mirror of
https://github.com/hwchase17/langchain
synced 2024-11-10 01:10:59 +00:00
docs: add retriever integrations template (#24836)
This commit is contained in:
ccurme
GitHub
parent
4da3d4b18e
commit
30f18c7b02
27
docs/docs/integrations/retrievers/index.mdx
Normal file
27
docs/docs/integrations/retrievers/index.mdx
Normal file
@ -0,0 +1,27 @@
|
||||
---
|
||||
sidebar_position: 0
|
||||
sidebar_class_name: hidden
|
||||
---
|
||||
|
||||
# Retrievers
|
||||
|
||||
A **retriever** is an interface that returns documents given an unstructured query.
|
||||
It is more general than a vector store.
|
||||
A retriever does not need to be able to store documents, only to return (or retrieve) them.
|
||||
Retrievers can be created from vector stores, but are also broad enough to include [Wikipedia search](/docs/integrations/retrievers/wikipedia/) and [Amazon Kendra](/docs/integrations/retrievers/amazon_kendra_retriever/).
|
||||
|
||||
Retrievers accept a string query as input and return a list of Document's as output.
|
||||
|
||||
For specifics on how to use retrievers, see the [relevant how-to guides here](/docs/how_to/#retrievers).
|
||||
|
||||
This table lists common retrievers.
|
||||
|
||||
|
||||
| Retriever | Namespace | Native async | Local |
|
||||
|-----------|-----------|---------------|------|
|
||||
| [AmazonKnowledgeBasesRetriever](https://api.python.langchain.com/en/latest/retrievers/langchain_aws.retrievers.bedrock.AmazonKnowledgeBasesRetriever.html) | langchain_aws.retrievers | ❌ | ❌ |
|
||||
| [AzureAISearchRetriever](https://api.python.langchain.com/en/latest/retrievers/langchain_community.retrievers.azure_ai_search.AzureAISearchRetriever.html) | langchain_community.retrievers | ✅ | ❌ |
|
||||
| [ElasticsearchRetriever](https://api.python.langchain.com/en/latest/retrievers/langchain_elasticsearch.retrievers.ElasticsearchRetriever.html) | langchain_elasticsearch | ❌ | ❌ |
|
||||
| [MilvusCollectionHybridSearchRetriever](https://api.python.langchain.com/en/latest/retrievers/langchain_milvus.retrievers.milvus_hybrid_search.MilvusCollectionHybridSearchRetriever.html) | langchain_milvus | ❌ | ❌ |
|
||||
| [TavilySearchAPIRetriever](https://api.python.langchain.com/en/latest/retrievers/langchain_community.retrievers.tavily_search_api.TavilySearchAPIRetriever.html) | langchain_community.retrievers | ❌ | ❌ |
|
||||
| [VertexAISearchRetriever](https://api.python.langchain.com/en/latest/vertex_ai_search/langchain_google_community.vertex_ai_search.VertexAISearchRetriever.html) | langchain_google_community.vertex_ai_search | ❌ | ❌ |
|
||||
@ -4,20 +4,70 @@
|
||||
"cell_type": "markdown",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"# Tavily Search API\n",
|
||||
"---\n",
|
||||
"sidebar_label: TavilySearchAPI\n",
|
||||
"---"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"# TavilySearchAPIRetriever\n",
|
||||
"\n",
|
||||
"## Overview\n",
|
||||
">[Tavily's Search API](https://tavily.com) is a search engine built specifically for AI agents (LLMs), delivering real-time, accurate, and factual results at speed.\n",
|
||||
"\n",
|
||||
"We can use this as a [retriever](/docs/how_to#retrievers). It will show functionality specific to this integration. After going through, it may be useful to explore [relevant use-case pages](/docs/how_to#qa-with-rag) to learn how to use this vectorstore as part of a larger chain.\n",
|
||||
"\n",
|
||||
"## Setup\n",
|
||||
"### Integration details\n",
|
||||
"\n",
|
||||
"The integration lives in the `langchain-community` package. We also need to install the `tavily-python` package itself.\n",
|
||||
"| Retriever | Namespace | Native async | Local |\n",
|
||||
"| :--- | :--- | :---: | :---: |\n",
|
||||
"[TavilySearchAPIRetriever](https://api.python.langchain.com/en/latest/retrievers/langchain_community.retrievers.tavily_search_api.TavilySearchAPIRetriever.html) | langchain_community.retrievers | ❌ | ❌ |\n",
|
||||
"\n",
|
||||
"```bash\n",
|
||||
"pip install -U langchain-community tavily-python\n",
|
||||
"```\n",
|
||||
"## Setup"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"If you want to get automated tracing from runs of individual tools, you can also set your [LangSmith](https://docs.smith.langchain.com/) API key by uncommenting below:"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": null,
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"# os.environ[\"LANGSMITH_API_KEY\"] = getpass.getpass(\"Enter your LangSmith API key: \")\n",
|
||||
"# os.environ[\"LANGSMITH_TRACING\"] = \"true\""
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"### Installation\n",
|
||||
"\n",
|
||||
"The integration lives in the `langchain-community` package. We also need to install the `tavily-python` package itself."
