refactor: 观测层双轴正交重构——spectra→observation、工具/API 收敛、安全韧性加固

将"以光谱为中心"的观测数据架构升级为 (Source × ProductType) 双轴正交模型,
光谱降级为与光变/测光/图像平级的产品类型之一;同步把分散的工具、API、缓存表
收敛为统一入口。新增 Gaia 光变曲线(EPOCH_PHOTOMETRY)支持。

【架构】services/spectra 整体替换为 services/observation(双轴正交)
- Source(LAMOST/Gaia/SDSS/DESI)× ProductType(Spectrum/LightCurve/Photometry/Image)
  正交组合,新增源/产品类型为纯加法(OCP)
- ObservationFetcher trait + ObservationRegistry:每个有效组合实现一个 fetcher,
  启动时注册;SDSS specobj/APOGEE 共用 key 按 subtype 二级路由
- cone 缓存逻辑模板方法化(trait 默认方法),消除各源 4 份重复代码
- 多文件 Artifact 模型:一个逻辑产物可含多文件(如 Gaia 光变 G/BP/RP 三波段各一 FITS)
- 统一编排 dispatch.rs:search(仅检索)/ download(检索+下载),支持坐标模式
  (cone→选源→下载)与标识符模式(直按 ID 下载)双输入

【Agent 工具整合】26 → 24
- 新增 find_observation:跨源×跨产品×双模式统一观测下载,取代 find_spectrum
- catalog_operation 升级为 6 合 1(search/describe/query/cone/export/lookup),
  取代独立的 query_vizier + cone_search
- citation_network + library_search 合并为 library.rs

【API 路由】
- 新增 /observation/{search,download,capabilities,list} 命名空间
- 移除 /catalog/{crossmatch,spectrum/download,spectrum/list}
- GET /observation/capabilities 暴露 registry 能力清单,前端动态渲染源/产品/版本
  下拉(不再硬编码各源支持矩阵)

【数据库迁移】
- 新表 observation_cache:新增 product 列 + artifacts_json(多文件产物),无 TTL
  (观测数据不可变,区别于 vizier_query_cache 的 7 天 TTL)
- 20260705140001:spectrum_cache 旧数据迁入 observation_cache,单文件→单元素 artifacts

【前端】
- 新 ObservationPanel(988 行):双视图(检索下载 / 缓存库),选项由 capabilities 动态生成
- 新 useObservation hook、ObservationResultCard、observation/constants、utils/apiError

【安全与韧性加固】
- sessions 锁 Mutex → RwLock(读多写少,降低争用)
- 新增 upload_rate_limiter;login_rate_limiter 容量保护(10000 上限,超限清最旧一半)
- bookmarklet API 密钥 SHA-1 → SHA-256;ADMIN_PASSWORD 长度上限 128
- *_TIMEOUT_SECS / EMBEDDING_DIM 非法值告警并回退默认;DB_POOL_SIZE 可配置(原硬编码 5)
- sqlite-vec 注册逻辑下沉至 utils::register_sqlite_vec_extension
This commit is contained in:
fmq
2026-07-07 01:34:02 +08:00
parent a156252bc3
commit 2f1fd19d74
71 changed files with 6229 additions and 4198 deletions
+339 -185
View File
@@ -1,9 +1,12 @@
// src/agent/tools/astro/research/vizier.rs
//
// QueryVizierTool —— VizieR TAP 星表查询(自由 ADQL + 便捷表查询)
// ConeSearchTool —— 锥形检索(按坐标查近邻天体)
//
// 对齐 QueryTargetTool 范式:单元结构体 + AgentTool 实现 + ctx.app_state 调 service
// CatalogOperationTool —— VizieR 星表统一操作工具
// search: 按关键词搜索星表目录
// describe: 查看表的列结构
// query: 执行 ADQL 或按表名查询,返回数据
// cone: 按坐标锥形检索
// export: 导出为 CSV 文件
// lookup: 通过 bibcode 查找关联数据表
use async_trait::async_trait;
use serde_json::json;
@@ -34,7 +37,6 @@ fn render_result_table(result: &VizierQueryResult, preview_rows: usize) -> Strin
return content;
}
// 表头
let headers: Vec<&str> = result.fields.iter().map(|f| f.name.as_str()).collect();
content.push_str(&format!("| {} |\n", headers.join(" | ")));
content.push_str(&format!(
@@ -46,7 +48,6 @@ fn render_result_table(result: &VizierQueryResult, preview_rows: usize) -> Strin
.join(" | ")
));
// 表体(限制预览行数)
let show = result.rows.len().min(preview_rows);
for row in result.rows.iter().take(show) {
let cells: Vec<String> = row
@@ -67,7 +68,6 @@ fn render_result_table(result: &VizierQueryResult, preview_rows: usize) -> Strin
));
}
// 列单位提示
let units: Vec<&str> = result
.fields
.iter()
@@ -80,161 +80,73 @@ fn render_result_table(result: &VizierQueryResult, preview_rows: usize) -> Strin
content
}
// ── QueryVizierTool ──
// ── CatalogOperationTool ──
pub struct QueryVizierTool;
