Hanyar Ka'ida don Binciken Kudin Tambayoyin GraphQL

1. Gabatarwa

GraphQL ya kawo juyin juya hali ga ƙirar API na yanar gizo ta barin abokan ciniki su ƙayyade ainihin bayanan da suke buƙata. Duk da haka, wannan bayyanawa tana haifar da haɗari mai mahimmanci ga masu samar da sabis. Tambaya ɗaya, mara kyau, na iya neman adadin bayanai mai yawa, wanda ke haifar da nauyin uwar garken da ya wuce kima, ƙarin kuɗi, da yuwuwar raunin Rashin Sabis (DoS). Nazarin gwaji ya nuna yawancin aiwatar da GraphQL suna cikin haɗari. Wannan takarda tana magance babban gibi: rashin hanyar ka'ida, daidai, da inganci don kimanta kudin tambaya kafin aiwatarwa.

2. Bayan Fage & Ayyukan Da Suka Danganta

Hanyoyin yanzu na binciken kudin GraphQL sun gaza:

• Bincike Mai Motsi: Yana aiwatar da tambayoyi ko bincika bayan gida. Daidai amma mai tsada sosai don tace buƙatun ainihin lokaci (misali, Hartig & Pérez, 2018).
• Binciken Tsaye Na Yanzu: Sau da yawa mai sauƙi (misali, ƙidaya kumburin tambaya). Sun kasa yin la'akari da ka'idojin gama gari na GraphQL kamar girman jerin, hujjojin tambaya, da nau'ikan mahadar/ƙungiya, wanda ke haifar da kimantawa da ya wuce kima da kuma ƙasa da kima (misali, Kundin Sarƙaƙiyar GraphQL).

Wannan aikin ya sanya kansa a matsayin na farko da ke ba da binciken tsaye da aka tabbatar da daidaitonsa wanda duka layi ne a cikin sarƙaƙiya kuma ana iya daidaita shi zuwa ka'idojin tsari na ainihin duniya.

3. Tsarin Ma'anar GraphQL

Tushen binciken shine sabon tsari, ƙaƙƙarfan tsari na ma'anar aiwatar da GraphQL. Wannan tsarin tsari ya ƙayyade daidai:

• Tsarin tambayoyi da tsare-tsare.
• Warware filayen, gami da abubuwa masu zurfi da jerin abubuwa.
• Tasirin hujjojin tambaya (misali, `first`, `limit`) akan girman sakamako.

Wannan tsarin ya wuce bayanin GraphQL na zahiri, yana ba da damar yin tunani na lissafi game da hanyoyin aiwatar da tambaya da kudaden da suka danganci su. Yana ɗaukar tsarin GraphQL a matsayin jadawali mai jagora na nau'ikan, inda filayen suke gefuna.

4. Ma'aunin Sarƙaƙiyar Tambayar GraphQL

Takardar ta ayyana ma'auni biyu na farko na kuɗi, suna nuna damuwar masu ruwa da tsaki daban-daban:

1. Kudin Uwar garken ($C_s$): Yana ƙirar aikin da ayyukan warwarewa suka yi. Aiki ne na zurfin tambaya, faɗi, da kimanta girman jerin. A zahiri, ana iya bayyana shi azaman jimla akan hanyoyin tambaya: $C_s(Q) = \sum_{p \in Paths(Q)} \prod_{f \in p} weight(f)$, inda $weight(f)$ ke kimanta ƙididdigar filin $f$.
2. Girman Amsa ($C_r$): Yana ƙirar ƙarar bayanai a cikin amsar JSON, yana shafar canja wurin cibiyar sadarwa kai tsaye. Yana da alaƙa da adadin kumburi a cikin bishiyar amsa.

Ana sanya waɗannan ma'auni sigogi ta hanyar saitawa mai sauƙi da mai haɓaka API ya bayar (misali, girman jerin na asali = 10, zurfi mafi girma = 7).

5. Binciken Kudin Tsaye na Lokaci-Layin

Babbar gudummawar fasaha ita ce algorithm da ke lissafin iyakar sama don $C_s$ da $C_r$ a cikin lokaci da sarari na O(n), inda n shine girman daftarin aiki na tambaya (kumburin AST).

Zanen Algorithm:

1. Fashe & Tabbatar: Ana fashe tambayar zuwa AST kuma a tabbatar da ita bisa tsari.
2. Bayyana AST: Kowane kumburi a cikin AST ana sanya masa bayanin kula tare da masu canjin kuɗi bisa nau'insa (abu, jerin abubuwa, scalar) da ma'auni da aka saita.
3. Yada Kudade: Ziyara ɗaya daga ƙasa zuwa sama tana yada kimantawar kuɗi daga ganyayen zuwa tushen, yana amfani da ninkawa don jerin abubuwa masu zurfi da jimlawa don filayen 'yan'uwa.
4. Cire Iyaka: Bayanin kumburin tushen ya ƙunshi iyakar kuɗi ta ƙarshe.

