際際滷

際際滷Share a Scribd company logo

Extracting Insights from the Topology of the JavaScript Package Ecosystem

Au Gai
Au Gai

The document summarizes research analyzing the topology of the JavaScript package ecosystem using topology data analysis (TDA). Key findings from the TDA include: 1) The number of package dependencies is a strong feature in the topology that separates packages. 2) Packages more likely to be used within the ecosystem are located separately from those meant for external application usage. 3) Comparison with archetypal analysis identified archetypes of packages based on common keywords and example packages for each archetype.

1 of 20
Download to read offline
Extracting Insights from the Topology of the JavaScript Package Ecosystem Authors : Nuttapon Lertwittayatrai+ Raula Gaikovina Kula* Saya Onoue* Hideaki Hata* Arnon Rungsawang+ Pattara Leelaprute+ Kenichi Matsumoto* + Kasetsart University, Thailand * Nara Institute of Science and Technology, Japan
Software Ecosystem ? Lungu* : a collection of software systems, which are developed and co-evolve in the same environment. 2 *Lungu, Mircea (2009). Reverse Engineering Software Ecosystems (Ph.D.). University of Lugano.
https://www.npmjs.com/ accessed 2017-12-02 Its Big..
Analysis of the ecosystem...
Complex Characteristics of Packages- as vectors 5 name: npm version: 5.6.0 license: Artistic-2.0 keywords: install,modules,package manager,package.json name: browserify version: 14.5.0 license: MIT keywords: browser,require,commonjs,commonj-esque,bundle,npm,javascript name: express version: 4.16.2 license: MIT keywords: express,framework,sinatra,web,rest,restful,router,app,api
Problem: How to analyze high dimensions of Package Characteristics? 6 http://setosa.io/ev/principal-component-analysis/ 2 vectors 3 vectors 4+ vectors name version name version license name version license keywords author ...
Shape of the Ecosystem http://slideplayer.com/slide/238747/ 7
Topology is a major area of mathematics -- spatial properties with deformations of objects. TDA is the result of a concerted effort to adapt topological methods to various applied problems, one of which is the study of large and high dimensional datasets. 8 Topology Data Analysis (TDA) Shape of Data P. Y. Lum, G. Singh, A. Lehman, T. Ishkanov, M. Vejdemo-Johansson, M. Alagappan, J. Carlsson, and G. Carlsson, Extracting insights from the shape of complex data using topology. from Nature
NBA TDA Players example 9P. Y. Lum, G. Singh, A. Lehman, T. Ishkanov, M. Vejdemo-Johansson, M. Alagappan, J. Carlsson, and G. Carlsson, Extracting insights from the shape of complex data using topology. from Nature
Shape of the ecosystem 10https://firebearstudio.com/blog/the-most-popular-packages-for-bower-and-npm.html To uncover insights and extract patterns of existing packages. Datasets gathered from software ecosystems are vastly high-dimensional, noisy and are generally challenging ...
TDA to understand shape of Application vs. Npm Packages A set of keywords that were likely to be related to either applications (GitHub strong) or the npm package ecosystem (npm strong). 11 GitHub strong npm strong gruntplugin util gulpplugin array express buffer react string authenticate file E. Wittern, P. Suter, and S. Rajagopalan, A look at the dynamics of the JavaScript package ecosystem, in Proceedings of the 13th International Workshop on Mining Software Repositories - MSR 16. New York, New York, USA: ACM Press, 2016, pp. 351361.
Extract Feature Vector (1/2) 12 f1 Author: Name of person who build this package. f2 Author Domain: Email domain of person who build this package. f3 License: License tell people know what organization that publish the package how they are permitted to use it. package.json
Extract Feature Vector (2/2) 13 f4 Tagged Keywords: An array of strings that helps people discover your package as its listed in npm search. f5 Version Released: Version form an identifier that is assumed to be completely unique. f6 Number of Dependencies: The number of mapped package dependencies to a version range.package.json
TDA High Dimension Mapping Dataset 14 f1 (Author) f2 (Author Domain) f3 (License) Knotter (TDA) Vector Space Model (VSM) f4 (Keywords) f5 (Version) f6 (No.dependencies) word2vec estimate version count dependencies package.json 151,000 packages https://github.com/rosinality/knotter
Results 15
Results The number of package dependencies is a strong feature in the topology 16
Results Packages that are more likely to be used within ecosystem are located separately from packages meant for application usage outside the ecosystem 17
Comparison with Archetypal Analysis 18 Archetype 1 (A1) has packages that contains keywords, such as web, plugin, test, http, express, node, api and server. Archetype 2 (A2) has packages that contains keywords like html, gulpplugin, css, javascript and gulp. Archetype 3 (A3) has lower packages compared to the other two archetypes.
Comparison with Archetypal Analysis 19 Archet ype Identified Packages A1 tar-parse, turtle-run, marked-sanitized, haversort, bmxplayjs A2 statsd-influxdb-backend, ardeidae, demo-blog-system, git-ssb- web, social-media-resolver A3 stream-viz, programify, polyclay-couch, meshblu-core-task-check- update-device-is-valid, apidoc-almond
20

