Refactor Orchestrator to DAG-based scheduler with Git Context merging

- Implemented `DagScheduler` in `workflow-orchestrator` to manage task dependencies and commit history.
- Added `vgcs.merge_commits` in `workflow-context` for smart 3-way merge of parallel task branches.
- Introduced generic `WorkflowTaskCommand` and `WorkflowTaskEvent` in `common-contracts`.
- Adapted `tushare-provider-service` with `generic_worker` to support Git-based context read/write.
- Updated NATS subjects to support wildcard routing for generic workflow commands.

crates/workflow-context/Cargo.toml

@@ -3,6 +3,10 @@ name = "workflow-context"
 version = "0.1.0"
 edition = "2024"
 
+[lib]
+name = "workflow_context"
+path = "src/lib.rs"
+
 [dependencies]
 git2 = { version = "0.18", features = ["vendored-openssl"] }
 sha2 = "0.10"
@@ -13,7 +17,7 @@ thiserror = "1.0"
 hex = "0.4"
 walkdir = "2.3"
 regex = "1.10"
-globset = "0.4.18"
+globset = "0.4"
 
 [dev-dependencies]
 tempfile = "3.8"

crates/workflow-context/src/traits.rs

@@ -17,6 +17,10 @@ pub trait ContextStore {
     
     /// Three-way merge (In-Memory), returns new Tree OID
     fn merge_trees(&self, req_id: &str, base: &str, ours: &str, theirs: &str) -> Result<String>;
+
+    /// Smart merge two commits, automatically finding the best common ancestor.
+    /// Returns the OID of the new merge commit.
+    fn merge_commits(&self, req_id: &str, our_commit: &str, their_commit: &str) -> Result<String>;
     
     /// Start a write transaction
     fn begin_transaction(&self, req_id: &str, base_commit: &str) -> Result<Box<dyn Transaction>>;

crates/workflow-context/src/vgcs.rs

@@ -150,23 +150,72 @@ impl ContextStore for Vgcs {
         Ok(oid.to_string())
     }
 
+    fn merge_commits(&self, req_id: &str, our_commit: &str, their_commit: &str) -> Result<String> {
+        let repo_path = self.get_repo_path(req_id);
+        let repo = Repository::open(&repo_path).context("Failed to open repo")?;
+
+        let our_oid = Oid::from_str(our_commit).context("Invalid our_commit")?;
+        let their_oid = Oid::from_str(their_commit).context("Invalid their_commit")?;
+
+        let base_oid = repo.merge_base(our_oid, their_oid).context("Failed to find merge base")?;
+        
+        let base_commit = repo.find_commit(base_oid)?;
+        let our_commit_obj = repo.find_commit(our_oid)?;
+        let their_commit_obj = repo.find_commit(their_oid)?;
+
+        // If base equals one of the commits, it's a fast-forward
+        if base_oid == our_oid {
+            return Ok(their_commit.to_string());
+        }
+        if base_oid == their_oid {
+            return Ok(our_commit.to_string());
+        }
+
+        let base_tree = base_commit.tree()?;
+        let our_tree = our_commit_obj.tree()?;
+        let their_tree = their_commit_obj.tree()?;
+
+        let mut index = repo.merge_trees(&base_tree, &our_tree, &their_tree, None)?;
+
+        if index.has_conflicts() {
+             return Err(anyhow!("Merge conflict detected between {} and {}", our_commit, their_commit));
+        }
+
+        let tree_oid = index.write_tree_to(&repo)?;
+        let tree = repo.find_tree(tree_oid)?;
+
+        let sig = Signature::now("vgcs-merge", "system")?;
+
+        let merge_commit_oid = repo.commit(
+            None, // Detached
+            &sig,
+            &sig,
+            &format!("Merge commit {} into {}", their_commit, our_commit),
+            &tree,
+            &[&our_commit_obj, &their_commit_obj],
+        )?;
+
+        Ok(merge_commit_oid.to_string())
+    }
+
     fn begin_transaction(&self, req_id: &str, base_commit: &str) -> Result<Box<dyn Transaction>> {
         let repo_path = self.get_repo_path(req_id);
         let repo = Repository::open(&repo_path).context("Failed to open repo")?;
         
-        let base_oid = Oid::from_str(base_commit).context("Invalid base_commit")?;
-        
         let mut index = Index::new()?;
         let mut base_commit_oid = None;
 
-        if !base_oid.is_zero() {
-            // Scope the borrow of repo
-            {
-                let commit = repo.find_commit(base_oid).context("Base commit not found")?;
-                let tree = commit.tree()?;
-                index.read_tree(&tree)?;
+        if !base_commit.is_empty() {
+            let base_oid = Oid::from_str(base_commit).context("Invalid base_commit")?;
+            if !base_oid.is_zero() {
+                // Scope the borrow of repo
+                {
+                    let commit = repo.find_commit(base_oid).context("Base commit not found")?;
+                    let tree = commit.tree()?;
+                    index.read_tree(&tree)?;
+                }
+                base_commit_oid = Some(base_oid);
             }
-            base_commit_oid = Some(base_oid);
         }
         
