asmhatre/spring-ai
1
0
Fork 0
Code Activity

Add rerank module: bi-encoder retrieval plus an ONNX Runtime cross-encoder reranker behind a Spring AI DocumentPostProcessor; hit@k and MRR by pool size and question kind, latency by candidates, batch size, threads and int8, and five reproduced traps. Local MiniLM models, no database or API key.

Browse Source 
Co-Authored-By: Claude Sonnet 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01Cfo2K4u4rF18QsPXPFnytf
This commit is contained in:
Claude
2026-10-24 18:00:05 +00:00
parent 2ccf1d8d19
commit a429aa40f6
34 changed files with 1617 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
README.md
+1
View File
@@ -27,5 +27,6 @@ Runnable companion code for the Spring AI articles on [ankurm.com](https://ankur
| [`agent-skills/`](agent-skills) | `SKILL.md` folders loaded on demand by the `Skill` tool from `spring-ai-agent-utils` 0.13.0: what the model sees before it loads anything (name and description only), the two-call load round trip, unknown, duplicate and description-less skills, and listing size against inlining every body (character counts). Scripted chat model, no live model. Spring Boot 4.1.1, Spring AI 2.0.1, Java 25. | [Agent Skills in Spring AI: Packaging Reusable Capabilities for Your Agents](https://ankurm.com/agent-skills-spring-ai-packaging-reusable-capabilities/) |
| [`rag-ingestion/`](rag-ingestion) | PDF, Word and HTML versions of one 12-section runbook through `PagePdfDocumentReader`, `TikaDocumentReader` and `JsoupDocumentReader`, then `TokenTextSplitter` at six sizes, a paragraph chunker, a sentence splitter and a section chunker; retrieval hit@1/3/5 on 37 fixed questions, metadata in the embedding, content-hash idempotent ingestion into pgvector and the re-embedding cost of one edit. Local all-MiniLM-L6-v2; a synthetic corpus, so differences of two or three questions are noise. Spring Boot 4.1.2, Spring AI 2.0.2, Java 25. | [Spring AI RAG Ingestion Pipeline: PDF, Word and HTML Readers with Chunking Strategies Compared](https://ankurm.com/spring-ai-rag-ingestion-pipeline-pdf-word-html-chunking/) |
| [`hybrid-search/`](hybrid-search) | Vector, `tsvector` and a Java Okapi BM25 over one PostgreSQL table (pgvector HNSW + GIN), fused with Reciprocal Rank Fusion in Java and in a single SQL statement, behind a `DocumentRetriever` for `RetrievalAugmentationAdvisor`. 60 questions in four kinds over 94 passages: hybrid helps exact error codes and hurts paraphrases; `k` tuned on the evaluation set did not hold up on held-out folds; plans and latency at 94 and 50,094 rows. Local all-MiniLM-L6-v2. Spring Boot 4.1.2, Spring AI 2.0.2, Java 25. | [Hybrid Search in Java: BM25 + Vector with PostgreSQL and Reciprocal Rank Fusion](https://ankurm.com/hybrid-search-java-pgvector-tsvector-reciprocal-rank-fusion/) |
| [`rerank/`](rerank) | A bi-encoder (all-MiniLM-L6-v2) retrieves candidates, a cross-encoder (ms-marco-MiniLM-L-6-v2) rescores them, and a `DocumentPostProcessor` keeps the top 5 for `RetrievalAugmentationAdvisor`. Both models run in-process through ONNX Runtime 1.31.0 with DJL tokenizers 0.39.0 (no database, no API key). 60 questions over 94 passages: hit@1 41 to 49 and MRR 0.78 to 0.87, almost all of it on exact error codes, with paraphrases slightly worse; pool size, batch size, threads and int8 against latency on 2 vCPUs; and five traps reproduced (missing `token_type_ids`, swapped pair order, logits are not probabilities, truncation, a pool that is too small). Spring Boot 4.1.2, Spring AI 2.0.2, Java 25. | [Reranking RAG Results in Java: Cross-Encoder Reranker with ONNX Runtime Before the LLM](https://ankurm.com/reranking-rag-results-java-cross-encoder-onnx-runtime-spring-ai/) |
Upgrading from Spring AI 1.x: [migration guide](https://ankurm.com/spring-ai-1-to-2-migration-guide/).