Enhancement: Blockchain Document Loader with better Metadata support (#3710)

This PR includes some minor alignment updates, including:

- metadata object extended to support contractAddress, blockchainType,
and tokenId
- notebook doc better aligned to standard langchain format
- startToken changed from int to str to support multiple hex value types
on the Alchemy API

The updated metadata will look like the below. It's possible for a
single contractAddress to exist across multiple blockchains (e.g.
Ethereum, Polygon, etc.) so it's important to include the
blockchainType.

```
 metadata = {"source": self.contract_address, 
                      "blockchain": self.blockchainType,
                      "tokenId": tokenId}
```
fix_agent_callbacks
Jon Saginaw 1 year ago committed by GitHub
parent 220a7076ac
commit f8d69e4e52
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23

@ -7,7 +7,7 @@
"id": "vm8vn9t8DvC_"
},
"source": [
"# Blockchain Document Loader"
"# Blockchain"
]
},
{
@ -31,7 +31,7 @@
"\n",
"Initially this Loader supports:\n",
"\n",
"\n",
"* Loading NFTs as Documents from NFT Smart Contracts (ERC721 and ERC1155)\n",
"* Ethereum Maninnet, Ethereum Testnet, Polgyon Mainnet, Polygon Testnet (default is eth-mainnet)\n",
"* Alchemy's getNFTsForCollection API\n",
"\n",
@ -39,14 +39,14 @@
"\n",
"* Additional APIs can be added (e.g. Tranction-related APIs)\n",
"\n",
"To run this notebook, the user will need:\n",
"\n",
"This Document Loader Requires:\n",
"\n",
"* An OpenAI key (for OpenAI models)\n",
"* A free [Alchemy API Key](https://www.alchemy.com/)\n",
"\n",
"The output takes the following format:\n",
"\n",
"\n"
"- pageContent= Individual NFT\n",
"- metadata={'source': '0x1a92f7381b9f03921564a437210bb9396471050c', 'blockchain': 'eth-mainnet', 'tokenId': '0x15'})"
]
},
{
@ -54,55 +54,16 @@
"cell_type": "markdown",
"metadata": {},
"source": [
"## Setup"
]
},
{
"cell_type": "code",
"execution_count": 48,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Note: you may need to restart the kernel to use updated packages.\n"
]
}
],
"source": [
"%pip install langchain -q"
"## Load NFTs into Document Loader"
]
},
{
"cell_type": "code",
"execution_count": 49,
"metadata": {},
"outputs": [],
"source": [
"from langchain.document_loaders import BlockchainDocumentLoader\n",
"from langchain.document_loaders.blockchain import BlockchainType\n",
"import os"
]
},
{
"cell_type": "code",
"execution_count": 50,
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"alchemyApiKey = \"get your own key from https://www.alchemy.com/\" \n",
"os.environ[\"ALCHEMY_API_KEY\"] = alchemyApiKey"
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "nzuPWRaBNCMx"
},
"source": [
"## Create a Blockchain Document Loader"
"alchemyApiKey = \"get from https://www.alchemy.com/ and set in environment variable ALCHEMY_API_KEY\""
]
},
{
@ -115,29 +76,18 @@
},
{
"cell_type": "code",
"execution_count": 24,
"execution_count": null,
"metadata": {
"id": "J3LWHARC-Kn0"
},
"outputs": [
{
"data": {
"text/plain": [