         Ok(Box::new(VgcsTransaction {

docs/3_project_management/tasks/pending/20251126_design_4_orchestrator.md

@@ -0,0 +1,117 @@
+# 设计方案 4: Workflow Orchestrator (调度器) [详细设计版]
+
+## 1. 定位与目标
+
+Orchestrator 负责 DAG 调度、RPC 分发和冲突合并。它使用 Rust 实现。
+
+**核心原则：**
+*   **Git as Source of Truth**: 所有的任务产出（数据、报告）均存储在 VGCS (Git) 中。
+*   **数据库轻量化**: 数据库仅用于存储系统配置、TimeSeries 缓存以及 Request ID 与 Git Commit Hash 的映射。Workflow 执行过程中不依赖数据库进行状态流转。
+*   **Context 隔离**: 每次 Workflow 启动均为全新的 Context（或基于特定 Snapshot），无全局共享 Context。
+
+## 2. 调度逻辑规范
+
+### 2.1 RPC Subject 命名
+基于 NATS。
+*   **Command Topic**: `workflow.cmd.{routing_key}`
+    *   e.g., `workflow.cmd.provider.tushare`
+    *   e.g., `workflow.cmd.analysis.report`
+*   **Event Topic**: `workflow.evt.task_completed` (统一监听)
+
+### 2.2 Command Payload Schema
+
+```json
+{
+  "request_id": "uuid-v4",
+  "task_id": "fetch_tushare",
+  "routing_key": "provider.tushare",
+  "config": {
+    "market": "CN",
+    "years": 5
+    // Provider Specific Config
+  },
+  "context": {
+    "base_commit": "a1b2c3d4...", // Empty for initial task
+    "mount_path": "/Raw Data/Tushare" 
+    // 指示 Worker 建议把结果挂在哪里
+  },
+  "storage": {
+    "root_path": "/mnt/workflow_data"
+  }
+}
+```
+
+### 2.3 Event Payload Schema
+
+```json
+{
+  "request_id": "uuid-v4",
+  "task_id": "fetch_tushare",
+  "status": "Completed",
+  "result": {
+    "new_commit": "e5f6g7h8...",
+    "error": null
+  }
+}
+```
+
+## 3. 合并策略 (Merge Strategy) 实现
+
+### 3.1 串行合并 (Fast-Forward)
+DAG: A -> B
+1.  A returns `C1`.
+2.  Orchestrator dispatch B with `base_commit = C1`.
+
+### 3.2 并行合并 (Three-Way Merge)
+DAG: A -> B, A -> C. (B, C parallel)
+1.  Dispatch B with `base_commit = C1`.
+2.  Dispatch C with `base_commit = C1`.
+3.  B returns `C2`. C returns `C3`.
+4.  **Merge Action**:
+    *   Wait for BOTH B and C to complete.
+    *   Call `VGCS.merge_trees(base=C1, ours=C2, theirs=C3)`.
+    *   Result: `TreeHash T4`.
+    *   Create Commit `C4` (Parents: C2, C3).
+5.  Dispatch D (dependent on B, C) with `base_commit = C4`.
+
+### 3.3 冲突处理
+如果 `VGCS.merge_trees` 返回 Error (Conflict):
+1.  Orchestrator 捕获错误。
+2.  标记 Workflow 状态为 `Conflict`.
+3.  (Future) 触发 `ConflictResolver` Agent，传入 C1, C2, C3。Agent 生成 C4。
+
+## 4. 状态机重构
+废弃旧的 `WorkflowStateMachine` 中关于 TaskType 的判断。
+引入 `CommitTracker`:
+```rust
+struct CommitTracker {
+    // 记录每个任务产出的 Commit
+    task_commits: HashMap<TaskId, String>,
+    // 记录当前主分支的 Commit (Latest Merged)
+    head_commit: String,
+}
+```
+
+## 5. 执行计划
+1.  **Contract**: 定义 Rust Structs for RPC Payloads.
+2.  **NATS**: 实现 Publisher/Subscriber。
+3.  **Engine**: 实现 Merge Loop。
+
+## 6. 数据持久化与缓存策略 (Persistence & Caching)
+
+### 6.1 数据库角色 (Database Role)
+数据库不再作为业务数据的“主存储”，其角色转变为：
+1.  **Configuration**: 存储系统运行所需的配置信息。
+2.  **Cache (Hot Data)**: 缓存 Data Provider 抓取的原始数据 (Time-series)，避免重复调用外部 API。
+3.  **Index**: 存储 `request_id` -> `final_commit_hash` 的映射，作为系统快照的索引。
+
+### 6.2 Provider 行为模式
+Provider 在接收到 Workflow Command 时：
+1.  **Check Cache**: 检查本地 DB/Cache 是否有有效数据。
+2.  **Fetch (If miss)**: 如果缓存未命中，调用外部 API 获取数据并更新缓存。
+3.  **Inject to Context**: 将数据写入当前的 VGCS Context (via `WorkerRuntime`)，生成新的 Commit。
+    *   *注意*: Provider 不直接将此次 Workflow 的结果“存”回数据库的业务表，数据库仅作 Cache 用。
+
+### 6.3 Orchestrator 行为
+Orchestrator 仅负责追踪 Commit Hash 的演变。
+Workflow 结束时，Orchestrator 将最终的 `Head Commit Hash` 关联到 `Request ID` 并持久化（即“Snapshot 落盘”）。

services/common-contracts/src/lib.rs

@@ -1,4 +1,5 @@
 pub mod dtos;
+#[cfg(feature = "persistence")]
 pub mod models;
 pub mod observability;
 pub mod messages;
@@ -9,3 +10,6 @@ pub mod registry;
 pub mod lifecycle;
 pub mod symbol_utils;
 pub mod persistence_client;
+pub mod abstraction;
+pub mod workflow_harness; // Export the harness
+pub mod workflow_types;

services/common-contracts/src/subjects.rs

@@ -10,6 +10,11 @@ pub trait SubjectMessage: Serialize + DeserializeOwned + Send + Sync {
 
