98,6 +99,17 @@
"metadata": {},
"outputs": [],
"source": [
+ "chitchat = Decision(\n",
+ " name=\"chitchat\",\n",
+ " utterances=[\n",
+ " \"how's the weather today?\",\n",
+ " \"how are things going?\",\n",
+ " \"lovely weather today\",\n",
+ " \"the weather is horrendous\",\n",
+ " \"let's go to the chippy\",\n",
+ " ],\n",
+ ")\n",
+ "\n",
"chitchat = Decision(\n",
" name=\"chitchat\",\n",
" utterances=[\n",
@@ -146,15 +158,7 @@
"cell_type": "code",
"execution_count": 5,
"metadata": {},
- "outputs": [
- {
- "name": "stdout",
- "output_type": "stream",
- "text": [
- "3454774732 not in encoder.idx_mapping\n"
- ]
- }
- ],
+ "outputs": [],
"source": [
"from semantic_router.layer import HybridDecisionLayer\n",
"\n",
@@ -167,18 +171,14 @@
"metadata": {},
"outputs": [
{
- "ename": "AxisError",
- "evalue": "axis 1 is out of bounds for array of dimension 1",
- "output_type": "error",
- "traceback": [
- "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
- "\u001b[0;31mAxisError\u001b[0m Traceback (most recent call last)",
- "\u001b[1;32m/Users/jamesbriggs/Documents/projects/aurelio-labs/semantic-router/docs/examples/hybrid-layer.ipynb Cell 15\u001b[0m line \u001b[0;36m1\n\u001b[0;32m----> <a href='vscode-notebook-cell:/Users/jamesbriggs/Documents/projects/aurelio-labs/semantic-router/docs/examples/hybrid-layer.ipynb#X20sZmlsZQ%3D%3D?line=0'>1</a>\u001b[0m dl(\u001b[39m\"\u001b[39;49m\u001b[39mdon\u001b[39;49m\u001b[39m'\u001b[39;49m\u001b[39mt you love politics?\u001b[39;49m\u001b[39m\"\u001b[39;49m)\n",
- "File \u001b[0;32m~/Documents/projects/aurelio-labs/semantic-router/semantic_router/layer.py:141\u001b[0m, in \u001b[0;36mHybridDecisionLayer.__call__\u001b[0;34m(self, text)\u001b[0m\n\u001b[1;32m 140\u001b[0m \u001b[39mdef\u001b[39;00m \u001b[39m__call__\u001b[39m(\u001b[39mself\u001b[39m, text: \u001b[39mstr\u001b[39m) \u001b[39m-\u001b[39m\u001b[39m>\u001b[39m \u001b[39mstr\u001b[39m \u001b[39m|\u001b[39m \u001b[39mNone\u001b[39;00m:\n\u001b[0;32m--> 141\u001b[0m results \u001b[39m=\u001b[39m \u001b[39mself\u001b[39;49m\u001b[39m.\u001b[39;49m_query(text)\n\u001b[1;32m 142\u001b[0m top_class, top_class_scores \u001b[39m=\u001b[39m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39m_semantic_classify(results)\n\u001b[1;32m 143\u001b[0m passed \u001b[39m=\u001b[39m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39m_pass_threshold(top_class_scores, \u001b[39mself\u001b[39m\u001b[39m.\u001b[39mscore_threshold)\n",
- "File \u001b[0;32m~/Documents/projects/aurelio-labs/semantic-router/semantic_router/layer.py:204\u001b[0m, in \u001b[0;36mHybridDecisionLayer._query\u001b[0;34m(self, text, top_k)\u001b[0m\n\u001b[1;32m 202\u001b[0m sim_d \u001b[39m=\u001b[39m np\u001b[39m.\u001b[39mdot(\u001b[39mself\u001b[39m\u001b[39m.\u001b[39mindex, xq_d\u001b[39m.\u001b[39mT) \u001b[39m/\u001b[39m (index_norm \u001b[39m*\u001b[39m xq_d_norm)\n\u001b[1;32m 203\u001b[0m \u001b[39m# calculate sparse vec similarity\u001b[39;00m\n\u001b[0;32m--> 204\u001b[0m sparse_norm \u001b[39m=\u001b[39m norm(\u001b[39mself\u001b[39;49m\u001b[39m.\u001b[39;49msparse_index, axis\u001b[39m=\u001b[39;49m\u001b[39m1\u001b[39;49m)\n\u001b[1;32m 205\u001b[0m xq_s_norm \u001b[39m=\u001b[39m norm(xq_s\u001b[39m.\u001b[39mT)\n\u001b[1;32m 206\u001b[0m sim_s \u001b[39m=\u001b[39m np\u001b[39m.\u001b[39mdot(\u001b[39mself\u001b[39m\u001b[39m.\u001b[39msparse_index, xq_s\u001b[39m.\u001b[39mT) \u001b[39m/\u001b[39m (sparse_norm \u001b[39m*\u001b[39m xq_s_norm)\n",
- "File \u001b[0;32m~/opt/anaconda3/envs/decision-layer/lib/python3.11/site-packages/numpy/linalg/linalg.py:2583\u001b[0m, in \u001b[0;36mnorm\u001b[0;34m(x, ord, axis, keepdims)\u001b[0m\n\u001b[1;32m 2580\u001b[0m \u001b[39melif\u001b[39;00m \u001b[39mord\u001b[39m \u001b[39mis\u001b[39;00m \u001b[39mNone\u001b[39;00m \u001b[39mor\u001b[39;00m \u001b[39mord\u001b[39m \u001b[39m==\u001b[39m \u001b[39m2\u001b[39m:\n\u001b[1;32m 2581\u001b[0m \u001b[39m# special case for speedup\u001b[39;00m\n\u001b[1;32m 2582\u001b[0m s \u001b[39m=\u001b[39m (x\u001b[39m.\u001b[39mconj() \u001b[39m*\u001b[39m x)\u001b[39m.\u001b[39mreal\n\u001b[0;32m-> 2583\u001b[0m \u001b[39mreturn\u001b[39;00m sqrt(add\u001b[39m.\u001b[39;49mreduce(s, axis\u001b[39m=\u001b[39;49maxis, keepdims\u001b[39m=\u001b[39;49mkeepdims))\n\u001b[1;32m 2584\u001b[0m \u001b[39m# None of the str-type keywords for ord ('fro', 'nuc')\u001b[39;00m\n\u001b[1;32m 2585\u001b[0m \u001b[39m# are valid for vectors\u001b[39;00m\n\u001b[1;32m 2586\u001b[0m \u001b[39melif\u001b[39;00m \u001b[39misinstance\u001b[39m(\u001b[39mord\u001b[39m, \u001b[39mstr\u001b[39m):\n",