"[Document(page_content=\"{'contract': {'address': '0xbc4ca0eda7647a8ab7c2061c2e118a18a936f13d'}, 'id': {'tokenId': '0x0000000000000000000000000000000000000000000000000000000000000000', 'tokenMetadata': {'tokenType': 'ERC721'}}, 'title': '', 'description': '', 'tokenUri': {'gateway': 'https://alchemy.mypinata.cloud/ipfs/QmeSjSinHpPnmXmspMjwiXyN6zS4E9zccariGR3jxcaWtq/0', 'raw': 'ipfs://QmeSjSinHpPnmXmspMjwiXyN6zS4E9zccariGR3jxcaWtq/0'}, 'media': [{'gateway': 'https://nft-cdn.alchemy.com/eth-mainnet/415d618f5fef7bfe683e02d4653c4289', 'thumbnail': 'https://res.cloudinary.com/alchemyapi/image/upload/thumbnailv2/eth-mainnet/415d618f5fef7bfe683e02d4653c4289', 'raw': 'ipfs://QmRRPWG96cmgTn2qSzjwr2qvfNEuhunv6FNeMFGa9bx6mQ', 'format': 'png', 'bytes': 133270}], 'metadata': {'image': 'ipfs://QmRRPWG96cmgTn2qSzjwr2qvfNEuhunv6FNeMFGa9bx6mQ', 'attributes': [{'value': 'Silver Hoop', 'trait_type': 'Earring'}, {'value': 'Orange', 'trait_type': 'Background'}, {'value': 'Robot', 'trait_type': 'Fur'}, {'value': 'Striped Tee', 'trait_type': 'Clothes'}, {'value': 'Discomfort', 'trait_type': 'Mouth'}, {'value': 'X Eyes', 'trait_type': 'Eyes'}]}, 'timeLastUpdated': '2023-04-18T04:05:27.817Z', 'contractMetadata': {'name': 'BoredApeYachtClub', 'symbol': 'BAYC', 'totalSupply': '10000', 'tokenType': 'ERC721', 'contractDeployer': '0xaba7161a7fb69c88e16ed9f455ce62b791ee4d03', 'deployedBlockNumber': 12287507, 'openSea': {'floorPrice': 68.16, 'collectionName': 'Bored Ape Yacht Club', 'safelistRequestStatus': 'verified', 'imageUrl': 'https://i.seadn.io/gae/Ju9CkWtV-1Okvf45wo8UctR-M9He2PjILP0oOvxE89AyiPPGtrR3gysu1Zgy0hjd2xKIgjJJtWIc0ybj4Vd7wv8t3pxDGHoJBzDB?w=500&auto=format', 'description': 'The Bored Ape Yacht Club is a collection of 10,000 unique Bored Ape NFTs— unique digital collectibles living on the Ethereum blockchain. Your Bored Ape doubles as your Yacht Club membership card, and grants access to members-only benefits, the first of which is access to THE BATHROOM, a collaborative graffiti board. Future areas and perks can be unlocked by the community through roadmap activation. Visit www.BoredApeYachtClub.com for more details.', 'externalUrl': 'http://www.boredapeyachtclub.com/', 'twitterUsername': 'BoredApeYC', 'discordUrl': 'https://discord.gg/3P5K3dzgdB', 'lastIngestedAt': '2023-03-21T03:54:33.000Z'}}}\", metadata={'tokenId': '0x0000000000000000000000000000000000000000000000000000000000000000'}),\n",
" Document(page_content=\"{'contract': {'address': '0xbc4ca0eda7647a8ab7c2061c2e118a18a936f13d'}, 'id': {'tokenId': '0x0000000000000000000000000000000000000000000000000000000000000001', 'tokenMetadata': {'tokenType': 'ERC721'}}, 'title': '', 'description': '', 'tokenUri': {'gateway': 'https://alchemy.mypinata.cloud/ipfs/QmeSjSinHpPnmXmspMjwiXyN6zS4E9zccariGR3jxcaWtq/1', 'raw': 'ipfs://QmeSjSinHpPnmXmspMjwiXyN6zS4E9zccariGR3jxcaWtq/1'}, 'media': [{'gateway': 'https://nft-cdn.alchemy.com/eth-mainnet/65558a4d0c5b0c56fbc50bf03f55e3fa', 'thumbnail': 'https://res.cloudinary.com/alchemyapi/image/upload/thumbnailv2/eth-mainnet/65558a4d0c5b0c56fbc50bf03f55e3fa', 'raw': 'ipfs://QmPbxeGcXhYQQNgsC6a36dDyYUcHgMLnGKnF8pVFmGsvqi', 'format': 'png', 'bytes': 171425}], 'metadata': {'image': 'ipfs://QmPbxeGcXhYQQNgsC6a36dDyYUcHgMLnGKnF8pVFmGsvqi', 'attributes': [{'value': 'Grin', 'trait_type': 'Mouth'}, {'value': 'Vietnam Jacket', 'trait_type': 'Clothes'}, {'value': 'Orange', 'trait_type': 'Background'}, {'value': 'Blue Beams', 'trait_type': 'Eyes'}, {'value': 'Robot', 'trait_type': 'Fur'}]}, 'timeLastUpdated': '2023-04-24T04:37:37.738Z', 'contractMetadata': {'name': 'BoredApeYachtClub', 'symbol': 'BAYC', 'totalSupply': '10000', 'tokenType': 'ERC721', 'contractDeployer': '0xaba7161a7fb69c88e16ed9f455ce62b791ee4d03', 'deployedBlockNumber': 12287507, 'openSea': {'floorPrice': 68.16, 'collectionName': 'Bored Ape Yacht Club', 'safelistRequestStatus': 'verified', 'imageUrl': 