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": null,
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"%pip install -qU langchain-community tavily-python"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"We also need to set our Tavily API key."
|
||||
]
|
||||
},
|
||||
@ -37,17 +87,20 @@
|
||||
"cell_type": "markdown",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"It's also helpful (but not needed) to set up [LangSmith](https://smith.langchain.com/) for best-in-class observability"
|
||||
"## Instantiation\n",
|
||||
"\n",
|
||||
"Now we can instantiate our retriever:"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": null,
|
||||
"execution_count": 1,
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"# os.environ[\"LANGCHAIN_TRACING_V2\"] = \"true\"\n",
|
||||
"# os.environ[\"LANGCHAIN_API_KEY\"] = getpass.getpass()"
|
||||
"from langchain_community.retrievers import TavilySearchAPIRetriever\n",
|
||||
"\n",
|
||||
"retriever = TavilySearchAPIRetriever(k=3)"
|
||||
]
|
||||
},
|
||||
{
|
||||
@ -59,42 +112,40 @@
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 8,
|
||||
"execution_count": 2,
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
"data": {
|
||||
"text/plain": [
|
||||
"[Document(page_content='Trending topics\\nTrending topics\\nThe Legend of Zelda™: Breath of the Wild\\nSelect a product\\nThe Legend of Zelda™: Breath of the Wild\\nThe Legend of Zelda™: Breath of the Wild\\nThe Legend of Zelda™: Breath of the Wild and The Legend of Zelda™: Breath of the Wild Expansion Pass Bundle\\nThis item will be sent to your system automatically after purchase or Nintendo Switch Game Voucher redemption. The Legend of Zelda: Breath of the Wild Expansion Pass\\nMore like this\\nSuper Mario Odyssey™\\nThe Legend of Zelda™: Tears of the Kingdom\\nMario + Rabbids® Kingdom Battle\\nThe Legend of Zelda™: Link’s Awakening\\nHollow Knight\\nThe Legend of Zelda™: Skyward Sword HD\\nStarlink: Battle for Atlas™ Digital Edition\\nDRAGON QUEST BUILDERS™ 2\\nDragon Quest Builders™\\nWARNING: If you have epilepsy or have had seizures or other unusual reactions to flashing lights or patterns, consult a doctor before playing video games. Saddle up with a herd of horse-filled games!\\nESRB rating\\nSupported play modes\\nTV\\nTabletop\\nHandheld\\nProduct information\\nRelease date\\nNo. of players\\nGenre\\nPublisher\\nESRB rating\\nSupported play modes\\nGame file size\\nSupported languages\\nPlay online, access classic NES™ and Super NES™ games, and more with a Nintendo Switch Online membership.\\n Two Game Boy games are now available for Nintendo Switch Online members\\n02/01/23\\nNintendo Switch Online member exclusive: Save on two digital games\\n09/13/22\\nOut of the Shadows … the Legend of Zelda: About Nintendo\\nShop\\nMy Nintendo Store orders\\nSupport\\nParents\\nCommunity\\nPrivacy\\n© Nintendo.', metadata={'title': 'The Legend of Zelda™: Breath of the Wild - Nintendo', 'source': 'https://www.nintendo.com/us/store/products/the-legend-of-zelda-breath-of-the-wild-switch/', 'score': 0.97451, 'images': None}),\n",
|
||||
" Document(page_content='The Legend of Zelda: Breath of the Wild is a masterpiece of open-world design and exploration, released on March 3, 2017 for Nintendo Switch. Find out the latest news, reviews, guides, videos, and more for this award-winning game on IGN.', metadata={'title': 'The Legend of Zelda: Breath of the Wild - IGN', 'source': 'https://www.ign.com/games/the-legend-of-zelda-breath-of-the-wild', 'score': 0.94496, 'images': None}),\n",
|
||||