 #[derive(Debug, Clone, PartialEq, Eq)]
 pub enum NatsSubject {
+    // --- Workflow Generic ---
+    WorkflowCommand(String), // Dynamic routing key: workflow.cmd.{routing_key}
+    WorkflowEventTaskCompleted, // workflow.evt.task_completed
+    WorkflowCommandWildcard, // workflow.cmd.>
+
     // --- Commands ---
     WorkflowCommandStart,
     WorkflowCommandSyncState,
@@ -37,6 +42,9 @@ pub enum NatsSubject {
 impl fmt::Display for NatsSubject {
     fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
         match self {
+            Self::WorkflowCommand(key) => write!(f, "workflow.cmd.{}", key),
+            Self::WorkflowEventTaskCompleted => write!(f, "workflow.evt.task_completed"),
+            Self::WorkflowCommandWildcard => write!(f, "workflow.cmd.>"),
             Self::WorkflowCommandStart => write!(f, "workflow.commands.start"),
             Self::WorkflowCommandSyncState => write!(f, "workflow.commands.sync_state"),
             Self::DataFetchCommands => write!(f, "data_fetch_commands"),
@@ -58,6 +66,8 @@ impl FromStr for NatsSubject {
 
     fn from_str(s: &str) -> Result<Self, Self::Err> {
         match s {
+            "workflow.evt.task_completed" => Ok(Self::WorkflowEventTaskCompleted),
+            "workflow.cmd.>" => Ok(Self::WorkflowCommandWildcard),
             "workflow.commands.start" => Ok(Self::WorkflowCommandStart),
             "workflow.commands.sync_state" => Ok(Self::WorkflowCommandSyncState),
             "data_fetch_commands" => Ok(Self::DataFetchCommands),
@@ -76,6 +86,10 @@ impl FromStr for NatsSubject {
                         return Ok(Self::WorkflowProgress(uuid));
                     }
                 }
+                if s.starts_with("workflow.cmd.") {
+                    let key = s.trim_start_matches("workflow.cmd.");
+                    return Ok(Self::WorkflowCommand(key.to_string()));
+                }
                 Err(format!("Unknown or invalid subject: {}", s))
             }
         }

services/common-contracts/src/workflow_types.rs

@@ -0,0 +1,110 @@
+use uuid::Uuid;
+use service_kit::api_dto;
+use serde::{Serialize, Deserialize};
+use crate::subjects::{NatsSubject, SubjectMessage};
+
+// --- Enums ---
+
+#[api_dto]
+#[derive(Copy, PartialEq, Eq, Hash)]
+pub enum WorkflowStatus {
+    Pending,
+    Running,
+    Completed,
+    Failed,
+    Cancelled,
+}
+
+#[api_dto]
+#[derive(Copy, PartialEq, Eq, Hash)]
+pub enum TaskStatus {
+    Pending,
+    Scheduled,
+    Running,
+    Completed,
+    Failed,
+    Skipped,
+    Cancelled,
+}
+
+// --- Data Structures ---
+
+/// Context information required for a worker to execute a task.
+/// This tells the worker "where" to work (Git Commit).
+#[api_dto]
+pub struct TaskContext {
+    /// The Git commit hash that this task should base its work on.
+    /// If None/Empty, it implies starting from an empty repo (or initial state).
+    pub base_commit: Option<String>,
+    
+    /// Suggested path where the worker should mount/write its output.
+    /// This is a hint; strict adherence depends on the worker implementation.
+    pub mount_path: Option<String>,
+}
+
+/// Configuration related to storage access.
+#[api_dto]
+pub struct StorageConfig {
+    /// The root path on the host/container where the shared volume is mounted.
+    /// e.g., "/mnt/workflow_data"
+    pub root_path: String,
+}
+
+// --- Commands ---
+
+/// A generic command sent by the Orchestrator to a Worker.
+/// The `routing_key` in the subject determines which worker receives it.
+#[api_dto]
+pub struct WorkflowTaskCommand {
+    pub request_id: Uuid,
+    pub task_id: String,
+    pub routing_key: String,
+    
+    /// Dynamic configuration specific to the worker (e.g., {"market": "CN", "years": 5}).
+    /// The worker is responsible for parsing this.
+    pub config: serde_json::Value,
+    
+    pub context: TaskContext,
+    pub storage: StorageConfig,
+}
+
+impl SubjectMessage for WorkflowTaskCommand {
+    fn subject(&self) -> NatsSubject {
+        // The actual subject is dynamic based on routing_key.
+        // This return value is primarily for default behavior or matching.
+        NatsSubject::WorkflowCommand(self.routing_key.clone())
+    }
+}
+
+// --- Events ---
+
+/// A generic event sent by a Worker back to the Orchestrator upon task completion (or failure).
+#[api_dto]
+pub struct WorkflowTaskEvent {
+    pub request_id: Uuid,
+    pub task_id: String,
+    pub status: TaskStatus,
+    
+    /// The result of the task execution.
+    pub result: Option<TaskResult>,
+}
+
+impl SubjectMessage for WorkflowTaskEvent {
+    fn subject(&self) -> NatsSubject {
+        NatsSubject::WorkflowEventTaskCompleted
+    }
+}
+
+#[api_dto]
+pub struct TaskResult {
+    /// The new Git commit hash generated by this task.
+    /// Should be None if the task failed.
+    pub new_commit: Option<String>,
+    
+    /// Error message if the task failed.
+    pub error: Option<String>,
+    
+    /// Optional metadata or summary of the result (not the full data).
+    pub summary: Option<serde_json::Value>,
+}
+