'https://i.seadn.io/gae/Ju9CkWtV-1Okvf45wo8UctR-M9He2PjILP0oOvxE89AyiPPGtrR3gysu1Zgy0hjd2xKIgjJJtWIc0ybj4Vd7wv8t3pxDGHoJBzDB?w=500&auto=format', 'description': 'The Bored Ape Yacht Club is a collection of 10,000 unique Bored Ape NFTs— unique digital collectibles living on the Ethereum blockchain. Your Bored Ape doubles as your Yacht Club membership card, and grants access to members-only benefits, the first of which is access to THE BATHROOM, a collaborative graffiti board. Future areas and perks can be unlocked by the community through roadmap activation. Visit www.BoredApeYachtClub.com for more details.', 'externalUrl': 'http://www.boredapeyachtclub.com/', 'twitterUsername': 'BoredApeYC', 'discordUrl': 'https://discord.gg/3P5K3dzgdB', 'lastIngestedAt': '2023-03-21T03:54:33.000Z'}}}\", metadata={'tokenId': '0x0000000000000000000000000000000000000000000000000000000000000001'})]"
]
},
"execution_count": 24,
"metadata": {},
"output_type": "execute_result"
}
],
"outputs": [],
"source": [
"contractAddress = \"0xbc4ca0eda7647a8ab7c2061c2e118a18a936f13d\" # Bored Ape Yacht Club contract address\n",
"\n",
"blockchainType = BlockchainType.ETH_MAINNET #default value, optional parameter\n",
"\n",
"blockchainLoader = BlockchainDocumentLoader(contractAddress)\n",
"blockchainLoader = BlockchainDocumentLoader(contract_address=contractAddress,\n",
" api_key=alchemyApiKey)\n",
"\n",
"nfts = blockchainLoader.load()\n",
"\n",
@ -154,265 +104,22 @@
},
{
"cell_type": "code",
"execution_count": 36,
"execution_count": null,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"[Document(page_content=\"{'contract': {'address': '0x448676ffcd0adf2d85c1f0565e8dde6924a9a7d9'}, 'id': {'tokenId': '0x01', 'tokenMetadata': {'tokenType': 'ERC1155'}}, 'title': 'Wyatt Horton #0001', 'description': 'A sleepy capybara', 'tokenUri': {'gateway': 'https://storage.googleapis.com/minted-nfts/smoothstack/avatars/metadata/1.json', 'raw': 'https://storage.googleapis.com/minted-nfts/smoothstack/avatars/metadata/1.json'}, 'media': [{'gateway': 'https://nft-cdn.alchemy.com/matic-mainnet/9085e06ff9f6c9074de91801d1c72d26', 'thumbnail': 'https://res.cloudinary.com/alchemyapi/image/upload/thumbnailv2/matic-mainnet/9085e06ff9f6c9074de91801d1c72d26', 'raw': 'https://storage.googleapis.com/minted-nfts/smoothstack/avatars/images/1.png', 'format': 'png', 'bytes': 769622}], 'metadata': {'name': 'Wyatt Horton #0001', 'description': 'A sleepy capybara', 'image': 'https://storage.googleapis.com/minted-nfts/smoothstack/avatars/images/1.png', 'attributes': [{'value': 'Avatar', 'trait_type': 'Type'}, {'value': 'Animal', 'trait_type': 'Category'}, {'value': 'Capybara', 'trait_type': 'Class'}, {'value': 'Fall 2022', 'trait_type': 'Collection'}, {'value': 'Furry', 'trait_type': 'Feature'}]}, 'timeLastUpdated': '2023-04-20T14:38:24.947Z', 'contractMetadata': {'name': 'Smoothstack - Avatars', 'symbol': 'SMTH', 'tokenType': 'ERC1155', 'contractDeployer': '0x23075b2523c6563b06920a302a8be4f90ef6e974', 'deployedBlockNumber': 34752389, 'openSea': {'lastIngestedAt': '2023-04-17T20:59:42.000Z'}}}\", metadata={'tokenId': '0x01'}),\n",