" Document(page_content='Reviewers also commented on the unexpected permutations of interactions between Link, villagers, pets, and enemies,[129][130][131] many of which were shared widely on social media.[132] A tribute to former Nintendo president Satoru Iwata, who died during development, also attracted praise.[129][134]\\nJim Sterling was more critical than most, giving Breath of the Wild a 7/10 score, criticizing the difficulty, weapon durability, and level design, but praising the open world and variety of content.[135] Other criticism focused on the unstable frame rate and the low resolution of 900p;[136] updates addressed some of these problems.[137][138]\\nSales\\nBreath of the Wild broke sales records for a Nintendo launch game in multiple regions.[139][140] In Japan, the Switch and Wii U versions sold a combined 230,000 copies in the first week of release, with the Switch version becoming the top-selling game released that week.[141] Nintendo reported that Breath of the Wild sold more than one million copies in the US that month—925,000 of which were for Switch, outselling the Switch itself.[145][146][147][148] Nintendo president Tatsumi Kimishima said that the attach rate on the Switch was \"unprecedented\".[149] Breath of the Wild had sold 31.15 million copies on the Switch by September 2023 and 1.70 million copies on the Wii U by December 2020.[150][151]\\nAwards\\nFollowing its demonstration at E3 2016, Breath of the Wild received several accolades from the Game Critics Awards[152] and from publications such as IGN and Destructoid.[153][154] It was listed among the best games at E3 by Eurogamer,[81] The game, he continued, would challenge the series\\' conventions, such as the requirement that players complete dungeons in a set order.[2][73] The next year, Nintendo introduced the game\\'s high-definition, cel-shaded visual style with in-game footage at its E3 press event.[74][75] Once planned for release in 2015, the game was delayed early in the year and did not show at that year\\'s E3.[76][77] Zelda series creator Shigeru Miyamoto reaffirmed that the game would still release for the Wii U despite the development of Nintendo\\'s next console, the Nintendo Switch.[78] The Switch version also has higher-quality environmental sounds.[53][54] Certain ideas that were planned for the game, like flying and underground dungeons were not implemented due to the Wii U’s limitations; they would eventually resurface in the game\\'s sequel.[55] Aonuma stated that the art design was inspired by gouache and en plein air art to help identify the vast world.[56] Takizawa has also cited the Jōmon period as an inspiration for the ancient Sheikah technology and architecture that is found in the game, due to the mystery surrounding the period.[57] Journalists commented on unexpected interactions between game elements,[129][130][131] with serendipitous moments proving popular on social media.[132] Chris Plante of The Verge predicted that whereas prior open-world games tended to feature prescribed challenges, Zelda would influence a new generation of games with open-ended problem-solving.[132] Digital Trends wrote that the game\\'s level of experimentation allowed players to interact with and exploit the environment in creative ways, resulting in various \"tricks\" still discovered years after release.[127]\\nReviewers lauded the sense of detail and immersion.[133][129] Kotaku recommended turning off UI elements in praise of the indirect cues that contextually indicate the same information, such as Link shivering in the cold or waypoints appearing when using the scope.[133]', metadata={'title': 'The Legend of Zelda: Breath of the Wild - Wikipedia', 'source': 'https://en.wikipedia.org/wiki/The_Legend_of_Zelda:_Breath_of_the_Wild', 'score': 0.93348, 'images': None})]"
|
||||
"[Document(metadata={'title': 'The Legend of Zelda: Breath of the Wild - Nintendo Switch Wiki', 'source': 'https://nintendo-switch.fandom.com/wiki/The_Legend_of_Zelda:_Breath_of_the_Wild', 'score': 0.9961155, 'images': []}, page_content='The Legend of Zelda: Breath of the Wild is an open world action-adventure game published by Nintendo for the Wii U and as a launch title for the Nintendo Switch, and was released worldwide on March 3, 2017. It is the nineteenth installment of the The Legend of Zelda series and the first to be developed with a HD resolution. The game features a gigantic open world, with the player being able to ...'),\n",
|
||||