Binciken yana magance fasalin GraphQL daidai kamar gutsure, masu canji, da hujjoji na cikin layi, yana haɗa su cikin lissafin kuɗi.

6. Kimantawa & Sakamako

An kimanta binciken akan sabon tarin tare 10,000 na ainihin tambayoyi-amsoshi daga APIs na kasuwanci na GraphQL guda biyu (GitHub da API na kamfani mai zaman kansa).

Fassarar Jadawali (Ra'ayi): Zanen zane-zane zai nuna ƙaƙƙarfan alaƙa mai kyau, layi mai kyau tsakanin Iyakar Sama da aka Lissafa (x-axis) da Ainihin Girman Amsa/Lokaci (y-axis) don hanyar da aka tsara, tare da maki da suka taru kusa da layin y=x. Maki don hanyar maras hankali za su kasance masu watsewa sosai, nesa da wannan layin.

7. Misalin Tsarin Bincike

Yanayi: API na blog tare da tambaya don samun posts da sharhinsu.

Saitin Tsari:

type Query {
  posts(limit: Int = 10): [Post!]!  # weight = 'limit' argument
}
type Post {
  title: String!
  comments(limit: Int = 5): [Comment!]! # weight = 'limit' argument
}
type Comment { text: String! }

Tambaya:

query {
  posts(limit: 2) {
    title
    comments(limit: 3) {
      text
    }
  }
}

Lissafin Kuɗi (Da Hannu):

• Girman jerin tushen `posts`: 2 (daga hujjar `limit`).
• Ga kowane `Post`, girman jerin `comments` mai zurfi: 3.
• Iyakar Sama na Kudin Uwar garken ($C_s$): $2 \times (1_{title} + 3 \times 1_{text}) = 2 \times 4 = 8$ kiran warwarewa.
• Iyakar Sama na Girman Amsa ($C_r$): $2_{posts} \times (1_{title} + 3_{comments}) = 8$ abubuwan JSON.

Binciken yana ziyartar tambayar sau ɗaya, yana amfani da waɗannan ƙa'idodin ninkawa, yana isa ga iyakar 8.

8. Aikace-aikace na Gaba & Hanyoyi

Binciken kuɗi na ka'ida ya buɗe hanyoyi da yawa:

• Iyakance Ƙimar Daidaitawa & Farashi: Matsa daga ƙirar farashi na tushen kuɗi (kamar AWS CloudWatch Logs Insights), inda abokan ciniki suka biya don sarƙaƙiyar lissafi, ba kawai kiran API ba.
• Ingantawa & Tsara Tambaya: Haɗa tare da masu tsara tambayar bayanai (misali, PostgreSQL, MongoDB) don GraphQL, kamar yadda masu inganta SQL ke amfani da kimanta kuɗi, kamar yadda aka bincika a cikin ayyuka kamar Hasura.
• Ƙirar Tsari Mai Ƙarfafawa: Kayan aiki don duba tsare-tsaren GraphQL yayin haɓakawa don raunin DoS, yana ba da shawarar iyakokin shafin ko iyakance zurfi, kama da ƙa'idodin ESLint don tsaro.
• Binciken Kudin GraphQL Tarayya: Faɗaɗa ƙirar don kimanta kuɗi a cikin tsarin tarayya (Apollo Federation), inda tambayoyi suka mamaye ƙananan jadawali da yawa, babban ƙalubale da ƙungiyar injiniyan Apollo ta lura.
• Haɗaɗɗiyar Koyon Injin: Yi amfani da bayanan tambaya/amsa na tarihi don koyo da ingantawa sigogin `nauyi` don filayen ta atomatik, matsawa daga saitawa na tsaye zuwa ƙirar kuɗi mai motsi, mai tuƙi da bayanai.

9. Nassoshi

1. Hartig, O., & Pérez, J. (2018). Ma'ana da Sarƙaƙiyar GraphQL. Proceedings of the World Wide Web Conference (WWW).
2. Facebook. (2021). Ƙayyadaddun GraphQL. https://spec.graphql.org/
3. Wittern, E., Cha, A., Davis, J. C., et al. (2019). Nazarin Gwaji na Tsare-tsaren GraphQL da Tasirin Tsaronsu. ICSE SEIP.
4. Gidauniyar GraphQL. (2022). Kayan Aikin Binciken Sarƙaƙiyar GraphQL.
5. GitHub. (2023). Takaddun API na GraphQL na GitHub. https://docs.github.com/en/graphql
6. Isola, P., Zhu, J., Zhou, T., & Efros, A. A. (2017). Fassarar Hotuna-zuwa-Hoto tare da Cibiyoyin Adawa na Sharadi (CycleGAN). CVPR.

10. Bincike na Kwararru & Zargi