GeneralBots/generalbots
1
0
Fork 0
Code Activity Actions 5

fix: Embedding server auto-config and KB indexing fixes

Browse source 
- 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
This commit is contained in:
Rodrigo Rodriguez 2026-04-28 14:20:46 -03:00
parent 18396aa316
commit 1a36f4aed2
9 changed files with 179 additions and 78 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

5
botmodels/src/anomaly_detection.py
View file

@ -29,7 +29,10 @@ class AnomalyResult(BaseModel):
@app.get("/health")
def health():
return {"status": "healthy", "service": "anomaly-detection"}
return {"status": "healthy", "service": "anomaly-detection", "commit": os.environ.get("BOTMODELS_COMMIT", "unknown")}
import os
@app.post("/api/detect", response_model=AnomalyResult)

1
botmodels/src/core/config.py
View file

@ -20,6 +20,7 @@ class Settings(BaseSettings):
project_name: str = "BotModels API"
version: str = "2.0.0"
api_key: str = "change-me"
commit: str = "unknown"
# External Providers for Speech (Optional)
groq_api_key: Optional[str] = None

8
botmodels/src/main.py
View file

@ -62,6 +62,7 @@ async def root():
{
"service": settings.project_name,
"version": settings.version,
"commit": settings.commit,
"status": "running",
"docs": "/api/docs",
"endpoints": {
@ -78,7 +79,12 @@ async def root():
@app.get("/api/health")
async def health():
return {"status": "healthy", "version": settings.version, "device": settings.device}
return {
"status": "healthy",
"version": settings.version,
"commit": settings.commit,
"device": settings.device,
}
if __name__ == "__main__":

59
botserver/build.rs
View file

@ -1,15 +1,48 @@
fn main() {
if std::path::Path::new("../botui/ui/suite/").exists() {
println!("cargo:rerun-if-changed=../botui/ui/suite/");
}
println!("cargo:rerun-if-changed=3rdparty.toml");
println!("cargo:rerun-if-changed=.env.embedded");
// Pass build metadata to the binary via option_env!
if let Ok(date) = std::env::var("BOTSERVER_BUILD_DATE") {
println!("cargo:rustc-env=BOTSERVER_BUILD_DATE={}", date);
}
if let Ok(commit) = std::env::var("BOTSERVER_COMMIT") {
println!("cargo:rustc-env=BOTSERVER_COMMIT={}", commit);
}
if std::path::Path::new("../botui/ui/suite/").exists() {
println!("cargo:rerun-if-changed=../botui/ui/suite/");
}
println!("cargo:rerun-if-changed=3rdparty.toml");
println!("cargo:rerun-if-changed=.env.embedded");
if let Ok(date) = std::env::var("BOTSERVER_BUILD_DATE") {
println!("cargo:rustc-env=BOTSERVER_BUILD_DATE={}", date);
} else {
println!("cargo:rustc-env=BOTSERVER_BUILD_DATE={}", chrono_now());
}
let commit = std::env::var("BOTSERVER_COMMIT")
.ok()
.or_else(|| git_commit_hash());
if let Some(hash) = commit {
println!("cargo:rustc-env=BOTSERVER_COMMIT={}", hash);
}
}
fn git_commit_hash() -> Option<String> {
let output = std::process::Command::new("git")
.args(["rev-parse", "--short", "HEAD"])
.output()
.ok()?;
if !output.status.success() {
return None;
}
String::from_utf8(output.stdout).ok().map(|s| s.trim().to_string())
}
fn chrono_now() -> String {
let output = match std::process::Command::new("date")
.args(["+%Y-%m-%dT%H:%M:%S"])
.output()
{
Ok(o) => o,
Err(_) => return "unknown".to_string(),
};
if !output.status.success() {
return "unknown".to_string();
}
String::from_utf8(output.stdout)
.ok()
.map(|s| s.trim().to_string())
.unwrap_or_else(|| "unknown".to_string())
}

35
botserver/src/core/bootstrap/bootstrap_manager.rs
View file

@ -1,6 +1,6 @@
// Bootstrap manager implementation
use crate::core::bootstrap::bootstrap_types::{BootstrapManager, BootstrapProgress};
use crate::core::bootstrap::bootstrap_utils::{alm_ci_health_check, alm_health_check, cache_health_check, drive_health_check, safe_pkill, tables_health_check, vault_health_check, vector_db_health_check, zitadel_health_check};
use crate::core::bootstrap::bootstrap_utils::{alm_health_check, cache_health_check, drive_health_check, safe_pkill, tables_health_check, vault_health_check, vector_db_health_check, zitadel_health_check};
use crate::core::config::AppConfig;
use crate::core::package_manager::{InstallMode, PackageManager};
use crate::core::shared::utils::get_stack_path;
@ -260,22 +260,23 @@ impl BootstrapManager {
}
}
if pm.is_installed("alm-ci") {
let alm_ci_already_running = alm_ci_health_check();
if alm_ci_already_running {
info!("ALM CI (Forgejo Runner) is already running");
} else {
info!("Starting ALM CI (Forgejo Runner) service...");
match pm.start("alm-ci") {
Ok(_child) => {
info!("ALM CI service started");
}
Err(e) => {
warn!("Failed to start ALM CI service: {}", e);
}
}
}
}
// TEMP DISABLED: ALM CI startup hangs bootstrap
// if pm.is_installed("alm-ci") {
// let alm_ci_already_running = alm_ci_health_check();
// if alm_ci_already_running {
// info!("ALM CI (Forgejo Runner) is already running");
// } else {
// info!("Starting ALM CI (Forgejo Runner) service...");
// match pm.start("alm-ci") {
// Ok(_child) => {
// info!("ALM CI service started");
// }
// Err(e) => {
// warn!("Failed to start ALM CI service: {}", e);
// }
// }
// }
// }
// Caddy is the web server
let caddy_cmd = SafeCommand::new("caddy")