services/tushare-provider-service/Cargo.toml

@@ -4,14 +4,16 @@ version = "0.1.0"
 edition = "2024"
 
 [dependencies]
-common-contracts = { path = "../common-contracts" }
+async-trait = "0.1.89"
+secrecy = { version = "0.8", features = ["serde"] }
+common-contracts = { path = "../common-contracts", default-features = false }
+workflow-context = { path = "../../crates/workflow-context" }
 
 anyhow = "1.0"
 async-nats = "0.45.0"
 axum = "0.8"
 config = "0.15.19"
 dashmap = "6.1.0"
-futures = "0.3"
 futures-util = "0.3.31"
 serde = { version = "1.0", features = ["derive"] }
 serde_json = "1.0"
@@ -23,11 +25,9 @@ uuid = { version = "1.6", features = ["v4", "serde"] }
 reqwest = { version = "0.12.24", features = ["json"] }
 url = "2.5.2"
 thiserror = "2.0.17"
-async-trait = "0.1.80"
 lazy_static = "1.5.0"
 regex = "1.10.4"
 chrono = "0.4.38"
 rust_decimal = "1.35.0"
 rust_decimal_macros = "1.35.0"
 itertools = "0.14.0"
-secrecy = { version = "0.8", features = ["serde"] }

services/tushare-provider-service/src/generic_worker.rs

@@ -0,0 +1,131 @@
+use anyhow::{Result, anyhow, Context};
+use tracing::{info, error, warn};
+use common_contracts::workflow_types::{WorkflowTaskCommand, WorkflowTaskEvent, TaskStatus, TaskResult};
+use common_contracts::subjects::{NatsSubject, SubjectMessage};
+use common_contracts::dtos::{CompanyProfileDto, TimeSeriesFinancialDto};
+use workflow_context::{WorkerContext, OutputFormat};
+use crate::state::AppState;
+use crate::tushare::TushareDataProvider;
+use serde_json::json;
+use std::sync::Arc;
+
+pub async fn handle_workflow_command(state: AppState, nats: async_nats::Client, cmd: WorkflowTaskCommand) -> Result<()> {
+    info!("Processing generic workflow command: task_id={}", cmd.task_id);
+
+    // 1. Parse Config
+    let symbol_code = cmd.config.get("symbol").and_then(|s| s.as_str()).unwrap_or("").to_string();
+    let market = cmd.config.get("market").and_then(|s| s.as_str()).unwrap_or("CN").to_string();
+    
+    if symbol_code.is_empty() {
+        return send_failure(&nats, &cmd, "Missing symbol in config").await;
+    }
+
+    // 2. Initialize Worker Context
+    // Note: We use the provided base_commit. If it's empty, it means start from scratch (or empty repo).
+    // We need to mount the volume.
+    let root_path = cmd.storage.root_path.clone();
+    
+    let mut ctx = match WorkerContext::init(&cmd.request_id.to_string(), &root_path, cmd.context.base_commit.as_deref()) {
+        Ok(c) => c,
+        Err(e) => return send_failure(&nats, &cmd, &format!("Failed to init context: {}", e)).await,
+    };
+
+    // 3. Fetch Data (with Cache)
+    let fetch_result = fetch_and_cache(&state, &symbol_code, &market).await;
+    
+    let (profile, financials) = match fetch_result {
+        Ok(data) => data,
+        Err(e) => return send_failure(&nats, &cmd, &format!("Fetch failed: {}", e)).await,
+    };
+
+    // 4. Write to VGCS
+    // Organize data in a structured way
+    let base_dir = format!("raw/tushare/{}", symbol_code);
+    
+    if let Err(e) = ctx.write_file(&format!("{}/profile.json", base_dir), &profile, OutputFormat::Json) {
+        return send_failure(&nats, &cmd, &format!("Failed to write profile: {}", e)).await;
+    }
+    
+    if let Err(e) = ctx.write_file(&format!("{}/financials.json", base_dir), &financials, OutputFormat::Json) {
+        return send_failure(&nats, &cmd, &format!("Failed to write financials: {}", e)).await;
+    }
+
+    // 5. Commit
+    let new_commit = match ctx.commit(&format!("Fetched Tushare data for {}", symbol_code)) {
+        Ok(c) => c,
+        Err(e) => return send_failure(&nats, &cmd, &format!("Commit failed: {}", e)).await,
+    };
+    
+    info!("Task {} completed. New commit: {}", cmd.task_id, new_commit);
+
+    // 6. Send Success Event
+    let event = WorkflowTaskEvent {
+        request_id: cmd.request_id,
+        task_id: cmd.task_id,
+        status: TaskStatus::Completed,
+        result: Some(TaskResult {
+            new_commit: Some(new_commit),
+            error: None,
+            summary: Some(json!({
+                "symbol": symbol_code,
+                "records": financials.len()
+            })),
+        }),
+    };
+    
+    publish_event(&nats, event).await
+}
+
+async fn fetch_and_cache(state: &AppState, symbol: &str, _market: &str) -> Result<(CompanyProfileDto, Vec<TimeSeriesFinancialDto>)> {
+    // 1. Get Provider (which holds the API token)
+    let provider = state.get_provider().await
+        .ok_or_else(|| anyhow!("Tushare Provider not initialized (missing API Token?)"))?;
+
+    // 2. Call fetch
+    let (profile, financials) = provider.fetch_all_data(symbol).await
+        .context("Failed to fetch data from Tushare")?;
+        
+    // 3. Write to DB Cache
+    // Note: PersistenceClient is not directly in AppState struct definition in `state.rs` I read.
+    // Let's check `state.rs` again. It implements TaskState which has `get_persistence_url`.
+    // We should instantiate PersistenceClient on the fly or add it to AppState.
+    
+    // For now, let's create a client on the fly to avoid changing AppState struct everywhere.
+    use common_contracts::persistence_client::PersistenceClient;
+    use common_contracts::workflow_harness::TaskState; // For get_persistence_url
+    
+    let persistence_url = state.get_persistence_url();
+    let p_client = PersistenceClient::new(persistence_url);
+    
+    if let Err(e) = p_client.save_company_profile(&profile).await {
+        warn!("Failed to cache company profile: {}", e);
+    }
+    
+    // Batch save financials logic is missing in PersistenceClient (based on context).
+    // If it existed, we would call it here.
+    
+    Ok((profile, financials))
+}
+
+async fn send_failure(nats: &async_nats::Client, cmd: &WorkflowTaskCommand, error_msg: &str) -> Result<()> {
+    error!("Task {} failed: {}", cmd.task_id, error_msg);
+    let event = WorkflowTaskEvent {
+        request_id: cmd.request_id,
+        task_id: cmd.task_id.clone(),
+        status: TaskStatus::Failed,
+        result: Some(TaskResult {
+            new_commit: None,
+            error: Some(error_msg.to_string()),
+            summary: None,
+        }),
+    };
+    publish_event(nats, event).await
+}
+
+async fn publish_event(nats: &async_nats::Client, event: WorkflowTaskEvent) -> Result<()> {
+    let subject = event.subject().to_string();
+    let payload = serde_json::to_vec(&event)?;
+    nats.publish(subject, payload.into()).await?;
+    Ok(())
+}
+