More Related Content

Extracting Insights from the Topology of the JavaScript Package Ecosystem

  • 1. Extracting Insights from the Topology of the JavaScript Package Ecosystem Authors : Nuttapon Lertwittayatrai+ Raula Gaikovina Kula* Saya Onoue* Hideaki Hata* Arnon Rungsawang+ Pattara Leelaprute+ Kenichi Matsumoto* + Kasetsart University, Thailand * Nara Institute of Science and Technology, Japan
  • 2. Software Ecosystem ? Lungu* : a collection of software systems, which are developed and co-evolve in the same environment. 2 *Lungu, Mircea (2009). Reverse Engineering Software Ecosystems (Ph.D.). University of Lugano.
  • 4. Analysis of the ecosystem...
  • 5. Complex Characteristics of Packages- as vectors 5 name: npm version: 5.6.0 license: Artistic-2.0 keywords: install,modules,package manager,package.json name: browserify version: 14.5.0 license: MIT keywords: browser,require,commonjs,commonj-esque,bundle,npm,javascript name: express version: 4.16.2 license: MIT keywords: express,framework,sinatra,web,rest,restful,router,app,api
  • 6. Problem: How to analyze high dimensions of Package Characteristics? 6 http://setosa.io/ev/principal-component-analysis/ 2 vectors 3 vectors 4+ vectors name version name version license name version license keywords author ...
  • 7. Shape of the Ecosystem http://slideplayer.com/slide/238747/ 7
  • 8. Topology is a major area of mathematics -- spatial properties with deformations of objects. TDA is the result of a concerted effort to adapt topological methods to various applied problems, one of which is the study of large and high dimensional datasets. 8 Topology Data Analysis (TDA) Shape of Data P. Y. Lum, G. Singh, A. Lehman, T. Ishkanov, M. Vejdemo-Johansson, M. Alagappan, J. Carlsson, and G. Carlsson, Extracting insights from the shape of complex data using topology. from Nature
  • 9. NBA TDA Players example 9P. Y. Lum, G. Singh, A. Lehman, T. Ishkanov, M. Vejdemo-Johansson, M. Alagappan, J. Carlsson, and G. Carlsson, Extracting insights from the shape of complex data using topology. from Nature
  • 10. Shape of the ecosystem 10https://firebearstudio.com/blog/the-most-popular-packages-for-bower-and-npm.html To uncover insights and extract patterns of existing packages. Datasets gathered from software ecosystems are vastly high-dimensional, noisy and are generally challenging ...
  • 11. TDA to understand shape of Application vs. Npm Packages A set of keywords that were likely to be related to either applications (GitHub strong) or the npm package ecosystem (npm strong). 11 GitHub strong npm strong gruntplugin util gulpplugin array express buffer react string authenticate file E. Wittern, P. Suter, and S. Rajagopalan, A look at the dynamics of the JavaScript package ecosystem, in Proceedings of the 13th International Workshop on Mining Software Repositories - MSR 16. New York, New York, USA: ACM Press, 2016, pp. 351361.
  • 12. Extract Feature Vector (1/2) 12 f1 Author: Name of person who build this package. f2 Author Domain: Email domain of person who build this package. f3 License: License tell people know what organization that publish the package how they are permitted to use it. package.json
  • 13. Extract Feature Vector (2/2) 13 f4 Tagged Keywords: An array of strings that helps people discover your package as its listed in npm search. f5 Version Released: Version form an identifier that is assumed to be completely unique. f6 Number of Dependencies: The number of mapped package dependencies to a version range.package.json
  • 14. TDA High Dimension Mapping Dataset 14 f1 (Author) f2 (Author Domain) f3 (License) Knotter (TDA) Vector Space Model (VSM) f4 (Keywords) f5 (Version) f6 (No.dependencies) word2vec estimate version count dependencies package.json 151,000 packages https://github.com/rosinality/knotter
  • 16. Results The number of package dependencies is a strong feature in the topology 16
  • 17. Results Packages that are more likely to be used within ecosystem are located separately from packages meant for application usage outside the ecosystem 17
  • 18. Comparison with Archetypal Analysis 18 Archetype 1 (A1) has packages that contains keywords, such as web, plugin, test, http, express, node, api and server. Archetype 2 (A2) has packages that contains keywords like html, gulpplugin, css, javascript and gulp. Archetype 3 (A3) has lower packages compared to the other two archetypes.
  • 19. Comparison with Archetypal Analysis 19 Archet ype Identified Packages A1 tar-parse, turtle-run, marked-sanitized, haversort, bmxplayjs A2 statsd-influxdb-backend, ardeidae, demo-blog-system, git-ssb- web, social-media-resolver A3 stream-viz, programify, polyclay-couch, meshblu-core-task-check- update-device-is-valid, apidoc-almond
  • 20. 20