feat(indexer): 优化 Milvus 索引器以支持批量处理和并发控制

[lonelymeko]

• 引入批次处理机制，提升大规模文档插入效率
• 添加并发控制功能，通过信号量限制最大并发数
• 修改默认配置，新增 BatchSize 和 MaxConcurrency 参数
• 调整 embedding 与数据转换逻辑，使其在批处理协程中执行
• 保留最终统一 Flush 操作，确保数据一致性
• 更新依赖项版本，引入 eino-ext 组件及相关库
• 升级多项间接依赖包，增强系统稳定性和兼容性
• 添加测试markdown 文件，在 milvus/examples/main.go引入md分割器，将分割后的docs交给新的 stroe 方法

What type of PR is this?

Check the PR title.

• [✅ ] This PR title match the format: <type>(optional scope): <description>
• [✅ ] The description of this PR title is user-oriented and clear enough for others to understand.
• Attach the PR updating the user documentation if the current PR requires user awareness at the usage level. User docs repo

(Optional) Translate the PR title into Chinese.

feat(milvus): 支持并发批量插入与性能优化

(Optional) More detailed description for this PR(en: English/zh: Chinese).

en: This PR introduces a high-performance batch processing mechanism for the Milvus Indexer, significantly improving the efficiency of large-scale document insertions.

Key Changes:

Batch Processing & Concurrency:
Introduced BatchSize and MaxConcurrency parameters in IndexerConfig.
Implemented internal batching logic in the Store method.
Added semaphore-based concurrency control to process Embedding, Conversion, and InsertRows in parallel goroutines.
Optimized the workflow to perform a unified Flush operation only after all batches are processed, avoiding performance bottlenecks caused by frequent flushing.
Refactoring:
Refactored the Store method to handle large document lists automatically, removing the need for callers to manually split data.
Dependency Upgrades:
Updated eino-ext and related indirect dependencies to enhance stability.
Examples & Testing:
Added examples/test.md (146 docs) and updated examples/main.go.
Introduced markdown splitter in the example to demonstrate the new capability.
Performance Comparison: Testing with examples/test.md (146 documents) on the same environment:

Before (Manual Serial Batching): ~2m 40s (160.3s)

After (Concurrent Batching): ~2.74s
Improvement: ~58x faster
Before (Before: Manual serial processing took 2m40s)

After (After: Concurrent processing took 2.74s)

zh(optional): 本 PR 为 Milvus Indexer 引入了高性能的批处理机制，极大提升了大规模文档插入的效率。

主要变更：

批处理与并发：
在 IndexerConfig 中新增了 BatchSize 和 MaxConcurrency 配置项。
在 Store 方法内部实现了自动分批逻辑。
添加了基于信号量的并发控制，支持并行执行 Embedding、数据转换和 InsertRows 操作。
优化了 Flush 逻辑，确保所有批次插入完成后统一执行一次 Flush，避免频繁 Flush 导致的性能阻塞。
重构：
重构了 Store 方法，调用方不再需要手动切分数据，可以直接传入大量文档。
依赖升级：
更新了 eino-ext 组件版本及相关间接依赖。
示例与测试：
新增 examples/test.md 测试文件（包含146个文档片段）并在 examples/main.go 中引入了 markdown 分割器进行验证。
性能对比： 使用 examples/test.md (146个文档片段) 在相同环境下测试：

优化前 (手动串行分批): 耗时约 2分40秒 (160.3s)

优化后 (内部并发分批): 耗时约 2.74秒

提升: 性能提升约 58 倍

(Optional) Which issue(s) this PR fixes:

Fixes #579

(optional) The PR that updates user documentation:

[CLAassistant]

All committers have signed the CLA.