services/tushare-provider-service/src/main.rs

@@ -15,7 +15,7 @@ use crate::error::{Result, AppError};
 use crate::state::AppState;
 use tracing::{info, warn};
 use common_contracts::lifecycle::ServiceRegistrar;
-use common_contracts::registry::ServiceRegistration;
+use common_contracts::registry::{ServiceRegistration, ProviderMetadata, ConfigFieldSchema, FieldType, ConfigKey};
 use std::sync::Arc;
 
 #[tokio::main]
@@ -52,6 +52,26 @@ async fn main() -> Result<()> {
             role: common_contracts::registry::ServiceRole::DataProvider,
             base_url: format!("http://{}:{}", config.service_host, port),
             health_check_url: format!("http://{}:{}/health", config.service_host, port),
+            metadata: Some(ProviderMetadata {
+                id: "tushare".to_string(),
+                name_en: "Tushare Pro".to_string(),
+                name_cn: "Tushare Pro (中国股市)".to_string(),
+                description: "Official Tushare Data Provider".to_string(),
+                icon_url: None,
+                config_schema: vec![
+                    ConfigFieldSchema {
+                        key: ConfigKey::ApiToken,
+                        label: "API Token".to_string(),
+                        field_type: FieldType::Password,
+                        required: true,
+                        placeholder: Some("Enter your token...".to_string()),
+                        default_value: None,
+                        description: Some("Get it from https://tushare.pro".to_string()),
+                        options: None,
+                    },
+                ],
+                supports_test_connection: true,
+            }),
         }
     );
 

services/tushare-provider-service/src/message_consumer.rs

@@ -1,6 +1,7 @@
 use crate::error::Result;
 use crate::state::{AppState, ServiceOperationalStatus};
 use common_contracts::messages::FetchCompanyDataCommand;
+use common_contracts::workflow_types::WorkflowTaskCommand; // Import
 use common_contracts::observability::ObservabilityTaskStatus;
 use common_contracts::subjects::NatsSubject;
 use futures_util::StreamExt;
@@ -29,9 +30,16 @@ pub async fn run(state: AppState) -> Result<()> {
         match async_nats::connect(&state.config.nats_addr).await {
             Ok(client) => {
                 info!("Successfully connected to NATS.");
-                if let Err(e) = subscribe_and_process(state.clone(), client).await {
-                    error!("NATS subscription error: {}. Reconnecting in 10s...", e);
+                // Use try_join or spawn multiple subscribers
+                let s1 = subscribe_legacy(state.clone(), client.clone());
+                let s2 = subscribe_workflow(state.clone(), client.clone());
+                
+                tokio::select! {
+                    res = s1 => if let Err(e) = res { error!("Legacy subscriber error: {}", e); },
+                    res = s2 => if let Err(e) = res { error!("Workflow subscriber error: {}", e); },
                 }
+                
+                // If any subscriber exits, wait and retry
             }
             Err(e) => {
                 error!("Failed to connect to NATS: {}. Retrying in 10s...", e);
@@ -41,14 +49,39 @@ pub async fn run(state: AppState) -> Result<()> {
     }
 }
 
