<![CDATA[Background: The rapid growth of web content has increased the complexity of retrieving relevant and high-quality information, especially in resource-constrained environments. Traditional keyword-based search engines often fail to capture semantic relationships and structural importance within web documents, leading to suboptimal retrieval performance.Aims: This study aims to develop a lightweight and modular web search engine, PyThinSearch, that integrates content-based and link-based ranking techniques to improve retrieval effectiveness and efficiency in low-resource and domain-specific environments.Method: The proposed system employs a hybrid ranking approach combining TF-IDF, PageRank, and HITS algorithms, along with anchor text analysis to enhance contextual relevance. The system is designed using a modular pipeline architecture consisting of data crawling, text preprocessing, indexing with inverted index, ranking, and query processing. Performance is evaluated using standard information retrieval metrics, including Precision, Recall, F1-score, Mean Average Precision (MAP), Normalized Discounted Cumulative Gain (NDCG), and response time.Result: The experimental results demonstrate that the hybrid ranking model consistently outperforms individual methods. The system achieves higher retrieval effectiveness, with improvements in Precision (0.78), Recall (0.75), MAP (0.77), and NDCG (0.80). Additionally, anchor text analysis significantly enhances performance in ambiguous queries, while the inverted index structure ensures efficient query response times suitable for small- to medium-scale datasets.Conclusion: PyThinSearch provides an effective and efficient solution for information retrieval by integrating textual relevance and structural importance within a lightweight and modular framework. The proposed system is well-suited for deployment in resource-constrained environments, although future work should focus on incorporating advanced NLP techniques and scalable architectures to improve performance in large-scale applications.]]>
