Generating and maintaining API documentation with integrity and consistency can be time-consuming and expensive for evolving APIs. To solve this problem, several approaches have been proposed to automatically generate high-quality API documentation based on a combination of knowledge from different web sources. However, current researches are weak in handling unpopular APIs and cannot generate structured API documentation. Hence, in this poster, we propose a hybrid technique(namely \textit{gDoc}) for the automatic generation of structured API documentation. We first present a fine-grained search-based strategy to generate the description for partial API parameters via computing the relevance between various APIs, ensuring the consistency of API documentation. Then, we employ the cross-modal pretraining Seq2Seq model M6 to generate a structured API document for each API, which treats the document generation problem as a translation problem. Finally, we propose a heuristic algorithm to extract practical parameter examples from API request logs. The experiments evaluated on the online system show that this work's approach significantly improves the effectiveness and efficiency of API document generation.

翻译：为不断演进的API生成和维护具有完整性与一致性的文档可能既耗时又昂贵。为解决此问题，已有多种方法被提出，通过结合不同网络来源的知识来自动生成高质量的API文档。然而，当前的研究在处理冷门API时表现薄弱，且无法生成结构化的API文档。因此，在本海报中，我们提出了一种混合技术（即\textit{gDoc}）用于自动生成结构化API文档。我们首先提出一种基于细粒度搜索的策略，通过计算不同API之间的相关性来生成部分API参数的描述，以确保API文档的一致性。随后，我们采用跨模态预训练的Seq2Seq模型M6为每个API生成结构化API文档，将文档生成问题视为翻译问题。最后，我们提出一种启发式算法从API请求日志中提取实用的参数示例。在线系统上的实验评估表明，本研究的方法显著提升了API文档生成的有效性与效率。