-async fn subscribe_and_process(
+async fn subscribe_workflow(state: AppState, client: async_nats::Client) -> Result<()> {
+    let subject = "workflow.cmd.provider.tushare".to_string();
+    let mut subscriber = client.subscribe(subject.clone()).await?;
+    info!("Workflow Consumer started on '{}'", subject);
+
+    while let Some(message) = subscriber.next().await {
+        // Check status check (omitted for brevity, assuming handled)
+        
+        let state = state.clone();
+        let client = client.clone();
+        
+        tokio::spawn(async move {
+            match serde_json::from_slice::<WorkflowTaskCommand>(&message.payload) {
+                Ok(cmd) => {
+                    if let Err(e) = crate::generic_worker::handle_workflow_command(state, client, cmd).await {
+                        error!("Generic worker handler failed: {}", e);
+                    }
+                },
+                Err(e) => error!("Failed to parse WorkflowTaskCommand: {}", e),
+            }
+        });
+    }
+    Ok(())
+}
+
+async fn subscribe_legacy(
     state: AppState,
     client: async_nats::Client,
 ) -> Result<()> {
     let subject = NatsSubject::DataFetchCommands.to_string();
     let mut subscriber = client.subscribe(subject.clone()).await?;
     info!(
-        "Consumer started, waiting for messages on subject '{}'",
+        "Legacy Consumer started, waiting for messages on subject '{}'",
         subject
     );
 

services/workflow-orchestrator-service/Cargo.toml

@@ -20,4 +20,8 @@ dashmap = "6.1.0"
 axum = "0.8.7"
 
 # Internal dependencies
-common-contracts = { path = "../common-contracts" }
+common-contracts = { path = "../common-contracts", default-features = false }
+workflow-context = { path = "../../crates/workflow-context" }
+
+[dev-dependencies]
+tempfile = "3"

services/workflow-orchestrator-service/src/config.rs

@@ -7,6 +7,7 @@ pub struct AppConfig {
     pub nats_addr: String,
     pub server_port: u16,
     pub data_persistence_service_url: String,
+    pub workflow_data_path: String,
 }
 
 impl AppConfig {
@@ -18,12 +19,14 @@ impl AppConfig {
             .context("SERVER_PORT must be a number")?;
         let data_persistence_service_url = env::var("DATA_PERSISTENCE_SERVICE_URL")
             .unwrap_or_else(|_| "http://data-persistence-service:3000/api/v1".to_string());
+        let workflow_data_path = env::var("WORKFLOW_DATA_PATH")
+            .unwrap_or_else(|_| "/mnt/workflow_data".to_string());
 
         Ok(Self {
             nats_addr,
             server_port,
             data_persistence_service_url,
+            workflow_data_path,
         })
     }
 }
-