pub struct CatalogOperationTool;
#[async_trait]
impl AgentTool for QueryVizierTool {
impl AgentTool for CatalogOperationTool {
fn name(&self) -> &str {
"query_vizier"
"catalog_operation"
}
fn description(&self) -> &str {
"通过 VizieR TAP 服务查询天文星表数据,支持两种模式:\
(1) 自由 ADQL 查询——传入 'adql' 参数执行标准 ADQL 语句;\
(2) 便捷表查询——传入 'table_name' + 可选 'columns' + 'limit' 直接取行。\
适用于:获取天体的精确测光/天体测量参数、查询星表中的近邻天体、交叉证认、批量拉取某类样本。\
结果自动缓存 7 天。不确定表名时先用 search_catalogs 搜索。\
ADQL 语法示例:SELECT TOP 10 ra, dec FROM \"I/355/gaiadr3\" WHERE parallax > 10"
"VizieR 星表统一操作。通过 action 选择:\n\
search — 按关键词搜索星表目录;\n\
describe — 查看表的列结构;\n\
query — 执行 ADQL 或按表名查询数据;\n\
cone — 按坐标锥形检索近邻天体;\n\
export — 下载表数据保存为 CSV;\n\
lookup — 通过 bibcode 查找关联数据表。"
}
fn parameters(&self) -> serde_json::Value {
json!({
"type": "object",
"properties": {
"adql": {
"action": {
"type": "string",
"description": "自由 ADQL 查询语句(与 table_name 二选一)。如 SELECT TOP 10 * FROM \"I/355/gaiadr3\""
"enum": ["search", "describe", "query", "cone", "export", "lookup"],
"description": "操作类型"
},
"table_name": {
"keyword": {
"type": "string",
"description": "VizieR 表名(便捷模式,与 adql 二选一),如 'I/355/gaiadr3'Gaia DR3)、'J/AJ/165/8/table2'"
},
"columns": {
"type": "string",
"description": "需要返回的列名(逗号分隔),为空时返回所有列"
},
"limit": {
"type": "integer",
"description": "最大返回行数(默认 20,上限 2000)",
"default": 20
}
}
})
}
fn group(&self) -> &str {
"as:research"
}
fn is_concurrency_safe(&self, _args: &serde_json::Value) -> bool {
true
}
fn is_readonly(&self) -> bool {
true
}
async fn execute(&self, args: serde_json::Value, ctx: &ToolContext) -> ToolOutput {
let state = &ctx.app_state;
// 解析参数:adql 优先,否则走 table_name 便捷模式
let result = if let Some(adql) = args.get("adql").and_then(|v| v.as_str()) {
let limit = args
.get("limit")
.and_then(|v| v.as_i64())
.unwrap_or(20)
.clamp(1, 2000);
info!(
"[QueryVizier] ADQL 查询 (limit={}): {}",
limit,
adql.chars().take(150).collect::<String>()
);
crate::services::cds::vizier::query_adql_cached(&state.db, &state.vizier, adql, limit)
.await
} else if let Some(table) = args.get("table_name").and_then(|v| v.as_str()) {
let limit = args
.get("limit")
.and_then(|v| v.as_i64())
.unwrap_or(20)
.clamp(1, 2000);
let columns: Vec<String> = args
.get("columns")
.and_then(|v| v.as_str())
.map(|c| c.split(',').map(|s| s.trim().to_string()).collect())
.unwrap_or_default();
info!("[QueryVizier] 表查询 table={} limit={}", table, limit);
crate::services::cds::vizier::query_table(
&state.db,
&state.vizier,
table,
&columns,
limit,
)
.await
} else {
return ToolOutput::error(
"需要提供 'adql'(自由 ADQL)或 'table_name'(便捷表查询)参数之一",
);
};
match result {
Ok(r) => {
let content = render_result_table(&r, 20);
ToolOutput::success(content, json!(r))
}
Err(e) => ToolOutput::error(format!("VizieR 查询失败: {}", e)),
}
}
}
// ── ConeSearchTool ──
pub struct ConeSearchTool;
#[async_trait]
impl AgentTool for ConeSearchTool {
fn name(&self) -> &str {
"cone_search"
}
fn description(&self) -> &str {
"锥形检索(Cone Search):按坐标在天文星表中检索近邻天体。\
需要指定目标星表(table 参数)。\
适用于:给定坐标找附近天体、获取某区域的星表数据、配合 query_target 解析名称后做区域查询。\
坐标系统为 J2000(ICRS),单位为度。结果自动缓存 7 天。\
不确定表名时先用 search_catalogs 搜索。"
}
fn parameters(&self) -> serde_json::Value {
json!({
"type": "object",
"properties": {
"ra": {
"type": "number",
"description": "赤经 RA(度,J2000/ICRS),范围 0~360"
},
"dec": {
"type": "number",
"description": "赤纬 Dec(度,J2000/ICRS),范围 -90~90"