- "\u001b[0;31mAxisError\u001b[0m: axis 1 is out of bounds for array of dimension 1"
- ]
+ "data": {
+ "text/plain": [
+ "'politics'"
+ ]
+ },
+ "execution_count": 6,
+ "metadata": {},
+ "output_type": "execute_result"
}
],
"source": [
@@ -187,178 +187,20 @@
},
{
"cell_type": "code",
- "execution_count": null,
- "metadata": {},
- "outputs": [],
- "source": []
- },
- {
- "cell_type": "code",
- "execution_count": null,
- "metadata": {},
- "outputs": [],
- "source": []
- },
- {
- "cell_type": "markdown",
- "metadata": {},
- "source": [
- "---\n",
- "\n",
- "#### Testing"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": null,
- "metadata": {},
- "outputs": [],
- "source": [
- "from semantic_router.encoders import BM25Encoder\n",
- "\n",
- "encoder = BM25Encoder()"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": null,
- "metadata": {},
- "outputs": [],
- "source": [
- "tests = [\"hello this is some text\", \"and more stuff\"]"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": null,
- "metadata": {},
- "outputs": [],
- "source": [
- "idx_list = encoder.model.get_params()['doc_freq']['indices']\n",
- "idx_list"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": null,
- "metadata": {},
- "outputs": [],
- "source": [
- "sparse_dicts = encoder.model.encode_documents(tests)\n",
- "sparse_dicts"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": null,
- "metadata": {},
- "outputs": [],
- "source": [
- "embeds = [0.0] * len(encoder.idx_mapping)"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": null,
+ "execution_count": 7,
"metadata": {},
- "outputs": [],
- "source": [
- "for output in sparse_dicts:\n",
- " indices = output[\"indices\"]\n",
- " values = output[\"values\"]\n",
- " for idx, val in zip(indices, values):\n",
- " position = encoder.idx_mapping[idx]\n",
- " embeds[position] = val"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": null,
- "metadata": {},
- "outputs": [],
- "source": [
- "encoder.idx_mapping"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": null,
- "metadata": {},
- "outputs": [],
- "source": [
- "encoded_output = encoder(tests)\n",
- "encoded_output"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": null,
- "metadata": {},
- "outputs": [],
- "source": [
- "import numpy as np\n",
- "\n",
- "\n",
- "sparse_vec = np.zeros(len(idx_list))\n",
- "idx_position_dict = {idx: i for i, idx in enumerate(idx_list)}\n",
- "\n",
- "for output in encoded_output:\n",
- " indices = output['indices']\n",
- " values = output['values']\n",
- " for idx, value in zip(indices, values):\n",
- " if idx in idx_position_dict:\n",
- " position = idx_position_dict[idx]\n",
- " sparse_vec[position] = value"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": null,
- "metadata": {},
- "outputs": [],
- "source": [
- "sparse_vec"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": null,
- "metadata": {},
- "outputs": [],
- "source": [
- "sparse_vec.shape"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {},
- "source": [
- "#### Finish Testing\n",
- "\n",
- "---"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {},
- "source": [
- "Now we can test it:"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": null,
- "metadata": {},
- "outputs": [],
- "source": [
- "dl(\"don't you love politics?\")"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": null,
- "metadata": {},
- "outputs": [],
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "'chitchat'"
+ ]
+ },
+ "execution_count": 7,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
"source": [
"dl(\"how's the weather today?\")"
]
@@ -367,23 +209,7 @@
"cell_type": "markdown",
"metadata": {},
"source": [
- "Both are classified accurately, what if we send a query that is unrelated to our existing `Decision` objects?"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": null,
- "metadata": {},
- "outputs": [],
- "source": [
- "dl(\"I'm interested in learning about llama 2\")"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {},
- "source": [
- "In this case, we return `None` because no matches were identified."
+ "---"
]
}
],
--
GitLab