" Document(metadata={'title': 'The Legend of Zelda: Breath of the Wild - Zelda Wiki', 'source': 'https://zelda.fandom.com/wiki/The_Legend_of_Zelda:_Breath_of_the_Wild', 'score': 0.9804313, 'images': []}, page_content='[]\\nReferences\\nThe Legend of Zelda \\xa0·\\nThe Adventure of Link \\xa0·\\nA Link to the Past (& Four Swords) \\xa0·\\nLink\\'s Awakening (DX; Nintendo Switch) \\xa0·\\nOcarina of Time (Master Quest; 3D) \\xa0·\\nMajora\\'s Mask (3D) \\xa0·\\nOracle of Ages \\xa0·\\nOracle of Seasons \\xa0·\\nFour Swords (Anniversary Edition) \\xa0·\\nThe Wind Waker (HD) \\xa0·\\nFour Swords Adventures \\xa0·\\nThe Minish Cap \\xa0·\\nTwilight Princess (HD) \\xa0·\\nPhantom Hourglass \\xa0·\\nSpirit Tracks \\xa0·\\nSkyward Sword (HD) \\xa0·\\nA Link Between Worlds \\xa0·\\nTri Force Heroes \\xa0·\\nBreath of the Wild \\xa0·\\nTears of the Kingdom\\nZelda (Game & Watch) \\xa0·\\nThe Legend of Zelda Game Watch \\xa0·\\nLink\\'s Crossbow Training \\xa0·\\nMy Nintendo Picross: Twilight Princess \\xa0·\\nCadence of Hyrule \\xa0·\\nGame & Watch: The Legend of Zelda\\nCD-i Games\\n Listings[]\\nCharacters[]\\nBosses[]\\nEnemies[]\\nDungeons[]\\nLocations[]\\nItems[]\\nTranslations[]\\nCredits[]\\nReception[]\\nSales[]\\nEiji Aonuma and Hidemaro Fujibayashi accepting the \"Game of the Year\" award for Breath of the Wild at The Game Awards 2017\\nBreath of the Wild was estimated to have sold approximately 1.3 million copies in its first three weeks and around 89% of Switch owners were estimated to have also purchased the game.[52] Sales of the game have remained strong and as of June 30, 2022, the Switch version has sold 27.14 million copies worldwide while the Wii U version has sold 1.69 million copies worldwide as of December 31, 2019,[53][54] giving Breath of the Wild a cumulative total of 28.83 million copies sold.\\n It also earned a Metacritic score of 97 from more than 100 critics, placing it among the highest-rated games of all time.[59][60] Notably, the game received the most perfect review scores for any game listed on Metacritic up to that point.[61]\\nIn 2022, Breath of the Wild was chosen as the best Legend of Zelda game of all time in their \"Top 10 Best Zelda Games\" list countdown; but was then placed as the \"second\" best Zelda game in their new revamped version of their \"Top 10 Best Zelda Games\" list in 2023, right behind it\\'s successor Tears of Video Game Canon ranks Breath of the Wild as one of the best video games of all time.[74] Metacritic ranked Breath of the Wild as the single best game of the 2010s.[75]\\nFan Reception[]\\nWatchMojo placed Breath of the Wild at the #2 spot in their \"Top 10 Legend of Zelda Games of All Time\" list countdown, right behind Ocarina of Time.[76] The Faces of Evil \\xa0·\\nThe Wand of Gamelon \\xa0·\\nZelda\\'s Adventure\\nHyrule Warriors Series\\nHyrule Warriors (Legends; Definitive Edition) \\xa0·\\nHyrule Warriors: Age of Calamity\\nSatellaview Games\\nBS The Legend of Zelda \\xa0·\\nAncient Stone Tablets\\nTingle Series\\nFreshly-Picked Tingle\\'s Rosy Rupeeland \\xa0·\\nTingle\\'s Balloon Fight DS \\xa0·\\n'),\n",
|
||||
" Document(metadata={'title': 'The Legend of Zelda: Breath of the Wild - Zelda Wiki', 'source': 'https://zeldawiki.wiki/wiki/The_Legend_of_Zelda:_Breath_of_the_Wild', 'score': 0.9627432, 'images': []}, page_content='The Legend of Zelda\\xa0•\\nThe Adventure of Link\\xa0•\\nA Link to the Past (& Four Swords)\\xa0•\\nLink\\'s Awakening (DX; Nintendo Switch)\\xa0•\\nOcarina of Time (Master Quest; 3D)\\xa0•\\nMajora\\'s Mask (3D)\\xa0•\\nOracle of Ages\\xa0•\\nOracle of Seasons\\xa0•\\nFour Swords (Anniversary Edition)\\xa0•\\nThe Wind Waker (HD)\\xa0•\\nFour Swords Adventures\\xa0•\\nThe Minish Cap\\xa0•\\nTwilight Princess (HD)\\xa0•\\nPhantom Hourglass\\xa0•\\nSpirit Tracks\\xa0•\\nSkyward Sword (HD)\\xa0•\\nA Link Between Worlds\\xa0•\\nTri Force Heroes\\xa0•\\nBreath of the Wild\\xa0•\\nTears of the Kingdom\\nZelda (Game & Watch)\\xa0•\\nThe Legend of Zelda Game Watch\\xa0•\\nHeroes of Hyrule\\xa0•\\nLink\\'s Crossbow Training\\xa0•\\nMy Nintendo Picross: Twilight Princess\\xa0•\\nCadence of Hyrule\\xa0•\\nVermin\\nThe Faces of Evil\\xa0•\\nThe Wand of Gamelon\\xa0•\\nZelda\\'s Adventure\\nHyrule Warriors (Legends; Definitive Edition)\\xa0•\\nHyrule Warriors: Age of Calamity\\nBS The Legend of Zelda\\xa0•\\nAncient Stone Tablets\\nFreshly-Picked Tingle\\'s Rosy Rupeeland\\xa0•\\nTingle\\'s Balloon Fight DS\\xa0•\\nToo Much Tingle Pack\\xa0•\\nRipened Tingle\\'s Balloon Trip of Love\\nSoulcalibur II\\xa0•\\nWarioWare Series\\xa0•\\nCaptain Rainbow\\xa0•\\nNintendo Land\\xa0•\\nScribblenauts Unlimited\\xa0•\\nMario Kart 8\\xa0•\\nSplatoon 3\\nSuper Smash Bros (Series)\\nSuper Smash Bros.