" Document(page_content=\"{'contract': {'address': '0x448676ffcd0adf2d85c1f0565e8dde6924a9a7d9'}, 'id': {'tokenId': '0x02', 'tokenMetadata': {'tokenType': 'ERC1155'}}, 'title': 'Dylan Leisler #0002', 'description': 'A chipper cat with a big, red bowtie', 'tokenUri': {'gateway': 'https://storage.googleapis.com/minted-nfts/smoothstack/avatars/metadata/2.json', 'raw': 'https://storage.googleapis.com/minted-nfts/smoothstack/avatars/metadata/2.json'}, 'media': [{'gateway': 'https://nft-cdn.alchemy.com/matic-mainnet/67c3c7ccef44b32bf2ce758e8e73dbcd', 'thumbnail': 'https://res.cloudinary.com/alchemyapi/image/upload/thumbnailv2/matic-mainnet/67c3c7ccef44b32bf2ce758e8e73dbcd', 'raw': 'https://storage.googleapis.com/minted-nfts/smoothstack/avatars/images/2.png', 'format': 'png', 'bytes': 1187749}], 'metadata': {'name': 'Dylan Leisler #0002', 'description': 'A chipper cat with a big, red bowtie', 'image': 'https://storage.googleapis.com/minted-nfts/smoothstack/avatars/images/2.png', 'attributes': [{'value': 'Avatar', 'trait_type': 'Type'}, {'value': 'Animal', 'trait_type': 'Category'}, {'value': 'Cat', 'trait_type': 'Class'}, {'value': 'Fall 2022', 'trait_type': 'Collection'}, {'value': 'Red Bowtie', 'trait_type': 'Feature'}]}, 'timeLastUpdated': '2023-04-23T13:38:29.316Z', 'contractMetadata': {'name': 'Smoothstack - Avatars', 'symbol': 'SMTH', 'tokenType': 'ERC1155', 'contractDeployer': '0x23075b2523c6563b06920a302a8be4f90ef6e974', 'deployedBlockNumber': 34752389, 'openSea': {'lastIngestedAt': '2023-04-17T20:59:42.000Z'}}}\", metadata={'tokenId': '0x02'})]"
]
},
"execution_count": 36,
"metadata": {},
"output_type": "execute_result"
}
],
"outputs": [],
"source": [
"contractAddress = \"0x448676ffCd0aDf2D85C1f0565e8dde6924A9A7D9\" # Polygon Mainnet contract address\n",
"\n",
"blockchainType = BlockchainType.POLYGON_MAINNET \n",
"\n",
"blockchainLoader = BlockchainDocumentLoader(contractAddress, blockchainType)\n",
"blockchainLoader = BlockchainDocumentLoader(contract_address=contractAddress, \n",
" blockchainType=blockchainType, \n",
" api_key=alchemyApiKey)\n",
"\n",
"nfts = blockchainLoader.load()\n",
"\n",
"nfts[:2]"
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {},
"source": [
"## (Optional) Using the Blockchain Document Loader"
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "_PGkFfMCB8J3"
},
"source": [
"### Setup Splitter and Index"
]
},
{
"cell_type": "code",
"execution_count": 37,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Note: you may need to restart the kernel to use updated packages.\n"
]
}
],
"source": [
"%pip install sentence_transformers chromadb openai tiktoken -q"
]
},
{
"cell_type": "code",
"execution_count": 38,
"metadata": {
"id": "BwxxopOCCABh"
},
"outputs": [],
"source": [
"from langchain.indexes import VectorstoreIndexCreator\n",
"from langchain.embeddings import HuggingFaceEmbeddings\n",
"from langchain.text_splitter import RecursiveCharacterTextSplitter"
]
},
{
"cell_type": "code",
"execution_count": 39,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "JE_myAulCDSZ",
"outputId": "99e16b6a-03b4-4e67-d4b4-9dd611a866ef"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"NUMBER OF DOCUMENTS: 424\n"
]
}
],
"source": [
"text_splitter = RecursiveCharacterTextSplitter(chunk_size=400, chunk_overlap=0)\n",
"\n",
"docs = text_splitter.split_documents(nfts)\n",
"print(\"NUMBER OF DOCUMENTS: \", len(docs))"
]
},
{
"cell_type": "code",
"execution_count": 40,
"metadata": {
"id": "d83yFuAuCKQS"
},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"Using embedded DuckDB without persistence: data will be transient\n"
]
}
],
"source": [
"index = VectorstoreIndexCreator(\n",
" embedding=HuggingFaceEmbeddings(),\n",
" text_splitter=text_splitter).from_loaders([blockchainLoader])"
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "y0VfObeXDEXB"
},
"source": [