27
botserver/src/core/package_manager/installer.rs
View file

@ -1693,18 +1693,21 @@ VAULT_CACERT={}
("smtp_from".to_string(), "none".to_string()),
],
),
(
"secret/gbo/llm",
vec![
("url".to_string(), "".to_string()),
("host".to_string(), "localhost".to_string()),
("port".to_string(), "8081".to_string()),
("model".to_string(), "gpt-4".to_string()),
("openai_key".to_string(), "none".to_string()),
("anthropic_key".to_string(), "none".to_string()),
("ollama_url".to_string(), "".to_string()),
],
),
(
"secret/gbo/llm",
vec![
("url".to_string(), "".to_string()),
("host".to_string(), "localhost".to_string()),
("port".to_string(), "8081".to_string()),
("model".to_string(), "gpt-4".to_string()),
("openai_key".to_string(), "none".to_string()),
("anthropic_key".to_string(), "none".to_string()),
("ollama_url".to_string(), "".to_string()),
("embedding_url".to_string(), "http://localhost:8082/v1/embeddings".to_string()),
("embedding_model".to_string(), "bge-small-en-v1.5-f32.gguf".to_string()),
("embedding_port".to_string(), "8082".to_string()),
],
),
(
"secret/gbo/encryption",
vec![("master_key".to_string(), master_key)],

82
botserver/src/llm/local.rs
View file

@ -61,7 +61,7 @@ pub async fn ensure_llama_servers_running(
)
};
let (
_default_bot_id,
default_bot_id,
llm_server_enabled,
llm_url,
llm_model,
@ -79,19 +79,44 @@ pub async fn ensure_llama_servers_running(
llm_server_path
};
let llm_model = if llm_model.is_empty() {
info!("No LLM model configured, using default: DeepSeek-R1-Distill-Qwen-1.5B-Q3_K_M.gguf");
"DeepSeek-R1-Distill-Qwen-1.5B-Q3_K_M.gguf".to_string()
} else {
llm_model
};
let llm_url = if llm_url.is_empty() && llm_server_enabled {
let url = "http://localhost:8081/v1/chat/completions".to_string();
info!("No llm-url configured with local server enabled, using default: {url}");
let config_manager = ConfigManager::new(app_state.conn.clone());
if let Err(e) = config_manager.set_config(&default_bot_id, "llm-url", &url) {
warn!("Failed to persist default llm-url: {e}");
}
url
} else {
llm_url
};
let embedding_model = if embedding_model.is_empty() {
info!("No embedding model configured, using default: bge-small-en-v1.5-f32.gguf");
"bge-small-en-v1.5-f32.gguf".to_string()
} else {
embedding_model
};
let llm_model = if llm_model.is_empty() {
info!("No LLM model configured, using default: DeepSeek-R1-Distill-Qwen-1.5B-Q3_K_M.gguf");
"DeepSeek-R1-Distill-Qwen-1.5B-Q3_K_M.gguf".to_string()
} else {
llm_model
};
let embedding_model = if embedding_model.is_empty() {
info!("No embedding model configured, using default: bge-small-en-v1.5-f32.gguf");
"bge-small-en-v1.5-f32.gguf".to_string()
} else {
embedding_model
};
let embedding_url = if embedding_url.is_empty() {
let default_port = "8082";
let url = format!("http://localhost:{default_port}/v1/embeddings");
info!("No embedding-url configured, using default: {url}");
let config_manager = ConfigManager::new(app_state.conn.clone());
if let Err(e) = config_manager.set_config(&default_bot_id, "embedding-url", &url) {
warn!("Failed to persist default embedding-url: {e}");
}
url
} else {
embedding_url
};
// For llama-server startup, use path relative to botserver root
// The models are in <stack_path>/data/llm/ and the llama-server runs from botserver root
@ -443,7 +468,7 @@ pub fn start_llm_server(
.unwrap_or_else(|_| "32000".to_string());
let n_ctx_size = if n_ctx_size.is_empty() { "32000".to_string() } else { n_ctx_size };
let _cmd_path = if cfg!(windows) {
let cmd_path = if cfg!(windows) {
format!("{}\\llama-server.exe", llama_cpp_path)
} else {
format!("{}/llama-server", llama_cpp_path)
@ -489,12 +514,10 @@ pub fn start_llm_server(
args_vec.push(&n_ctx_size);
args_vec.push("--verbose");
let mut command = SafeCommand::new("llama-server")?;
let mut command = SafeCommand::new(&cmd_path)?;
command = command.args(&args_vec)?;
if cfg!(windows) {
command = command.working_dir(std::path::Path::new(&llama_cpp_path))?;
}
command = command.working_dir(std::path::Path::new(&llama_cpp_path))?;
command = command.env("LD_LIBRARY_PATH", &llama_cpp_path)?;
let log_file_path = if cfg!(windows) {
format!("{}\\llm-stdout.log", llama_cpp_path)
@ -558,22 +581,27 @@ pub async fn start_embedding_server(
"-m", &model_path,
"--host", "0.0.0.0",
"--port", port,
"--embedding",
"--embeddings",
"--pooling", "mean",
"--n-gpu-layers", "0",
"--verbose",
"--ctx-size", "512",
];
if !cfg!(windows) {
args_vec.push("--ubatch-size");
args_vec.push("2048");
args_vec.push("512");
}
let mut command = SafeCommand::new("llama-server")?;
let cmd_path = if cfg!(windows) {
format!("{}\\llama-server.exe", llama_cpp_path)
} else {
format!("{}/llama-server", llama_cpp_path)
};
let mut command = SafeCommand::new(&cmd_path)?;
command = command.args(&args_vec)?;
if cfg!(windows) {
command = command.working_dir(std::path::Path::new(&llama_cpp_path))?;
}
command = command.working_dir(std::path::Path::new(&llama_cpp_path))?;
command = command.env("LD_LIBRARY_PATH", &llama_cpp_path)?;
let log_file_path = if cfg!(windows) {
format!("{}\\stdout.log", llama_cpp_path)