\\xa0•\\nSuper Smash Bros. Melee\\xa0•\\nSuper Smash Bros. Brawl\\xa0•\\nSuper Smash Bros. for Nintendo 3DS / Wii U\\xa0•\\n It also earned a Metacritic score of 97 from more than 100 critics, placing it among the highest-rated games of all time.[60][61] Notably, the game received the most perfect review scores for any game listed on Metacritic up to that point.[62]\\nAwards\\nThroughout 2016, Breath of the Wild won several awards as a highly anticipated game, including IGN\\'s and Destructoid\\'s Best of E3,[63][64] at the Game Critic Awards 2016,[65] and at The Game Awards 2016.[66] Following its release, Breath of the Wild received the title of \"Game of the Year\" from the Japan Game Awards 2017,[67] the Golden Joystick Awards 2017,<ref\"Our final award is for the Ultimate Game of the Year. Official website(s)\\nOfficial website(s)\\nCanonicity\\nCanonicity\\nCanon[citation needed]\\nPredecessor\\nPredecessor\\nTri Force Heroes\\nSuccessor\\nSuccessor\\nTears of the Kingdom\\nThe Legend of Zelda: Breath of the Wild guide at StrategyWiki\\nBreath of the Wild Guide at Zelda Universe\\nThe Legend of Zelda: Breath of the Wild is the nineteenth main installment of The Legend of Zelda series. Listings\\nCharacters\\nBosses\\nEnemies\\nDungeons\\nLocations\\nItems\\nTranslations\\nCredits\\nReception\\nSales\\nBreath of the Wild was estimated to have sold approximately 1.3 million copies in its first three weeks and around 89% of Switch owners were estimated to have also purchased the game.[53] Sales of the game have remained strong and as of September 30, 2023, the Switch version has sold 31.15 million copies worldwide while the Wii U version has sold 1.7 million copies worldwide as of December 31, 2021,[54][55] giving Breath of the Wild a cumulative total of 32.85 million copies sold.\\n The Legend of Zelda: Breath of the Wild\\nThe Legend of Zelda: Breath of the Wild\\nThe Legend of Zelda: Breath of the Wild\\nDeveloper(s)\\nDeveloper(s)\\nPublisher(s)\\nPublisher(s)\\nNintendo\\nDesigner(s)\\nDesigner(s)\\n')]"
|
||||
]
|
||||
},
|
||||
"execution_count": 8,
|
||||
"execution_count": 2,
|
||||
"metadata": {},
|
||||
"output_type": "execute_result"
|
||||
}
|
||||
],
|
||||
"source": [
|
||||
"from langchain_community.retrievers import TavilySearchAPIRetriever\n",
|
||||
"query = \"what year was breath of the wild released?\"\n",
|
||||
"\n",
|
||||
"retriever = TavilySearchAPIRetriever(k=3)\n",
|
||||
"\n",
|
||||
"retriever.invoke(\"what year was breath of the wild released?\")"
|
||||
"retriever.invoke(query)"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"## Chaining\n",
|
||||
"## Use within a chain\n",
|
||||
"\n",
|
||||
"We can easily combine this retriever in to a chain."
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 11,
|
||||
"execution_count": 3,
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
@ -110,40 +161,50 @@
|
||||
"\n",
|
||||
"Question: {question}\"\"\"\n",
|
||||
")\n",
|
||||
"\n",
|
||||
"llm = ChatOpenAI(model=\"gpt-3.5-turbo-0125\")\n",
|
||||
"\n",
|
||||
"\n",
|
||||
"def format_docs(docs):\n",
|
||||
" return \"\\n\\n\".join(doc.page_content for doc in docs)\n",
|
||||
"\n",
|
||||
"\n",
|
||||
"chain = (\n",
|
||||
" RunnablePassthrough.assign(context=(lambda x: x[\"question\"]) | retriever)\n",
|
||||
" {\"context\": retriever | format_docs, \"question\": RunnablePassthrough()}\n",
|
||||
" | prompt\n",
|
||||
" | ChatOpenAI(model=\"gpt-4-1106-preview\")\n",
|
||||
" | llm\n",
|
||||
" | StrOutputParser()\n",
|
||||
")"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 13,
|
||||
"execution_count": 4,
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
"data": {
|
||||
"text/plain": [
|
||||
"'As of the end of 2020, \"The Legend of Zelda: Breath of the Wild\" sold over 21.45 million copies worldwide.'"
|
||||
"'As of August 2020, The Legend of Zelda: Breath of the Wild had sold over 20.1 million copies worldwide on Nintendo Switch and Wii U.'"