"## Setup Models and Chains"
]
},
{
"cell_type": "code",
"execution_count": 42,
"metadata": {},
"outputs": [],
"source": [
"openAiKey = \"put OpenAI key here\"\n",
"os.environ[\"OPENAI_API_KEY\"] = openAiKey"
]
},
{
"cell_type": "code",
"execution_count": 31,
"metadata": {
"id": "hiNjDzP9C4pA"
},
"outputs": [],
"source": [
"from langchain.chains import RetrievalQA\n",
"from langchain.llms import OpenAI"
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "u-xDlKPaC_xg"
},
"source": [
"### Retrieval Chain"
]
},
{
"cell_type": "code",
"execution_count": 43,
"metadata": {
"id": "BqP00JovC9R4"
},
"outputs": [],
"source": [
"llmOpenAI = OpenAI()\n",
"\n",
"chainQA = RetrievalQA.from_chain_type(llm=llmOpenAI, \n",
" chain_type=\"map_reduce\",\n",
" retriever=index.vectorstore.as_retriever(), \n",
" verbose=True,\n",
" input_key=\"question\")"
]
},
{
"cell_type": "code",
"execution_count": 44,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 122
},
"id": "2Y3cVVKZDVNq",
"outputId": "dfeea416-5193-47cf-e9dc-c17a5c1cd780"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n",
"\n",
"\u001b[1m> Entering new RetrievalQA chain...\u001b[0m\n",
"\n",
"\u001b[1m> Finished chain.\u001b[0m\n"
]
},
{
"data": {
"text/plain": [
"' Popular attributes include \"Avatar\" (Type), \"Character\" (Category), and \"Human\" or \"Wizard\" (Class).'"
]
},
"execution_count": 44,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"chainQA.run(\"What are some of the popular attributes?\")"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 122
},
"id": "7o6ArPo9DXbz",
"outputId": "b1f8ad43-27c7-4cdb-95a7-8c8bd6381c5a"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n",
"\n",
"\u001b[1m> Entering new RetrievalQA chain...\u001b[0m\n",
"\n",
"\u001b[1m> Finished chain.\u001b[0m\n"
]
},
{
"data": {
"application/vnd.google.colaboratory.intrinsic+json": {
"type": "string"
},
"text/plain": [
"' There are 10,000 unique Bored Ape NFTs.'"
]
},
"execution_count": 32,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"chainQA.run(\"How many NFTs are there?\")"
]
}
],
"metadata": {

@ -26,10 +26,14 @@ class BlockchainDocumentLoader(BaseLoader):
The Loader uses the Alchemy API to interact with the blockchain.
The API returns 100 NFTs per request and can be paginated using the
startToken parameter.
ALCHEMY_API_KEY environment variable must be set to use this loader.
Future versions of this loader can:
- Support additional Alchemy APIs (e.g. getTransactions, etc.)
- Support additional blockain APIs (e.g. Infura, Opensea, etc.)
"""
def __init__(
@ -37,7 +41,7 @@ class BlockchainDocumentLoader(BaseLoader):
contract_address: str,
blockchainType: BlockchainType = BlockchainType.ETH_MAINNET,
api_key: str = "docs-demo",
startToken: int = 0,
startToken: str = "",
):
self.contract_address = contract_address
self.blockchainType = blockchainType.value
@ -75,6 +79,10 @@ class BlockchainDocumentLoader(BaseLoader):
for item in items:
content = str(item)
tokenId = item["id"]["tokenId"]
metadata = {"tokenId": tokenId}
metadata = {
"source": self.contract_address,
"blockchain": self.blockchainType,
"tokenId": tokenId,
}
result.append(Document(page_content=content, metadata=metadata))
return result

@ -26,7 +26,7 @@ def test_get_nfts_with_pagination() -> None:
contract_address = (
"0x1a92f7381b9f03921564a437210bb9396471050c" # CoolCats contract address
)
startToken = 20
startToken = "20"
result = BlockchainDocumentLoader(
contract_address,

Loading…
Cancel
Save