services/workflow-orchestrator-service/src/dag_scheduler.rs

@@ -0,0 +1,264 @@
+use std::collections::HashMap;
+use uuid::Uuid;
+use common_contracts::workflow_types::{TaskStatus, TaskContext};
+use common_contracts::messages::TaskType;
+use workflow_context::{Vgcs, ContextStore};
+use anyhow::Result;
+use tracing::info;
+use serde::{Serialize, Deserialize};
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct CommitTracker {
+    /// Maps task_id to the commit hash it produced.
+    pub task_commits: HashMap<String, String>,
+    /// The latest merged commit for the whole workflow (if linear).
+    /// Or just a reference to the "main" branch tip.
+    pub head_commit: String,
+}
+
+impl CommitTracker {
+    pub fn new(initial_commit: String) -> Self {
+        Self {
+            task_commits: HashMap::new(),
+            head_commit: initial_commit,
+        }
+    }
+
+    pub fn record_commit(&mut self, task_id: &str, commit: String) {
+        self.task_commits.insert(task_id.to_string(), commit.clone());
+        // Note: head_commit update strategy depends on whether we want to track 
+        // a single "main" branch or just use task_commits for DAG resolution.
+        // For now, we don't eagerly update head_commit unless it's a final task.
+    }
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct DagScheduler {
+    pub request_id: Uuid,
+    pub nodes: HashMap<String, DagNode>,
+    /// TaskID -> List of downstream TaskIDs
+    pub forward_deps: HashMap<String, Vec<String>>,
+    /// TaskID -> List of upstream TaskIDs
+    pub reverse_deps: HashMap<String, Vec<String>>,
+    
+    pub commit_tracker: CommitTracker,
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct DagNode {
+    pub id: String,
+    pub task_type: TaskType, // Kept for UI/Observability, not for logic
+    pub status: TaskStatus,
+    pub config: serde_json::Value,
+    pub routing_key: String,
+}
+
+impl DagScheduler {
+    pub fn new(request_id: Uuid, initial_commit: String) -> Self {
+        Self {
+            request_id,
+            nodes: HashMap::new(),
+            forward_deps: HashMap::new(),
+            reverse_deps: HashMap::new(),
+            commit_tracker: CommitTracker::new(initial_commit),
+        }
+    }
+
+    pub fn add_node(&mut self, id: String, task_type: TaskType, routing_key: String, config: serde_json::Value) {
+        self.nodes.insert(id.clone(), DagNode {
+            id,
+            task_type,
+            status: TaskStatus::Pending,
+            config,
+            routing_key,
+        });
+    }
+
+    pub fn add_dependency(&mut self, from: &str, to: &str) {
+        self.forward_deps.entry(from.to_string()).or_default().push(to.to_string());
+        self.reverse_deps.entry(to.to_string()).or_default().push(from.to_string());
+    }
+
+    /// Get all tasks that have no dependencies (roots)
+    pub fn get_initial_tasks(&self) -> Vec<String> {
+        self.nodes.values()
+            .filter(|n| n.status == TaskStatus::Pending && self.reverse_deps.get(&n.id).map_or(true, |deps| deps.is_empty()))
+            .map(|n| n.id.clone())
+            .collect()
+    }
+
+    pub fn update_status(&mut self, task_id: &str, status: TaskStatus) {
+        if let Some(node) = self.nodes.get_mut(task_id) {
+            node.status = status;
+        }
+    }
+    
+    pub fn record_result(&mut self, task_id: &str, new_commit: Option<String>) {
+        if let Some(c) = new_commit {
+            self.commit_tracker.record_commit(task_id, c);
+        }
+    }
+
+    /// Determine which tasks are ready to run given that `completed_task_id` just finished.
+    pub fn get_ready_downstream_tasks(&self, completed_task_id: &str) -> Vec<String> {
+        let mut ready = Vec::new();
+        if let Some(downstream) = self.forward_deps.get(completed_task_id) {
+            for next_id in downstream {
+                if self.is_ready(next_id) {
+                    ready.push(next_id.clone());
+                }
+            }
+        }
+        ready
+    }
+
+    fn is_ready(&self, task_id: &str) -> bool {
+        let node = match self.nodes.get(task_id) {
+            Some(n) => n,
+            None => return false,
+        };
+
+        if node.status != TaskStatus::Pending {
+            return false;
+        }
+
+        if let Some(deps) = self.reverse_deps.get(task_id) {
+            for dep_id in deps {
+                match self.nodes.get(dep_id).map(|n| n.status) {
+                    Some(TaskStatus::Completed) => continue,
+                    _ => return false, // Dependency not completed
+                }
+            }
+        }
+        true
+    }
+
+    /// Resolve the context (Base Commit) for a task.
+    /// If multiple dependencies, perform Merge or Fast-Forward.
+    pub fn resolve_context(&self, task_id: &str, vgcs: &Vgcs) -> Result<TaskContext> {
+        let deps = self.reverse_deps.get(task_id).cloned().unwrap_or_default();
+        
+        if deps.is_empty() {
+            // Root task: Use initial commit (usually empty string or base snapshot)
+            return Ok(TaskContext {
+                base_commit: Some(self.commit_tracker.head_commit.clone()),
+                mount_path: None,
+            });
+        }
+
+        // Collect parent commits
+        let mut parent_commits = Vec::new();
+        for dep_id in &deps {
+            if let Some(c) = self.commit_tracker.task_commits.get(dep_id) {
+                if !c.is_empty() {
+                    parent_commits.push(c.clone());
+                }
+            }
+        }
+
+        if parent_commits.is_empty() {
+            // All parents produced no commit? Fallback to head or empty.
+            return Ok(TaskContext {
+                base_commit: Some(self.commit_tracker.head_commit.clone()),
+                mount_path: None,
+            });
+        }
+
+        // Merge Strategy
+        let final_commit = self.merge_commits(vgcs, parent_commits)?;
+
+        Ok(TaskContext {
+            base_commit: Some(final_commit),
+            mount_path: None, // Or determine based on config
+        })
+    }
+
+    /// Merge logic:
+    /// 1 parent -> Return it.
+    /// 2+ parents -> Iteratively merge using smart merge_commits
+    fn merge_commits(&self, vgcs: &Vgcs, commits: Vec<String>) -> Result<String> {
+        if commits.is_empty() {
+            return Ok(String::new());
+        }
+        if commits.len() == 1 {
+            return Ok(commits[0].clone());
+        }
+        
+        let mut current_head = commits[0].clone();
+        
+        for i in 1..commits.len() {
+            let next_commit = &commits[i];
+            if current_head == *next_commit {
+                continue;
+            }
+            
+            // Use the smart merge_commits which finds the common ancestor automatically
+            // Note: This handles Fast-Forward implicitly (merge_commits checks for ancestry)
+            info!("Merging commits: Ours={}, Theirs={}", current_head, next_commit);
+            current_head = vgcs.merge_commits(&self.request_id.to_string(), &current_head, next_commit)?;
+        }
+
+        Ok(current_head)
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use tempfile::TempDir;
+    use workflow_context::{Vgcs, ContextStore, Transaction};
+    use common_contracts::messages::TaskType;
+    use serde_json::json;
+
+    #[test]
+    fn test_dag_merge_strategy() -> Result<()> {
+        let temp_dir = TempDir::new()?;
+        let vgcs = Vgcs::new(temp_dir.path());
+        let req_id = Uuid::new_v4();
+        let req_id_str = req_id.to_string();
+
+        vgcs.init_repo(&req_id_str)?;
+
+        // 0. Create Initial Commit (Common Ancestor)
+        let mut tx = vgcs.begin_transaction(&req_id_str, "")?;
+        let init_commit = Box::new(tx).commit("Initial Commit", "system")?;
+
+        // 1. Setup DAG
+        let mut dag = DagScheduler::new(req_id, init_commit.clone());
+        dag.add_node("A".to_string(), TaskType::DataFetch, "key.a".into(), json!({}));
+        dag.add_node("B".to_string(), TaskType::DataFetch, "key.b".into(), json!({}));
+        dag.add_node("C".to_string(), TaskType::Analysis, "key.c".into(), json!({}));
+
+        // C depends on A and B
+        dag.add_dependency("A", "C");
+        dag.add_dependency("B", "C");
+
+        // 2. Simulate Task A Execution -> Commit A (Based on Init)
+        let mut tx = vgcs.begin_transaction(&req_id_str, &init_commit)?;
+        tx.write("file_a.txt", b"Content A")?;
+        let commit_a = Box::new(tx).commit("Task A", "worker")?;
+        dag.record_result("A", Some(commit_a.clone()));
+        dag.update_status("A", TaskStatus::Completed);
+
+        // 3. Simulate Task B Execution -> Commit B (Based on Init)
+        let mut tx = vgcs.begin_transaction(&req_id_str, &init_commit)?;
+        tx.write("file_b.txt", b"Content B")?;
+        let commit_b = Box::new(tx).commit("Task B", "worker")?;
+        dag.record_result("B", Some(commit_b.clone()));
+        dag.update_status("B", TaskStatus::Completed);
+
+        // 4. Resolve Context for C
+        // Should merge A and B
+        let ctx = dag.resolve_context("C", &vgcs)?;
+        let merged_commit = ctx.base_commit.expect("Should have base commit");
+
+        // Verify merged content
+        let files = vgcs.list_dir(&req_id_str, &merged_commit, "")?;
+        let file_names: Vec<String> = files.iter().map(|f| f.name.clone()).collect();
+        
+        assert!(file_names.contains(&"file_a.txt".to_string()));
+        assert!(file_names.contains(&"file_b.txt".to_string()));
+        
+        Ok(())
+    }
+}