|
||||
]
|
||||
},
|
||||
"execution_count": 13,
|
||||
"execution_count": 4,
|
||||
"metadata": {},
|
||||
"output_type": "execute_result"
|
||||
}
|
||||
],
|
||||
"source": [
|
||||
"chain.invoke({\"question\": \"how many units did bretch of the wild sell in 2020\"})"
|
||||
"chain.invoke(\"how many units did bretch of the wild sell in 2020\")"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": null,
|
||||
"cell_type": "markdown",
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": []
|
||||
"source": [
|
||||
"## API reference\n",
|
||||
"\n",
|
||||
"For detailed documentation of all `TavilySearchAPIRetriever` features and configurations head to the [API reference](https://api.python.langchain.com/en/latest/retrievers/langchain_community.retrievers.tavily_search_api.TavilySearchAPIRetriever.html)."
|
||||
]
|
||||
}
|
||||
],
|
||||
"metadata": {
|
||||
@ -162,7 +223,7 @@
|
||||
"name": "python",
|
||||
"nbconvert_exporter": "python",
|
||||
"pygments_lexer": "ipython3",
|
||||
"version": "3.10.12"
|
||||
"version": "3.10.4"
|
||||
}
|
||||
},
|
||||
"nbformat": 4,
|
||||
|
||||
@ -0,0 +1,217 @@
|
||||
{
|
||||
"cells": [
|
||||
{
|
||||
"cell_type": "raw",
|
||||
"id": "afaf8039",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"---\n",
|
||||
"sidebar_label: __ModuleName__\n",
|
||||
"---"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"id": "e49f1e0d",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"# __ModuleName__Retriever\n",
|
||||
"\n",
|
||||
"## Overview\n",
|
||||
"- TODO: Make sure API reference link is correct.\n",
|
||||
"\n",
|
||||
"This will help you getting started with the __ModuleName__ [retriever](/docs/concepts/#retrievers). For detailed documentation of all __ModuleName__Retriever features and configurations head to the [API reference](https://api.python.langchain.com/en/latest/retrievers/__module_name__.retrievers.__ModuleName__.__ModuleName__Retriever.html).\n",
|
||||
"\n",
|
||||
"### Integration details\n",
|
||||
"\n",
|
||||
"| Retriever | Namespace | Native async | Local |\n",
|
||||
"| :--- | :--- | :---: | :---: |\n",
|
||||
"[__ModuleName__Retriever](https://api.python.langchain.com/en/latest/retrievers/__package_name__.retrievers.__module_name__.__ModuleName__Retriever.html) | __package_name__.retrievers | ❌ | ❌ |\n",
|
||||
"\n",
|
||||
"\n",
|
||||
"## Setup\n",
|
||||
"\n",
|
||||
"- TODO: Update with relevant info."
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"id": "72ee0c4b-9764-423a-9dbf-95129e185210",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"If you want to get automated tracing from runs of individual tools, you can also set your [LangSmith](https://docs.smith.langchain.com/) API key by uncommenting below:"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": null,
|
||||
"id": "a15d341e-3e26-4ca3-830b-5aab30ed66de",
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"# os.environ[\"LANGSMITH_API_KEY\"] = getpass.getpass(\"Enter your LangSmith API key: \")\n",
|
||||
"# os.environ[\"LANGSMITH_TRACING\"] = \"true\""
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"id": "0730d6a1-c893-4840-9817-5e5251676d5d",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"### Installation\n",
|
||||
"\n",
|
||||
"This retriever lives in the `__package_name__` package:"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": null,
|
||||
"id": "652d6238-1f87-422a-b135-f5abbb8652fc",
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"%pip install -qU __package_name__"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"id": "a38cde65-254d-4219-a441-068766c0d4b5",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"## Instantiation\n",
|
||||
"\n",
|
||||
"Now we can instantiate our retriever:\n",
|
||||
"\n",
|
||||
"- TODO: Update model instantiation with relevant params."
