Spaces:
Sleeping
Sleeping
File size: 7,135 Bytes
cfd3735 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 |
{
"cells": [
{
"attachments": {},
"cell_type": "markdown",
"id": "ab66dd43",
"metadata": {},
"source": [
"# Pinecone Hybrid Search\n",
"\n",
"This notebook goes over how to use a retriever that under the hood uses Pinecone and Hybrid Search.\n",
"\n",
"The logic of this retriever is taken from [this documentaion](https://docs.pinecone.io/docs/hybrid-search)"
]
},
{
"cell_type": "code",
"execution_count": 75,
"id": "393ac030",
"metadata": {},
"outputs": [],
"source": [
"from langchain.retrievers import PineconeHybridSearchRetriever"
]
},
{
"cell_type": "markdown",
"id": "aaf80e7f",
"metadata": {},
"source": [
"## Setup Pinecone"
]
},
{
"attachments": {},
"cell_type": "markdown",
"id": "95d5d7f9",
"metadata": {},
"source": [
"You should only have to do this part once.\n",
"\n",
"Note: it's important to make sure that the \"context\" field that holds the document text in the metadata is not indexed. Currently you need to specify explicitly the fields you do want to index. For more information checkout Pinecone's [docs](https://docs.pinecone.io/docs/manage-indexes#selective-metadata-indexing)."
]
},
{
"cell_type": "code",
"execution_count": 76,
"id": "3b8f7697",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"WhoAmIResponse(username='load', user_label='label', projectname='load-test')"
]
},
"execution_count": 76,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"import os\n",
"import pinecone\n",
"\n",
"api_key = os.getenv(\"PINECONE_API_KEY\") or \"PINECONE_API_KEY\"\n",
"# find environment next to your API key in the Pinecone console\n",
"env = os.getenv(\"PINECONE_ENVIRONMENT\") or \"PINECONE_ENVIRONMENT\"\n",
"\n",
"index_name = \"langchain-pinecone-hybrid-search\"\n",
"\n",
"pinecone.init(api_key=api_key, enviroment=env)\n",
"pinecone.whoami()"
]
},
{
"cell_type": "code",
"execution_count": 77,
"id": "cfa3a8d8",
"metadata": {},
"outputs": [],
"source": [
" # create the index\n",
"pinecone.create_index(\n",
" name = index_name,\n",
" dimension = 1536, # dimensionality of dense model\n",
" metric = \"dotproduct\", # sparse values supported only for dotproduct\n",
" pod_type = \"s1\",\n",
" metadata_config={\"indexed\": []} # see explaination above\n",
")"
]
},
{
"cell_type": "markdown",
"id": "e01549af",
"metadata": {},
"source": [
"Now that its created, we can use it"
]
},
{
"cell_type": "code",
"execution_count": 78,
"id": "bcb3c8c2",
"metadata": {},
"outputs": [],
"source": [
"index = pinecone.Index(index_name)"
]
},
{
"attachments": {},
"cell_type": "markdown",
"id": "dbc025d6",
"metadata": {},
"source": [
"## Get embeddings and sparse encoders\n",
"\n",
"Embeddings are used for the dense vectors, tokenizer is used for the sparse vector"
]
},
{
"cell_type": "code",
"execution_count": 79,
"id": "2f63c911",
"metadata": {},
"outputs": [],
"source": [
"from langchain.embeddings import OpenAIEmbeddings\n",
"embeddings = OpenAIEmbeddings()"
]
},
{
"attachments": {},
"cell_type": "markdown",
"id": "96bf8879",
"metadata": {},
"source": [
"To encode the text to sparse values you can either choose SPLADE or BM25. For out of domain tasks we recommend using BM25.\n",
"\n",
"For more information about the sparse encoders you can checkout pinecone-text library [docs](https://pinecone-io.github.io/pinecone-text/pinecone_text.html)."
]
},
{
"cell_type": "code",
"execution_count": 80,
"id": "c3f030e5",
"metadata": {},
"outputs": [],
"source": [
"from pinecone_text.sparse import BM25Encoder\n",
"# or from pinecone_text.sparse import SpladeEncoder if you wish to work with SPLADE\n",
"\n",
"# use default tf-idf values\n",
"bm25_encoder = BM25Encoder().default()"
]
},
{
"attachments": {},
"cell_type": "markdown",
"id": "23601ddb",
"metadata": {},
"source": [
"The above code is using default tfids values. It's highly recommended to fit the tf-idf values to your own corpus. You can do it as follow:\n",
"\n",
"```python\n",
"corpus = [\"foo\", \"bar\", \"world\", \"hello\"]\n",
"\n",
"# fit tf-idf values on your corpus\n",
"bm25_encoder.fit(corpus)\n",
"\n",
"# store the values to a json file\n",
"bm25_encoder.dump(\"bm25_values.json\")\n",
"\n",
"# load to your BM25Encoder object\n",
"bm25_encoder = BM25Encoder().load(\"bm25_values.json\")\n",
"```"
]
},
{
"cell_type": "markdown",
"id": "5462801e",
"metadata": {},
"source": [
"## Load Retriever\n",
"\n",
"We can now construct the retriever!"
]
},
{
"cell_type": "code",
"execution_count": 81,
"id": "ac77d835",
"metadata": {},
"outputs": [],
"source": [
"retriever = PineconeHybridSearchRetriever(embeddings=embeddings, sparse_encoder=bm25_encoder, index=index)"
]
},
{
"cell_type": "markdown",
"id": "1c518c42",
"metadata": {},
"source": [
"## Add texts (if necessary)\n",
"\n",
"We can optionally add texts to the retriever (if they aren't already in there)"
]
},
{
"cell_type": "code",
"execution_count": 82,
"id": "98b1c017",
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"100%|██████████| 1/1 [00:02<00:00, 2.27s/it]\n"
]
}
],
"source": [
"retriever.add_texts([\"foo\", \"bar\", \"world\", \"hello\"])"
]
},
{
"cell_type": "markdown",
"id": "08437fa2",
"metadata": {},
"source": [
"## Use Retriever\n",
"\n",
"We can now use the retriever!"
]
},
{
"cell_type": "code",
"execution_count": 83,
"id": "c0455218",
"metadata": {},
"outputs": [],
"source": [
"result = retriever.get_relevant_documents(\"foo\")"
]
},
{
"cell_type": "code",
"execution_count": 84,
"id": "7dfa5c29",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Document(page_content='foo', metadata={})"
]
},
"execution_count": 84,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"result[0]"
]
}
],
"metadata": {
"kernelspec": {
"display_name": ".venv",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.9.13"
},
"vscode": {
"interpreter": {
"hash": "7ec0d8babd8cabf695a1d94b1e586d626e046c9df609f6bad065d15d49f67f54"
}
}
},
"nbformat": 4,
"nbformat_minor": 5
}
|