Documentation-Guided API Sequence Search without Worrying about the Text-API Semantic Gap

Wei, Hongwei; Su, Xiaohong; Zheng, Weining; Tao, Wenxin

doi:10.1109/saner56733.2023.00040

Cited by 2 publications

(1 citation statement)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…API reference documentation is a type of developer documentation that provides detailed information about how to access and use the functions, classes, and methods exposed by an API (Wei et al, 2023). It acts as a definitive guide for software developers who need to integrate and work with the API, providing them with explanations of individual functions, data types, parameters required format, and endpoint URLs (S. Wang et al, 2023a).…”

Section: Developer Documentationmentioning

confidence: 99%

Technical communication in China : studies on the user experience of technical documentation

Gao

View full text Add to dashboard Cite

Conducting global technical communication requires strong language and communication skills, especially in English and other widely spoken languages.According to our findings, 62% of participants employed two or more languages in their daily work. English, French, German, Japanese, Russian, and Spanish are the most used foreign languages. 38% engaged in only one language, either Chinese or English. Geographic locationParticipants' workplaces were distributed in 13 provinces and municipalities (or 20 cities) of China's 34 provincial-level administrative regions. Two thirds of technical communicators lived in three cities: Beijing (26%), Shanghai (25%), and Shenzhen (15%). 90% of participants worked in China's eastern and southern coastal region, which is the most economically developed area in comparison to the country's central, western, and northeast regions. All 20 cities are municipalities, provincial capitals, or other big cities with thriving economies, diverse workforces, and high wages. Professional experience and work practiceWe studied participants' past and present practices of technical communication in the following aspects. Previous professionHalf of the participants (54%) began working in this field soon after graduating from college or graduate school. In other words, technical communicators were half of practitioners' first jobs. Among people who switched to technical communication from another field, we see a strong connection between their prior and current jobs. Engineers and translators ranked second (19%) and third (15%), respectively, as the most likely previous jobs. Other previous jobs include management and marketing specialists, product trainers, editors, consultants, English teachers, etc. It shows that most people's previous professions required dealing with various aspects of enterprise products or services, such as design, management, training, and document translation and editing, making the transition to technical communication easy.

show abstract

Section: Developer Documentationmentioning

confidence: 99%

Technical communication in China : studies on the user experience of technical documentation

Gao

View full text Add to dashboard Cite

show abstract

Deep API Sequence Generation via Golden Solution Samples and API Seeds

Huang,

Wang,

Wang

et al. 2024

ACM Trans. Softw. Eng. Methodol.

View full text Add to dashboard Cite

Automatic API recommendation can accelerate developers’ programming, and has been studied for years. There are two orthogonal lines of approaches for this task, i.e., information retrieval-based (IR-based) approaches and sequence to sequence (seq2seq) model based approaches. Although these approaches were reported to have remarkable performance, our observation finds two major drawbacks, i.e., IR-based approaches lack the consideration of relations among the recommended APIs, and seq2seq models do not model the API’s semantic meaning. To alleviate the above two problems, we propose APIGens, which is a retrieval-enhanced large language model (LLM) based API recommendation approach to recommend an API sequence for a natural language query. The approach first retrieves similar programming questions in history based on the input natural language query, and then scores the results based on API documents via a scorer model. Finally, these results are used as samples for few-shot learning of LLM. To reduce the risk of encountering local optima, we also extract API seeds from the retrieved results to increase the search scope during the LLM generation process. The results show that our approach can achieve 48.41% ROUGE@10 on API sequence recommendation and the 82.61% MAP on API set recommendation, largely outperforming the state-of-the-art baselines.

show abstract

Documentation-Guided API Sequence Search without Worrying about the Text-API Semantic Gap

Cited by 2 publications

References 28 publications

Technical communication in China : studies on the user experience of technical documentation

Technical communication in China : studies on the user experience of technical documentation

Deep API Sequence Generation via Golden Solution Samples and API Seeds

Contact Info

Product

Resources

About