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": null,
|
||||
"id": "70cc8e65-2a02-408a-bbc6-8ef649057d82",
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"from __module_name__ import __ModuleName__Retriever\n",
|
||||
"\n",
|
||||
"retriever = __ModuleName__Retriever(\n",
|
||||
" # ...\n",
|
||||
")"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"id": "5c5f2839-4020-424e-9fc9-07777eede442",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"## Usage"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": null,
|
||||
"id": "51a60dbe-9f2e-4e04-bb62-23968f17164a",
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"query = \"...\"\n",
|
||||
"\n",
|
||||
"retriever.invoke(query)"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"id": "dfe8aad4-8626-4330-98a9-7ea1ca5d2e0e",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"## Use within a chain"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": null,
|
||||
"id": "23e11cc9-abd6-4855-a7eb-799f45ca01ae",
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"from langchain_core.output_parsers import StrOutputParser\n",
|
||||
"from langchain_core.prompts import ChatPromptTemplate\n",
|
||||
"from langchain_core.runnables import RunnablePassthrough\n",
|
||||
"from langchain_openai import ChatOpenAI\n",
|
||||
"\n",
|
||||
"prompt = ChatPromptTemplate.from_template(\n",
|
||||
" \"\"\"Answer the question based only on the context provided.\n",
|
||||
"\n",
|
||||
"Context: {context}\n",
|
||||
"\n",
|
||||
"Question: {question}\"\"\"\n",
|
||||
")\n",
|
||||
"\n",
|
||||
"llm = ChatOpenAI(model=\"gpt-3.5-turbo-0125\")\n",
|
||||
"\n",
|
||||
"\n",
|
||||
"def format_docs(docs):\n",
|
||||
" return \"\\n\\n\".join(doc.page_content for doc in docs)\n",
|
||||
"\n",
|
||||
"\n",
|
||||
"chain = (\n",
|
||||
" {\"context\": retriever | format_docs, \"question\": RunnablePassthrough()}\n",
|
||||
" | prompt\n",
|
||||
" | llm\n",
|
||||
" | StrOutputParser()\n",
|
||||
")"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": null,
|
||||
"id": "d47c37dd-5c11-416c-a3b6-bec413cd70e8",
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"chain.invoke(\"...\")"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"id": "d1ee55bc-ffc8-4cfa-801c-993953a08cfd",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"## TODO: Any functionality or considerations specific to this retriever\n",
|
||||
"\n",
|
||||
"Fill in or delete if not relevant."
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "markdown",
|
||||
"id": "3a5bb5ca-c3ae-4a58-be67-2cd18574b9a3",
|
||||
"metadata": {},
|
||||
"source": [
|
||||
"## API reference\n",
|
||||
"\n",
|
||||
"For detailed documentation of all __ModuleName__Retriever features and configurations head to the [API reference](https://api.python.langchain.com/en/latest/retrievers/__module_name__.retrievers.__ModuleName__.__ModuleName__Retriever.html)."
|
||||
]
|
||||
}
|
||||
],
|
||||
"metadata": {
|
||||
"kernelspec": {
|
||||
"display_name": "Python 3 (ipykernel)",
|
||||
"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": "3.10.4"
|
||||
}
|
||||
},
|
||||
"nbformat": 4,
|
||||
"nbformat_minor": 5
|
||||
}
|
||||
@ -0,0 +1,89 @@
|
||||
"""__ModuleName__ retrievers."""
|
||||
|
||||
from typing import List
|
||||
|
||||
from langchain_core.documents import Document
|
||||
from langchain_core.retrievers import BaseRetriever
|
||||
|
||||
|
||||
class __ModuleName__Retriever(BaseRetriever):
|
||||
# TODO: Replace all TODOs in docstring. See example docstring:
|
||||
# https://github.com/langchain-ai/langchain/blob/master/libs/community/langchain_community/retrievers/tavily_search_api.py#L17
|
||||
"""__ModuleName__ retriever.
|
||||
|
||||
# TODO: Replace with relevant packages, env vars, etc.
|
||||
Setup:
|
||||
Install ``__package_name__`` and set environment variable ``__MODULE_NAME___API_KEY``.
|
||||
|
||||
.. code-block:: bash
|
||||
|
||||
pip install -U __package_name__
|
||||
export __MODULE_NAME___API_KEY="your-api-key"
|
||||
|
||||
# TODO: Populate with relevant params.
|
||||
Key init args:
|
||||
arg 1: type
|
||||
description
|
||||
arg 2: type
|
||||
description
|
||||
|
||||
# TODO: Replace with relevant init params.
|
||||
Instantiate:
|
||||
.. code-block:: python
|
||||
|
||||
from __package_name__ import __ModuleName__Retriever
|
||||
|
||||
retriever = __ModuleName__Retriever(
|
||||
# ...
|
||||
)
|
||||
|
||||
Usage:
|
||||
.. code-block:: python
|
||||
|
||||
query = "..."
|
||||
|
||||
retriever.invoke(query)
|
||||
|
||||
.. code-block:: python
|
||||
|
||||
# TODO: Example output.
|
||||
|
||||
Use within a chain:
|
||||
.. code-block:: python
|
||||
|
||||
from langchain_core.output_parsers import StrOutputParser
|
||||
from langchain_core.prompts import ChatPromptTemplate
|
||||
from langchain_core.runnables import RunnablePassthrough
|
||||
from langchain_openai import ChatOpenAI
|
||||
|
||||
prompt = ChatPromptTemplate.from_template(
|
||||
\"\"\"Answer the question based only on the context provided.
|
||||
|
||||
Context: {context}
|
||||
|
||||
Question: {question}\"\"\"
|
||||
)
|
||||
|
||||
llm = ChatOpenAI(model="gpt-3.5-turbo-0125")
|
||||
|
||||
def format_docs(docs):
|
||||
return "\n\n".join(doc.page_content for doc in docs)
|
||||
|
||||
chain = (
|
||||
{"context": retriever | format_docs, "question": RunnablePassthrough()}
|
||||
| prompt
|
||||
| llm
|
||||
| StrOutputParser()
|
||||
)
|
||||
|
||||
chain.invoke("...")
