Ukusebenzisa i-Jupyter Lab, ngithuthele iningi lomsebenzi wami ezincwadini ze-Marimo zokubhalela, incwajana entsha yePython ebhekana nezinkinga eziningi zokuma ezinde ngendabuko. Le ndatshana ihlanganisa izizathu ngemuva kokuguqulwa kwami nokuthi i-marimo ifanela kanjani ngokwemvelo ekuhambeni kwami kokusebenza kwamanje, ukubonga okugcwele kwe-jupyter yokwakha i-ecosybook yezincwadi, ucwaningo kanye nemfundo.
Isizinda esincane
Ngichithe iminyaka engaphakathi kwezincwadi. Isikhathi eside, ukusetha kwami bekuyinhlanganisela ye-jupyter amabhuloki / ilebhu yokuhlolwa kwasendaweni kanye ne-Google colab yefu noma i-GPU umsebenzi wabo wamahhala. Ngiphinde ngagunyaza amabhulogi amaningi ekwenzeni okuningi kwalezi zincwajana, lapho ngihlanganyele khona amathiphu wokudlula ngendlela ejwayelekile yokuzisebenzisa.
Eseshilo lokho, kube nezinkinga ezithile ngalezi zincwadi zendabuko, ukuthengiswa okukhulu kakhulu kwezobuciko (noma kunalokho, ukuntuleka kwawo). Umbuso ofihliwe, ukwenziwa kwe-oda, kanye nama-reluns ebhukwana kwakungamanye amaphuzu ami abuhlungu ngalezi zincwadi zokubhala.
Lokho kwashintsha lapho ngithola I-Marimo yasolwandleEnye indlela evulekile ye-juphitterbooks lapho ungagijima khona amaseli, wenze amakhono, bese udala iziza njengoba nje ubuzokwenza eJupyter. Kepha ngaphansi, kusebenza ngendlela ehluke kakhulu futhi kulungisa izingqinamba eziningi zokuma isikhathi eside. Umehluko osemqoka ukuthi i-Marimo Notebook ifayela nje le-python, elenza kube yincwajana engcono nemvelo engcono yokufaka amakhodi kunoma ngubani osebenzisa i-Python. Kulesi sihloko, ngabelana ngezizathu eziyishumi ngemuva kokushintsha kwami nokuthi kungani ukusebenza ngezincwajana kubembisene kakhulu kimi kusukela kuthuthela eMarimo.
Uma ufuna ukulandela, faka uMarimo njengoba kukhonjisiwe ngezansi. Lo myalo wethula umhleli weMarimo futhi uvula incwadi yokubhalela ngaphakathi Imodi yeSandboxokusebenza njengemvelo ebonakalayo engasindi eligcina incwadi yakho yokubhalela ihlukaniswe nohlelo lwakho lonke. Ungasebenzisa futhi
marimo.newUkuqala incwadi entsha yokubhalela endaweni yokudlala online. Ngemiyalo enemininingwane, bheka umcabango.
pip install marimo
marimo edit --sandbox notebook.py1. Ukusebenza kabusha okugcina kulungiswa izindaba zombuso ezifihliwe
Izincwadi zendabuko azisebenzi. Lokhu kusho ukuthi uma ngishintsha okuthile kuseli elilodwa, akukho okunye okuvuselelwa ngokwawo. Imvamisa ngigcina ngibagijima ngiphinde ngibagijime ukuze ngigcine imiphumela ihambisane. Isibonelo, cabanga ngencwadi ye-jupyter engezansi lapho ngakwesobunxele ngichaza ukuguquguquka okubili, a na- bengeza ndawonye ukuze uthole ukuguquguquka kwesithathu cbese ukhombisa inani le cokusho ukuthi 30.
Manje, ukube bengizoshintsha a kuya ku-20 njengoba kukhonjisiwe ngakwesokudla, c ngabe usakhombisa 30 ngaphandle kokuthi ngiphinde ngagijimela esitokisini. Ukuthi ukungahambi kahle phakathi kwekhodi nokuphuma kuyinkinga yezwe efihlekile efihlekile.