},
"radius_deg": {
"type": "number",
"description": "检索半径(度),默认 0.1,范围 0~5",
"default": 0.1
"description": "search 时的搜索关键词,如 'Gaia DR3'"
},
"table": {
"type": "string",
"description": "目标星表(必填),如 'I/355/gaiadr3'Gaia DR3)、'II/246/out'2MASS"
"description": "VizieR 表名,describe/query/cone/export 时使用,如 'I/355/gaiadr3'"
},
"max_records": {
"adql": {
"type": "string",
"description": "query/export 时的 ADQL 语句(与 table 二选一)"
},
"columns": {
"type": "string",
"description": "query/export + table 模式下指定列(逗号分隔,默认 *)"
},
"coords": {
"type": "object",
"description": "cone 时的坐标参数",
"properties": {
"ra": { "type": "number", "description": "赤经 RA(度)" },
"dec": { "type": "number", "description": "赤纬 Dec(度)" },
"radius_deg": { "type": "number", "description": "检索半径(度),默认 0.1", "default": 0.1 },
"strategy": { "type": "string", "enum": ["nearest", "all"], "description": "nearest(默认)/ all", "default": "nearest" }
},
"required": ["ra", "dec"]
},
"limit": {
"type": "integer",
"description": "最大返回行数(默认 50,上限 2000)",
"default": 50
},
"bibcode": {
"type": "string",
"description": "lookup 时的 ADS bibcode"
}
},
"required": ["ra", "dec", "table"]
"required": ["action"]
})
}
@@ -247,59 +159,301 @@ impl AgentTool for ConeSearchTool {
}
fn is_readonly(&self) -> bool {
true
false
}
async fn execute(&self, args: serde_json::Value, ctx: &ToolContext) -> ToolOutput {
let action = match args.get("action").and_then(|v| v.as_str()) {
Some(a) => a,
None => return ToolOutput::error("缺少必需参数 'action'"),
};
let state = &ctx.app_state;
let ra = match args.get("ra").and_then(|v| v.as_f64()) {
Some(v) => v,
None => return ToolOutput::error("缺少必需参数 'ra'(赤经,度)"),
};
let dec = match args.get("dec").and_then(|v| v.as_f64()) {
Some(v) => v,
None => return ToolOutput::error("缺少必需参数 'dec'(赤纬,度)"),
};
let radius = args
.get("radius_deg")
.and_then(|v| v.as_f64())
.unwrap_or(0.1);
let table = match args.get("table").and_then(|v| v.as_str()) {
Some(t) => t,
None => return ToolOutput::error("缺少必需参数 'table'(目标星表)"),
};
let max_records = args
.get("max_records")
.and_then(|v| v.as_i64())
.unwrap_or(50)
.clamp(1, 2000);
info!(
"[ConeSearch] ra={} dec={} radius={}° table={}",
ra, dec, radius, table
);
match crate::services::cds::vizier::cone_search(
&state.db,
&state.vizier,
ra,
dec,
radius,
table,
max_records,
)
.await
{
Ok(r) => {
let mut content = format!(
"Cone Search 结果 (中心 ra={}, dec={}, radius={}°)\n\n",
ra, dec, radius
);
content.push_str(&render_result_table(&r, 20));
ToolOutput::success(content, json!(r))
}
Err(e) => ToolOutput::error(format!("Cone Search 失败: {}", e)),
match action {
"search" => do_catalog_search(state, &args).await,
"describe" => do_catalog_describe(state, &args).await,
"query" => do_catalog_query(state, &args).await,
"cone" => do_catalog_cone(state, &args).await,
"export" => do_catalog_export(state, &args).await,
"lookup" => do_catalog_lookup(state, &args).await,
other => ToolOutput::error(format!(
"未知 action '{}',可选: search/describe/query/cone/export/lookup",
other
)),
}
}
}
async fn do_catalog_search(
state: &crate::api::AppState,
args: &serde_json::Value,
) -> ToolOutput {
let keyword = match args.get("keyword").and_then(|v| v.as_str()) {
Some(k) => k,
None => return ToolOutput::error("search 需要 'keyword' 参数"),
};
let limit = args
.get("limit")
.and_then(|v| v.as_i64())