|
||||
|
||||
.. code-block:: python
|
||||
|
||||
# TODO: Example output.
|
||||
|
||||
""" # noqa: E501
|
||||
|
||||
# TODO: This method must be implemented to retrieve documents.
|
||||
def _get_relevant_documents(self, query: str) -> List[Document]:
|
||||
raise NotImplementedError()
|
||||
@ -1,4 +1,4 @@
|
||||
"""__ModuleName__ chat models."""
|
||||
"""__ModuleName__ toolkits."""
|
||||
|
||||
from typing import List
|
||||
from langchain_core.tools import BaseTool, BaseToolKit
|
||||
@ -6,7 +6,7 @@ from langchain_core.tools import BaseTool, BaseToolKit
|
||||
|
||||
class __ModuleName__Toolkit(BaseToolKit):
|
||||
# TODO: Replace all TODOs in docstring. See example docstring:
|
||||
# https://github.com/langchain-ai/langchain/blob/a6d1fb4275801a4850e62b6209cfbf096a24f93f/libs/community/langchain_community/agent_toolkits/sql/toolkit.py#L20
|
||||
# https://github.com/langchain-ai/langchain/blob/c123cb2b304f52ab65db4714eeec46af69a861ec/libs/community/langchain_community/agent_toolkits/sql/toolkit.py#L19
|
||||
"""__ModuleName__ toolkit.
|
||||
|
||||
# TODO: Replace with relevant packages, env vars, etc.
|
||||
@ -66,6 +66,6 @@ class __ModuleName__Toolkit(BaseToolKit):
|
||||
|
||||
""" # noqa: E501
|
||||
|
||||
# TODO: This method must be implemented to generate chat responses.
|
||||
# TODO: This method must be implemented to list tools.
|
||||
def get_tools(self) -> List[BaseTool]:
|
||||
raise NotImplementedError()
|
||||
|
||||
@ -15,7 +15,71 @@ class SearchDepth(Enum):
|
||||
|
||||
|
||||
class TavilySearchAPIRetriever(BaseRetriever):
|
||||
"""Tavily Search API retriever."""
|
||||
"""Tavily Search API retriever.
|
||||
|
||||
Setup:
|
||||
Install ``langchain-community`` and set environment variable ``TAVILY_API_KEY``.
|
||||
|
||||
.. code-block:: bash
|
||||
|
||||
pip install -U langchain-community
|
||||
export TAVILY_API_KEY="your-api-key"
|
||||
|
||||
Key init args:
|
||||
k: int
|
||||
Number of results to include.
|
||||
include_generated_answer: bool
|
||||
Include a generated answer with results
|
||||
include_raw_content: bool
|
||||
Include raw content with results.
|
||||
include_images: bool
|
||||
Return images in addition to text.
|
||||
|
||||
Instantiate:
|
||||
.. code-block:: python
|
||||
|
||||
from langchain_community.retrievers import TavilySearchAPIRetriever
|
||||
|
||||
retriever = TavilySearchAPIRetriever(k=3)
|
||||
|
||||
Usage:
|
||||
.. code-block:: python
|
||||
|
||||
query = "what year was breath of the wild released?"
|
||||
|
||||
retriever.invoke(query)
|
||||
|
||||
Use within a chain:
|
||||
.. code-block:: python
|
||||
|
||||
from langchain_core.output_parsers import StrOutputParser
|
||||
from langchain_core.prompts import ChatPromptTemplate
|
||||
from langchain_core.runnables import RunnablePassthrough
|
||||
from langchain_openai import ChatOpenAI
|
||||
|
||||
prompt = ChatPromptTemplate.from_template(
|
||||
\"\"\"Answer the question based only on the context provided.
|
||||
|
||||
Context: {context}
|
||||
|
||||
Question: {question}\"\"\"
|
||||
)
|
||||
|
||||
llm = ChatOpenAI(model="gpt-3.5-turbo-0125")
|
||||
|
||||
def format_docs(docs):
|
||||
return "\n\n".join(doc.page_content for doc in docs)
|
||||
|
||||
chain = (
|
||||
{"context": retriever | format_docs, "question": RunnablePassthrough()}
|
||||
| prompt
|
||||
| llm
|
||||
| StrOutputParser()
|
||||
)
|
||||
|
||||
chain.invoke("how many units did bretch of the wild sell in 2020")
|
||||
|
||||
""" # noqa: E501
|
||||
|
||||
k: int = 10
|
||||
include_generated_answer: bool = False
|
||||
|
||||
Loading…
Reference in New Issue
Block a user