UMarimo ulungisa lokhu Ukubulawa okusebenzayo, lapho wonke iseli eyingxenye ye igrafu yokuncika (dag). Izinguquko engqamuzaneni elilodwa ngokuzenzakalelayo zibangela izibuyekezo kumaseli ancike. Ngakho-ke, nini a Izinguquko, uMarimo uhlola kabusha noma yimaphi amaseli ancike. Bona kanjani c Kuvuselelwa ngokuzenzakalelayo lapho inani le a ishintshiwe.
Lokhu kugcina ikhodi yami kanye nemiphumela ebonakalayo ngokuphelele ku-sync futhi kusho ukuthi ngingabelana ngezincwajana zami nabanye ngaphandle kokukhathazeka nge-oda lapho benza khona.
Kepha Ingabe wonke amaseli okuvuselelwa ngokuzenzakalelayo? Hhayi njalo. Amanye amaseli abiza ukugijima, njengalabo abahilela ukuqeqeshwa kwe-ML noma ukucubungula kwedatha esindayo. Izindaba ezimnandi ukuthi uMarimo uyazisekela Ukubulawa okuvilaokukuvumela ukuthi ucime ukuvuselelwa okuzenzakalelayo futhi kubangele amaseli anjalo kuphela lapho kudingeka.
2
Njengoba ama-marimo amanothi akhipha amaseli asuselwa ebudlelwaneni obuguquguqukayo hhayi i-oda labo ekhasini, lokhu kunginikeza inkululeko yokuhlela ikhodi yami ngendlela enengqondo kimi. Isibonelo, ngingabeka yonke imisebenzi yami yokungenisa ngaphakathi kanye nemisebenzi ezansi ukuze ngigcine indawo yokusebenza ihlanzekile.
Lokhu kuvula amacala athile okusebenzisa okumnandi. Ngikusebenzise lokhu ukudala imibuzo emincane enezinyathelo ezenziwe nge-auto lapho ukhiye wokuphendula uhlala ezansi kwencwadi yokubhalela. Imibuzo isasebenza kahle ngoba i-oda leseli alinandaba, futhi izimpendulo azibonakali ngokushesha kulowo muntu othatha imibuzo. (Futhi yebo, uMarimo uza nemodi emnyama nayo 😃.)
3. Ngokuyisisekelo ifayela le-python ngaphansi kwe-hood
Njengoba yonke i-Marimo Notebook igcinwa njengefayela le-Python elihlanzekile, lixazulula inkinga yami ende ngokuma kwe-git. Ekugcineni ngikwazi ukulandelela izinguquko zihlanzekile ngaphandle kokubhekana ne-messy JSON Diff. Incwajana yokubhalela ibuye ibe nezimo. Ngiyakwazi ukuyisebenzisa njengesiskripthi, wugijime kusuka kulayini womyalo, noma uyiguqule ibe uhlelo lokusebenza oluncane olusebenzayo ngaphandle kokushintsha noma yini.
Ngezansi, uhlangothi lwesokunene lubonisa ukubuka kwe-Marimo NotebookBook, ngenkathi uhlangothi lwesobunxele lubonisa incwadi efanayo evulwe njengefayela elijwayelekile le-python ku-umhleli.
4. Ngingakwazi kalula ukuguqula amabhukwana ami abe kuzinhlelo zokusebenza
Ukuguqula incwajana ye-Marimo ku-App esebenzisana nayo futhi kulula. Ngiyakwazi ukushintsha phakathi Hlela Imodi na- Ukubuka kohlelo lokusebenza Ngokuchofoza okukodwa, futhi angikaze ngiphinde ngibhale ikhodi yami noma ngingeze noma yini eyengeziwe.
Nayi incwadi yokubhala eyisibonelo eguqula amazinga okushisa avela Celsius kwa- Fahrenheit. Incwajana efanayo ingabukwa ngaphakathi Imodi yohlelo lokusebenza usebenzisa ukwakheka okuzenzakalelayo okuqondile.
Lesi sici siwusizo ikakhulukazi ekwakheni amabhulogi asebenzayo, tutorials, kanye nokuhamba kwezemfundo kwemitapo yolwazi. Uma wake wafunda Ai aclorablesuyawazi umuzwa wokuzibandakanya nokuqukethwe kunokuba uwufunde nje. UMarimo unikeza umuzwa ofanayo futhi Ukwakheka kwegridi Kwenza kube ngcono kakhulu njengoba ngikwazi ukuhudula bese ngidonsa imiphumela yokuhlela uhlelo lokusebenza indlela engifuna ngayo.