.unwrap_or(10)
.clamp(1, 30) as usize;
let catalog = crate::services::cds::vizier::VizierCatalog::new(&state.db, &state.vizier);
let results = match catalog.search(keyword, limit).await {
Ok(r) => r,
Err(e) => return ToolOutput::error(format!("搜索失败: {}", e)),
};
if results.is_empty() {
return ToolOutput::success(format!("未找到与 '{}' 匹配的星表", keyword), json!([]));
}
let mut content = format!(
"搜索 '{}' 匹配到 {} 个星表(按数据量降序):\n\n",
keyword,
results.len()
);
let items: Vec<serde_json::Value> = results
.iter()
.map(|entry| {
let nrows_str = entry.nrows.map(|n| format!("{}", n)).unwrap_or_else(|| "行数未知".into());
content.push_str(&format!("- `{}` — {} ({})\n", entry.table_name, entry.description, nrows_str));
json!({"table": entry.table_name, "description": entry.description, "nrows": entry.nrows})
})
.collect();
ToolOutput::success(content, json!(items))
}
async fn do_catalog_describe(
state: &crate::api::AppState,
args: &serde_json::Value,
) -> ToolOutput {
let table = match args.get("table").and_then(|v| v.as_str()) {
Some(t) => t,
None => return ToolOutput::error("describe 需要 'table' 参数"),
};
let catalog = crate::services::cds::vizier::VizierCatalog::new(&state.db, &state.vizier);
let columns = match catalog.describe(table).await {
Ok(c) => c,
Err(e) => return ToolOutput::error(format!("查询表结构失败: {}", e)),
};
if columns.is_empty() {
return ToolOutput::success(format!("表 '{}' 未找到列定义", table), json!([]));
}
let mut content = format!("表 `{}` 共 {} 列:\n\n", table, columns.len());
let items: Vec<serde_json::Value> = columns
.iter()
.map(|col| {
let unit_str = col.unit.as_deref().unwrap_or("");
let desc_str = col.description.as_deref().unwrap_or("");
content.push_str(&format!("- `{}` ({}) {}{}\n", col.column_name, col.datatype, unit_str, desc_str));
json!({"name": col.column_name, "datatype": col.datatype, "unit": col.unit, "description": col.description})
})
.collect();
ToolOutput::success(content, json!(items))
}
async fn do_catalog_query(
state: &crate::api::AppState,
args: &serde_json::Value,
) -> ToolOutput {
let limit = args
.get("limit")
.and_then(|v| v.as_i64())
.unwrap_or(50)
.clamp(1, 2000);
let result = if let Some(adql) = args.get("adql").and_then(|v| v.as_str()) {
info!(
"[CatalogOp:query] ADQL (limit={}): {}",
limit,
adql.chars().take(150).collect::<String>()
);
crate::services::cds::vizier::query_adql_cached(&state.db, &state.vizier, adql, limit)
.await
} else if let Some(table) = args.get("table").and_then(|v| v.as_str()) {
let columns: Vec<String> = args
.get("columns")
.and_then(|v| v.as_str())
.map(|c| c.split(',').map(|s| s.trim().to_string()).collect())
.unwrap_or_default();
info!("[CatalogOp:query] table={} limit={}", table, limit);
crate::services::cds::vizier::query_table(
&state.db,
&state.vizier,
table,
&columns,
limit,
)
.await
} else {
return ToolOutput::error("query 需要 'adql' 或 'table' 参数之一");
};
match result {
Ok(r) => {
let content = render_result_table(&r, 20);
ToolOutput::success(content, json!(r))
}
Err(e) => ToolOutput::error(format!("查询失败: {}", e)),
}
}
async fn do_catalog_cone(
state: &crate::api::AppState,
args: &serde_json::Value,
) -> ToolOutput {
let coords = match args.get("coords").and_then(|v| v.as_object()) {
Some(c) => c,
None => return ToolOutput::error("cone 需要 'coords' 参数(含 ra/dec"),
};
let ra = match coords.get("ra").and_then(|v| v.as_f64()) {
