milvus fts bm25

[Denise2004]

This PR adds complete full-text search feature support to VectorDBBench, enabling the benchmarking tool to evaluate Milvus's BM25 full-text search performance. This feature is based on the MS MARCO dataset and supports test cases of various scales.

Main Achievements

Adapted 3 FTS Performance Test Cases: Support for 100K, 5M, and 8.8M scale MS MARCO datasets. Only submitting 100K version here.

Complete FTS Dataset Management: Support for reading, parsing, and batch processing of TSV format data files

Milvus FTS Client Integration: Implement full-text document insertion and BM25 search functionality.

FTS-Specific Evaluation Metrics: Added calculation of Recall@K, NDCG@K, MRR and other metrics

Frontend Interface Support: Added FTS test case configuration and parameter settings in the Web UI

Core Features

1. Dataset Support

• Added FtsDatasetManager to manage MS MARCO datasets
• Support for TSV format collection, queries, qrels files

2. Milvus FTS Integration

• Implemented insert_fulltext() method: Support batch insertion of full-text documents
• Implemented search_fulltext() method: Full-text search based on BM25 algorithm
• Support for 8 configuration parameters: index algorithm, BM25 parameters, tokenizer, stop words, etc.

3. Test Execution Engine

• Added SerialFtsInsertRunner: FTS document insertion executor
• Extended SerialSearchRunner and MultiProcessingSearchRunner: Support FTS search testing
• Support for both serialized search and concurrent search modes

4. Evaluation System

• Added calc_recall_fts() and calc_ndcg_fts() functions
• Support for Recall@K, NDCG@K, MRR, QPS, Latency and other metrics

[sre-ci-robot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: Denise2004
To complete the pull request process, please assign xuanyang-cn after the PR has been reviewed.
You can assign the PR to them by writing /assign @xuanyang-cn in a comment when ready.

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these files:

• OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[zhuwenxing]

https://github.com/allenai/ir_datasets/

can we use this python lib to manage the datasets for fts? it contains most ir datasets!