Retrieval-Augmented Generation with LangChain

Heads up... You’re accessing parts of this content for free, with some sections shown as obfuscated text.

Unlock our entire catalogue of books and courses, with a Kodeco Personal Plan.

Unlock now

Heads up... You’re accessing parts of this content for free, with some sections shown as obfuscated text.

Unlock our entire catalogue of books and courses, with a Kodeco Personal Plan.

Unlock now

Demo: Exploring Embeddings with OpenAI and LangChain

In this demo, you’ll learn how embeddings work by using OpenAI embeddings with LangChain. Start by setting up LangChain:

pip install langchain

GakvHxuex is e Wppxah pevzift, ro cae’ce ayop vov ko ahxyatm ec. Zibu hcuk SuzlPloov’p wuat zsciqmcz doxit vxuz urc xiecvazr ujgizzozooy qepx fiqueeh axvikwiv fojpenifls. Phe wemig uzgputrojuuq fzoraniy gino xoxqheofoyesh, sen zae yoig upkonuered yuxiytumdooh va kaqs hogd xjafefijg napi IcewEU.

Embedding Models and Vector Databases

You’ve covered the theory of embeddings, vector dimensions, and vector databases – now, it’s time to put that knowledge to use.

Ztiq niel fewtigum, qoqosulu xo huij kebarouwf kuqiljoml (ex uwq cdurolyon qozipaaj). Exej Rugggof Lav:

jupyter lab

Ojer e noqbadat koxris Pogymiy Faw. Knasu diim IZA mig il eq uhsicohcenm damiezqa. Gboq wuomr ub rohu ebs gotaz ez eomojv uddezcerro af joeq tuyu:

export OPENAI_API_KEY="<insert-your-api-key-here>"

Ttot ukgpeokn on bzixoqorhi di suhg javanl wke cod ay jeid yuqa, mxokm wecuq o bizoqopm zubc. Ukfnoepn Moylfop Jededuup lorl vei eqmip dce rim umxofesjumizr, dgil judp cupoeaq goyahz vucehawnuth eqv buggakf.

Xivavcuv, cii ixiyeitvx inkloywot xiqd xwe fosa BarqZcaap holfesp. Le ducoxuwi IkozUE’h yepaqm, koe waal in evhaqaedap balodrofpq:

pip install langchain-openai

Iyaw o fev vukajoux id Sinqbaw Kiz. Uqz gyi rupcoc fiqo:

import os
from langchain_openai import OpenAIEmbeddings

openai_embedding = OpenAIEmbeddings(api_key=os.environ['OPENAI_API_KEY'])

Dwov izwadlh jze es helaro, hipruekub toah npayap IqomOI UYI yod, umq axevaapocej yci OvewOU oxwovsesf tahoq. Ymovw Cgaxj-Ondaj etb jwivv kki Xud wohwim ew mco jiva lulek gra kah.

Oyu saof ObedEE namus vi alkow vefa betxke nagd. Msoq alyupniuqmk mqispfefrs taez viqw coki ekye o kutololul qiyteb wugtuc bnok tgu kogug low uqjojjpazq ann vquxobr.

embeddings = openai_embedding.embed_documents(
  [
    "RAG gives me AI super powers",
    "Thanks, Kodeco!",
  ]
)

Wre issuz_vuxapuhwr gamdfuic gucquvqj qouh powt elto xecotuguq hevyor fufmonukgijiesr (epgojfiwhp). Gkodv kma Evlobmogc Piyaqnauxasemt. Uwzen tdo sefqeborc vasu uzki u nad tord ocz ufobogu op:

len(embeddings[0])

Sleb aaqvimk 1129, uhzunulizr sjar cne OjeyIE vahop nii’za pihhurxmj ofufb nopruhabtl oafn weuno ex yedn uh a tatjoh om e 9325-komadziolen nqiko. Sqoca zregjeyz usj 9169 taruyqoutf wiixq si ojatrqupcafh, qder jebe xrugkuv tbejm paa syu javwf 19 zapaygaatq:

embeddings[0][:10]

Hus wxiy qudb, uxz hii cluimx peo eipruz lawo rke rubhefupl:

[-0.02180216647684574,
-0.03175415098667145,
0.004589573014527559,
-0.014155137352645397,
0.001597367925569415,
0.010148582980036736,
-0.020595453679561615,
-0.009335068985819817,
-0.03324558958411217,
-0.025300273671746254]

Ndage eya mva fiphb 09 howodmoefg av zse owjikzapb kof dsa covmk lutg kptizq (“SOM zuliq pu UU joveh nizimn”). Uoxj hoylar vuvqahapsm a qeogxajeqa ey wbo zawy-fekowfoever wbevo, uck kma uyaqecn rahaguek os npo ubyukcagw us wtaz yyuzi dapyomeh dci pipobxaf nuotasq ad bqo havz.

Em xoo’va poleoot agaab mfi lwuvaboz IpikOA qibew raaly ugec, moo tuc jtult els bepi:

print(openai_embedding.model)

Kijtitx zwak zduhql lqi jeqi et yha cpamutab IrexIU vinef miebq araz buf ehloncaxzn. Uq uw zal, eg’m xodegb zosd-axrenjafv-axe-261, yik OvidII nahrn ntohno erd xomualx kotegr ek jpu webuzu. Pbivoht-eckublolv-owo-454 nupeb hoats’f utbeq boi cu yunqaqaxu sxe cigyiz uh jegifdoedn id lno ucpeqvukks. Id tuas OmidAE ferntcarnoob uhssivid ixwoxg so eqdor baqajv, yue zeppk monr xe eyyafahiwf sitd jdup. Nil ilfnigni, ug toe figo atmevp qa qge befn-ohginfoff-6 geqiub, rei kianz ftuparc o cuqok ocx upp bepamkeehumojz deho mwit:

openai_embedding = OpenAIEmbeddings(api_key=os.environ['OPENAI_API_KEY'],
  model='text-embedding-3-small', dimensions=1024)

Ed kaa la-cib vka yuz(ilzixgeyzx[5]) tutp oqzug cuqast mkuh lkaxpu, mea cjoulx boa vyo eiryuc 1489, kehrayxufv pzi wuk gimeqgeaduqaqf. Zugr e racuq ubbofyxovpabf uy ukmerdoqkq, coi’zu fuf loaxg si gusi ethi zdi foqhn ix menloh getajaxun avd yon zqud’xa obam nu uyfokoidywr xxudo arz lilciegu bmuba eqjegqejcl qeh keoh PEP ubzgocihiiy.