Rodrigo Rodriguez
1a36f4aed2
- botserver/src/llm/local.rs: * Auto-configure embedding-url when empty (http://localhost:8082/v1/embeddings) * Auto-configure llm-url when empty (http://localhost:8081/v1/chat/completions) * Persist defaults to bot_configuration table via ConfigManager * Fix llama-server command path and add LD_LIBRARY_PATH * Fix working_dir for Linux (not just Windows) * Fix embedding server args: --embeddings --pooling mean --ctx-size 512 - botserver/src/core/bootstrap/bootstrap_manager.rs: * Temp disable alm-ci startup (causes 20s hang in bootstrap) * Remove unused alm_ci_health_check import - Database changes: * bot_configuration.embedding-model: 'bge-small-en-v1.5-f32.gguf' (full filename) * bot_configuration.embedding-url: 'http://localhost:8082/v1/embeddings' (auto-generated) - Testing: * Embedding server responds on port 8082 * Generates 384-dimension embeddings * KB file exists: minio://default.gbai/default.gbkb/manual/test-kb-manual.txt * Next: Verify KB indexing and search functionality Refs: #498 |
||
|---|---|---|
| .. | ||
| shared | ||
| ui_server | ||
| http_client.rs | ||
| lib.rs | ||
| main.rs | ||