Some(v) => v,
None => return ToolOutput::error("coords 缺少 'ra'"),
};
let dec = match coords.get("dec").and_then(|v| v.as_f64()) {
Some(v) => v,
None => return ToolOutput::error("coords 缺少 'dec'"),
};
let radius = coords
.get("radius_deg")
.and_then(|v| v.as_f64())
.unwrap_or(0.1);
let nearest = coords.get("strategy").and_then(|v| v.as_str()) != Some("all");
let table = match args.get("table").and_then(|v| v.as_str()) {
Some(t) => t,
None => return ToolOutput::error("cone 需要 'table' 参数"),
};
let max_records = args
.get("limit")
.and_then(|v| v.as_i64())
.unwrap_or(50)
.clamp(1, 2000);
info!(
"[CatalogOp:cone] ra={} dec={} radius={}° table={} strategy={}",
ra, dec, radius, table, if nearest { "nearest" } else { "all" }
);
match crate::services::cds::vizier::cone_search(
&state.db,
&state.vizier,
ra,
dec,
radius,
table,
max_records,
nearest,
)
.await
{
Ok(r) => {
let mut content = format!(
"Cone Search 结果 (ra={}, dec={}, radius={}°)\n\n",
ra, dec, radius
);
content.push_str(&render_result_table(&r, 20));
ToolOutput::success(content, json!(r))
}
Err(e) => ToolOutput::error(format!("Cone Search 失败: {}", e)),
}
}
async fn do_catalog_export(
state: &crate::api::AppState,
args: &serde_json::Value,
) -> ToolOutput {
let limit = args
.get("limit")
.and_then(|v| v.as_i64())
.unwrap_or(100)
.clamp(1, 5000);
let adql = args.get("adql").and_then(|v| v.as_str());
let table = args.get("table").and_then(|v| v.as_str());
let columns = args.get("columns").and_then(|v| v.as_str());
if adql.is_none() && table.is_none() {
return ToolOutput::error("export 需要 'adql' 或 'table' 参数");
}
let catalog = crate::services::cds::vizier::VizierCatalog::new(&state.db, &state.vizier);
let result = match catalog.export_to_file(
state.config.library_dir.to_str().unwrap_or("."),
adql,
table,
columns,
limit,
).await {
Ok(r) => r,
Err(e) => return ToolOutput::error(format!("导出失败: {}", e)),
};
ToolOutput::success(
format!(
"已导出 {} 行数据到 `{}`\n文件大小: {}",
result.row_count,
result.path.display(),
result.size_bytes,
),
json!({
"path": result.path.to_str(),
"rows": result.row_count,
"columns": result.column_count,
}),
)
}
async fn do_catalog_lookup(
state: &crate::api::AppState,
args: &serde_json::Value,
) -> ToolOutput {
let bibcode = match args.get("bibcode").and_then(|v| v.as_str()) {
Some(b) => b,
None => return ToolOutput::error("lookup 需要 'bibcode' 参数"),
};
let limit = args
.get("limit")
.and_then(|v| v.as_i64())
.unwrap_or(10)
.clamp(1, 30) as usize;
let catalog = crate::services::cds::vizier::VizierCatalog::with_ads(
&state.db,
&state.vizier,
&state.ads,
);
let results = match catalog.lookup(bibcode, limit).await {
Ok(r) => r,
Err(e) => return ToolOutput::error(format!("查找失败: {}", e)),
};
if results.is_empty() {
return ToolOutput::success(
format!("文献 '{}' 未在 CDS/VizieR 中找到关联数据表", bibcode),
json!([]),
);
}
let mut content = format!(
"文献 '{}' 关联 {} 个 VizieR 数据表:\n\n",
bibcode,
results.len()
);
let items: Vec<serde_json::Value> = results
.iter()
.map(|entry| {
let nrows_str = entry.nrows.map(|n| format!("{}", n)).unwrap_or_default();
content.push_str(&format!("- `{}` — {} ({})\n", entry.table_name, entry.description, nrows_str));
json!({"table": entry.table_name, "description": entry.description, "nrows": entry.nrows})
})
.collect();
ToolOutput::success(content, json!(items))
}