I-5
UMarimo uza nomphathi wephakheji owakhelwe ngaphakathi ongivumela ukuthi ngifake imitapo yolwazi engekho ngqo kusuka encwadini yokubhalela. Uma imodyuli ingatholakali, uMarimo ukhombisa ngokushesha okukhulu, futhi ngingayifaka ngokuchofoza. Lokhu kususa okujwayelekile emuva kanye ne-the-front ngama-terminals.
UMarimo uphinde anikeze ukuzala kwami. Lapho ngisebenzisa incwadi yokubhalela nge --sandbox Ifulegi, ilandelela wonke amaphakheji nezinhlobo zazo ngqo futhi isikonga ngaphakathi kwefayela le-notebook njengokuphawula okuphezulu. Asikho isidingo sokwehlukana requirements.txt Ifayela ngoba incwadi yokubhalela isivele ithwala konke okudingayo.
I-6. Ukusetshenziswa okwakhelwe ngaphakathi kokusebenza nge-dafaframes.
Uma uchitha isikhathi esiningi sokuhlola idatha, lokhu kungenye yezici ezizwakala ngokushesha njenge-medsend. Uma ubonisa i-DataFrame eMarimo, ukubukwa okuzenzakalelayo kusebenza ngokugcwele futhi kuza namathuluzi amaningana awusizo aphuma ebhokisini, njengokuphenya, ukuphenya, ama-histogams we-automal for colourns, nokuhlunga okulula nokuhlunga okuvela kubahenqo zekholomu. Lokhu kusheshisa inqubo ye-EDA kimi, njengoba ukuhlola okubonakalayo kuvame ukunika ukuqonda okungcono.
UMarimo futhi ungivumele ngivule i-dataFRAME ibe yinto yokufaka yokufaka. Ukusebenzisa mo.ui.table,Ngiyakwazi ukukhetha imigqa bese ngisebenzisa lokho kukhethwa kumanani aphansi. Ukukhethwa kokuvuselela kabusha okusebenzayo, okusho ukuthi futhi ezinye izibuyekezo ze-notebook zizenzakalelayo.
Uma ngifuna ukuhlanganyela incwadi yokubhalela nge-a Isihleli se-UI DataFrame Lokho kubavumela ukuba bahudule futhi baphonse ukuhlanganiswa nokuguqulwa ngaphandle kokubhala noma iyiphi ikhodi. Kukhiqizwa ngisho nekhodi efanayo ye-Python kuzo zonke izinyathelo.
Bese ulinda, akusikho konke. Kukhona futhi a I-Data Explorer Ukusetshenziswa okungivumela ukuba ngikhale ngokubonakalayo okubonakalayo ngqo kusuka kudathaframe.
Konke lokhu kwenza idatha yedatha yesigaba sokuqala ngokushesha okukhulu. Esikhundleni semitapo yolwazi ye-juggling noma ikhodi yokubhala ye-boilerplate ukuze nje uqonde idatha yedatha, ngikwazi ukuqala ukuhlaziya nangokubona ngeso lengqondo khona lapho ngaphakathi kwencwajana.
7. Bona imibhalo njengoba uthayipha
Enye yezinto ezingaphansi kakhulu ze-Marimo Notebooks for Me bekuyiphaneli yedokhumenti ebukhoma. Ngiyabona imisebenzi ye-docstrings, amapharamitha, nezibonelo ngaphandle kokubiza izinsiza zosizo.
Iphaneli futhi iyasebenzisana, ngakho-ke ngikwazi ukupheqa ngokusebenzisa amadokhumenti amade, hlola izibonelo, bese ukopisha ama-snippets kubhukwana lami. Uma i-docstring ihlehliswe izibonelo, uMarimo ubathobela futhi angenze ukubakopisha ngqo esitokisini esisha, obekusihle kakhulu ngenkathi ngifunda imitapo yolwazi emisha.
