Gukoresha Agentic AI ya DevOps Automation hamwe na Kiro na LangChain

Intangiriro

Muri iki gihe cyihuta cyane cyiterambere rya software, amakipe ya DevOps ahora ashakisha uburyo bwo kunoza imikorere no kwizerwa byoguhuza kwabo no gukomeza kohereza (CI / CD). Bumwe mu buryo butanga ikizere ni ugukoresha ibikoresho bya AI, bifasha kurema ibintu byubwenge bishobora gukora kandi bigahindura imirimo ya DevOps. Muri iki kiganiro, tuzasuzuma uburyo bwo gushyira mubikorwa ibikoresho bikoreshwa na AI dukoresheje Kiro na LangChain, ibice bibiri bigezweho bigenda bihindura murwego rwo gutangiza DevOps.

Agentic AI ni iki?

Agentic AI bivuga ubwoko bwubwenge bwubuhanga bugenewe gukora imirimo yigenga, nkumuntu wumuntu. Mu rwego rwa DevOps, agent ya AI irashobora gukoreshwa mugutangiza imirimo nko kubaka, kugerageza, no gukoresha porogaramu za software. Mugukoresha imashini yiga imashini hamwe no gutunganya ururimi karemano (NLP), abakozi ba AI barashobora kwigira kuburambe no guhuza nibidukikije bihinduka, bigatuma biba byiza kubikorwa bya DevOps bigoye.

Kumenyekanisha Kiro na LangChain

Kiro na LangChain nuburyo bubiri bufungura-isoko yabugenewe yo kubaka sisitemu ya AI. Kiro ni urwego rushingiye kuri Rust rutanga urubuga rukomeye kandi runini rwo kubaka abakozi bafite ubwenge, naho LangChain ni urwego rushingiye kuri Python rwibanda kuri NLP no gusobanukirwa ururimi.

Kiro: Urwego rushingiye kumurongo wo kubaka abakozi b'ubwenge

Kiro nuburyo bugezweho bushingiye kuri Rust butanga urubuga rwuzuye rwo kubaka abakozi bafite ubwenge. Hamwe na Kiro, abitezimbere barashobora gukora abakozi bashobora gukorana na sisitemu zitandukanye, serivisi, namakuru yamakuru, kandi bagakora imirimo nka:

• Gutangiza ibikorwa byakazi
• Gusesengura amakuru no gufata ibyemezo
• Gushyikirana nabantu hamwe nubundi buryo

Ubwubatsi bwa Kiro bwagenewe kuba bunini cyane kandi bworoshye, bukaba ari amahitamo meza kumishinga minini ya DevOps yo gutangiza.

LangChain: Imikorere ishingiye kuri Python ya NLP no Gusobanukirwa Ururimi

LangChain ni Python ishingiye kubikorwa byibanda kuri NLP no gusobanukirwa ururimi. Hamwe na LangChain, abitezimbere barashobora gukora abakozi bashobora kumva no kubyara imvugo imeze nkabantu, kandi bagakora imirimo nka:

• Gutunganya ururimi karemano (NLP)
• Gusesengura inyandiko no gusesengura amarangamutima
• Guhindura ururimi n'ibisekuru

Ubwubatsi bwa LangChain bwagenewe kuba modular cyane kandi bwagutse, bigatuma ihitamo neza mukubaka NLP yihariye hamwe nubushobozi bwo gusobanukirwa ururimi.

Gushyira mubikorwa abakozi bakora AI hamwe na Kiro na LangChain

Kugaragaza imbaraga za Kiro na LangChain, reka dusuzume urugero rworoshye rwo kubaka agent ikoreshwa na AI itangiza umuyoboro wa CI / CD. Muriyi ngero, tuzakoresha Kiro kugirango dukore agent ishobora gukorana nububiko bwa Git, kubaka no kugerageza kode, no kohereza porogaramu kumurongo wibicu.

Hano hari kode yoroshye yerekana uburyo bwo gukora umukozi wa Kiro utangiza umuyoboro wa CI / CD: `` python Kiro kuva muri langchain itumiza LLAMA

umukozi = kiro. Byihutirwa ( izina = "ci_cd_agent", ibisobanuro = "Yikora umuyoboro wa CI / CD", ibikorwa = [ kiro. Igisubizo ( izina = "git_clone", ibisobanuro = "Clones ububiko bwa Git", code = "" " git repo = git.Repo.clone ("https://github.com/user/repo.git") "" " ), kiro. Igisubizo ( izina = "kubaka", ibisobanuro = "Yubaka kode", code = "" " gutumiza mu mahanga subprocess.run (["gukora", "kubaka"]) "" " ), kiro. Igisubizo ( izina = "ikizamini", ibisobanuro = "Ikora ibizamini", code = "" " gutumiza mu mahanga subprocess.run (["gukora", "ikizamini"]) "" " ), kiro. Igisubizo ( izina = "kohereza", ibisobanuro = "Kohereza porogaramu", code = "" " gutumiza mu mahanga subprocess.run (["gukora", "kohereza"]) "" " ) ] )