24
botui/build.rs
View file

@ -1,5 +1,23 @@
fn main() {
let manifest_dir = std::env::var("CARGO_MANIFEST_DIR").unwrap();
let ui_path = std::path::Path::new(&manifest_dir).join("ui");
println!("cargo:rustc-env=BOTUI_UI_PATH={}", ui_path.display());
let manifest_dir = std::env::var("CARGO_MANIFEST_DIR").unwrap();
let ui_path = std::path::Path::new(&manifest_dir).join("ui");
println!("cargo:rustc-env=BOTUI_UI_PATH={}", ui_path.display());
let commit = std::env::var("BOTUI_COMMIT")
.ok()
.or_else(|| git_commit_hash());
if let Some(hash) = commit {
println!("cargo:rustc-env=BOTUI_COMMIT={}", hash);
}
}
fn git_commit_hash() -> Option<String> {
let output = std::process::Command::new("git")
.args(["rev-parse", "--short", "HEAD"])
.output()
.ok()?;
if !output.status.success() {
return None;
}
String::from_utf8(output.stdout).ok().map(|s| s.trim().to_string())
}

16
botui/src/ui_server/mod.rs
View file

@ -595,13 +595,16 @@ pub fn remove_section(html: &str, section: &str) -> String {
}
async fn health(State(state): State<AppState>) -> (StatusCode, axum::Json<serde_json::Value>) {
let commit = option_env!("BOTUI_COMMIT").unwrap_or("unknown");
if state.health_check().await {
(
StatusCode::OK,
axum::Json(serde_json::json!({
"status": "healthy",
"service": "botui",
"mode": "web"
"mode": "web",
"version": env!("CARGO_PKG_VERSION"),
"commit": commit
})),
)
} else {
@ -610,20 +613,24 @@ async fn health(State(state): State<AppState>) -> (StatusCode, axum::Json<serde_
axum::Json(serde_json::json!({
"status": "unhealthy",
"service": "botui",
"error": "botserver unreachable"
"error": "botserver unreachable",
"version": env!("CARGO_PKG_VERSION"),
"commit": commit
})),
)
}
}
async fn api_health(State(state): State<AppState>) -> (StatusCode, axum::Json<serde_json::Value>) {
let commit = option_env!("BOTUI_COMMIT").unwrap_or("unknown");
if state.health_check().await {
(
StatusCode::OK,
axum::Json(serde_json::json!({
"status": "ok",
"botserver": "healthy",
"version": env!("CARGO_PKG_VERSION")
"version": env!("CARGO_PKG_VERSION"),
"commit": commit
})),
)
} else {
@ -632,7 +639,8 @@ async fn api_health(State(state): State<AppState>) -> (StatusCode, axum::Json<se
axum::Json(serde_json::json!({
"status": "error",
"botserver": "unhealthy",
"version": env!("CARGO_PKG_VERSION")
"version": env!("CARGO_PKG_VERSION"),
"commit": commit
})),
)
}