8. SQL ngaphakathi kwencwadi yokubhalela
I-Marimo inokusekelwa okukhulu kwe-SQL. Ngingahlala ebhukwini elilodwa futhi ngixube i-SQL ne-Python ngendlela yemvelo. Ngingabhala umbuzo we-SQL ngaphakathi kweseli ye-SQL, wugijime, futhi uthole umbono ofanayo wethebula elihlanzekile elinikeza i-Marimo yedatha.
Ngiyakwazi futhi ukuqhuba i-SQL kumafayili endawo nge-Duckdb. A alula select * from "file.csv" Isebenza nje. Uma ngifuna ukuxhuma database yangempela, ngingenza lokho kusuka kuphaneli eseceleni. I-Marimo ikuvumela ukuthi ngingeze imithombo yedatha ehlukile, futhi uma kungeze, amatafula avela ngqo kubhukwana.
9. UMolab: Amabhukwana e-Marimo efwini
UMolab ungowamabhukwana eMarimo, ukuthi yi-Google Colab enjani ye-jupyterbooks. Uthola konke ubuhle be-Marimo Norteoks efwini. Ungavula i-molab at molab.marimo.iodala izincwadi ezintsha zokubhalela, noma uvule labo owenzile ngaphambili. Kuyindlela enhle lapho ngifuna ukusongela incwajana yokubhalela ngokushesha noma ukugwema ukwenza izinto emshinini wami. Amaphakheji amaningi athandwayo wePython afakwe kwangaphambili, amabhukwana anokugcina okuphikelelayo, futhi ngiyakwazi ukwaba, ukulanda, noma ngiwalayishe kalula.
Enye into ehlanzekile ukuhlanganiswa kwe-GitHub. Ungaya ku-URL ye-molab bese ubeka esikhundleni molab.marimo.io/notebooks ne- molab.marimo.io/github Futhi kuzokuvumela uvule noma yiliphi incwadinook, i-jupyter noma i-marimo, ephethwe ku-github eMolab. Kodwa-ke, akukho ukwesekwa kwe-GPU okwamanje, njengaseColob, kepha ngiyaqiniseka ukuthi yinto ethile iqembu elizosebenza kulo.
10. Ukuhlanganiswa kwe-LLMS ngokwezifiso
UMarimo futhi unezici ezithile eziwusizo ze-AI ezisize ezisizwayo ze-AI ezakhiwe kuzo, eziye zangisiza kakhulu. Ukuze uthembeke, iJupyter futhi inokunwetshwa kweJupyterlab kwezimpawu ze-AI ezibizwa nge-AI Jupytai (Futhi yebo, ngibhale ngalokho), kepha ukusekelwa kwe-AI okwakhelwe eMarimo kususa ingxabano enkulu.
Ngisebenzisa i-AI eMarimo ikakhulukazi ngezinto ezimbili. Owokuqala ukukhiqiza amabhuku amasha lapho ngifuna ukuzama umtapo wezincwadi omusha noma wenze idemo esheshayo. Kungisindisa isikhathi ngoba akudingeki ukuthi ngizingele amathoyizi amatatomu noma setha yonke into kusuka ekuqaleni.
Okwesibili kuhlanza futhi kubangela iphutha. Umsizi we-AI kaMarimo unesimo esigcwele seNcwadi yami yokubhalela, ngakho-ke ingafunda ikhodi yami futhi angisize ngilungise noma ngiyihlanze ngakwesokudla ngaphakathi kwencwajana.
Nami ngithanda ukuthi ngikwazi ukukhetha amamodeli ahlukile engxoxweni futhi ngihlele, bese ngibeka eyami imithetho yokuthi ngifuna kanjani ikhodi ukulotshwa. Futhi uma ngifuna ukugwema amamodeli okuphathelene, u-Ollama usebenza kahle njengomhlinzeki futhi.
Ukugcina
Angixhunyiwe eqenjini le-Marimo futhi lokhu akuyona into ekhokhelwayo, kodwa isipiliyoni sami ekusebenziseni kumaphrojekthi wangempela. Lokhu kufundwe isikhathi eside ngoba bengifuna ukuhlanganyela ukuthi uMarimo wenze kanjani incwadi yami yokubhala isebenze kahle. Ngikhipha kuphela izingxenye ezingisizile kakhulu. Amadokhumenti ahamba ajulile kakhulu futhi ayindawo engcono kakhulu yokuhlola yonke into iMarimo okufanele ayinikele.