Kora icyitegererezo cya LangChain LLAMA

llama = LLAMA ( izina = "llama_model", ibisobanuro = "LLAMA icyitegererezo cyo gusobanukirwa ururimi", code = "" " gutumiza mu mahanga icyitegererezo = ururimi.LLAMA ( izina = "llama_model", ibisobanuro = "LLAMA icyitegererezo cyo gusobanukirwa ururimi", code = "" " gutumiza mu mahanga icyitegererezo = ururimi.LLAMA ( izina = "llama_model", ibisobanuro = "LLAMA icyitegererezo cyo gusobanukirwa ururimi", code = "" " def inzira_text (inyandiko): # Tunganya inyandiko ukoresheje icyitegererezo cya LLAMA garuka inyandiko "" " ) "" " ) "" " )

Kora umukozi wa Kiro ukoresha moderi ya LangChain LLAMA

umukozi = kiro. Byihutirwa ( izina = "ci_cd_agent", ibisobanuro = "Yikora umuyoboro wa CI / CD", ibikorwa = [ kiro. Igisubizo ( izina = "git_clone", ibisobanuro = "Clones ububiko bwa Git", code = "" " git repo = git.Repo.clone ("https://github.com/user/repo.git") "" " ), kiro. Igisubizo ( izina = "kubaka", ibisobanuro = "Yubaka kode", code = "" " gutumiza mu mahanga subprocess.run (["gukora", "kubaka"]) "" " ), kiro. Igisubizo ( izina = "ikizamini", ibisobanuro = "Ikora ibizamini", code = "" " gutumiza mu mahanga subprocess.run (["gukora", "ikizamini"]) "" " ), kiro. Igisubizo ( izina = "kohereza", ibisobanuro = "Kohereza porogaramu", code = "" " gutumiza mu mahanga subprocess.run (["gukora", "kohereza"]) "" " ) ], ururimi_model = llama )

Koresha umukozi

umukozi.run () `` ` Iyi kodegisi yerekana uburyo bwo gukora umukozi wa Kiro utangiza umuyoboro wa CI / CD ukoresheje icyitegererezo cya LangChain LLAMA kugirango wumve ururimi. Intumwa irashobora gukorana nububiko bwa Git, kubaka no kugerageza kode, no kohereza porogaramu kumurongo wibicu.

Hejuru 5-10 Amabwiriza Yakoreshejwe Byinshi / Udukoryo

Dore top 5-10 zikoreshwa cyane / ibisobanuro kuri Kiro na LangChain:

1. kiroutil init: Itangiza umushinga mushya wa Kiro.
2. kiroutil kubaka: Yubaka umukozi wa Kiro.
3. langchain init: Itangiza umushinga mushya wa LangChain.
4. langchain kubaka: Yubaka icyitegererezo cya LangChain.
5. kiroutil run: Ikoresha umukozi wa Kiro.
6. Ikizamini cya langchain: Gerageza icyitegererezo cya LangChain.
8. langchain gusuzuma: Isuzuma icyitegererezo cya LangChain.
9. kiroutil debug: Gukemura umukozi wa Kiro.
10. langchain visualize: Yerekana icyitegererezo cya LangChain.

Kubona amashusho

Hano hari amashusho yububiko bwa Kiro na LangChain ukoresheje Chart.js: json { "ubwoko": "akabari", "amakuru": { "ibirango": ["Kiro", "LangChain"], "imibare": [ { "label": "Ubwubatsi", "amakuru": [100, 80], "backgroundColor": ["# 4CAF50", "# FF9800"] } ] }, "amahitamo": { "umutwe": { "kwerekana": ukuri, "inyandiko": "Ubwubatsi bwa Kiro na LangChain" }, "umugani": { "kwerekana": ukuri } } } ` Iyi visualisation yerekana urwego rwohejuru rwubatswe rwa Kiro na LangChain, rugaragaza imbaraga nintege nke zabo.

Umwanzuro

Muri iki kiganiro, twasuzumye imbaraga za agentic AI yo gukoresha mudasobwa ya DevOps dukoresheje Kiro na LangChain. Mugukoresha ubu buryo bugezweho, abitezimbere barashobora gukora ibintu byubwenge bishobora gukoresha no guhuza imiyoboro ya CI / CD, kunoza imikorere ya software no kugabanya ingaruka zamakosa. Hamwe na Kiro na LangChain, abitezimbere barashobora gukora NLP yihariye hamwe nubushobozi bwo gusobanukirwa ururimi, kandi bagakorana na sisitemu zitandukanye, serivisi, ninkomoko yamakuru. Twatanze kandi kode yoroshye yerekana uburyo bwo gukora agent ya Kiro itangiza umuyoboro wa CI / CD ukoresheje icyitegererezo cya LangChain LLAMA kugirango wumve ururimi.