services/workflow-orchestrator-service/src/lib.rs

@@ -0,0 +1,7 @@
+pub mod api;
+pub mod config;
+pub mod message_consumer;
+pub mod persistence;
+pub mod state;
+pub mod workflow;
+pub mod dag_scheduler;

services/workflow-orchestrator-service/src/main.rs

@@ -2,20 +2,15 @@ use anyhow::Result;
 use tracing::info;
 use tracing_subscriber::EnvFilter;
 use std::sync::Arc;
-
-mod config;
-mod state;
-mod message_consumer;
-mod workflow;
-mod persistence;
-mod api;
+use workflow_orchestrator_service::{config, state, message_consumer, api};
 
 #[tokio::main]
 async fn main() -> Result<()> {
     // Initialize tracing
     tracing_subscriber::fmt()
-        .with_env_filter(EnvFilter::from_default_env())
-        .init();
+    // .with_env_filter(EnvFilter::from_default_env())
+    .with_env_filter("info")
+    .init();
 
     info!("Starting workflow-orchestrator-service...");
 
@@ -38,8 +33,7 @@ async fn main() -> Result<()> {
         }
     });
 
-    // Start HTTP Server (for health check & potential admin/debug)
-    // Note: The main trigger API is now NATS commands, but we might expose some debug endpoints
+    // Start HTTP Server
     let app = api::create_router(state.clone());
     let addr = format!("0.0.0.0:{}", config.server_port);
     let listener = tokio::net::TcpListener::bind(&addr).await?;

services/workflow-orchestrator-service/src/message_consumer.rs

@@ -4,114 +4,50 @@ use anyhow::Result;
 use tracing::{info, error};
 use futures::StreamExt;
 use crate::state::AppState;
-use common_contracts::messages::{
-    StartWorkflowCommand, SyncStateCommand, 
-    FinancialsPersistedEvent, DataFetchFailedEvent,
-};
+use common_contracts::messages::StartWorkflowCommand;
+use common_contracts::workflow_types::WorkflowTaskEvent;
 use common_contracts::subjects::NatsSubject;
-use crate::persistence::{DashMapWorkflowRepository, NatsMessageBroker};
 use crate::workflow::WorkflowEngine;
 
 pub async fn run(state: Arc<AppState>, nats: Client) -> Result<()> {
     info!("Message Consumer started. Subscribing to topics...");
 
-    let mut workflow_sub = nats.subscribe(NatsSubject::WorkflowCommandsWildcard.to_string()).await?;
-    let mut data_sub = nats.subscribe(NatsSubject::DataEventsWildcard.to_string()).await?;
-    let mut analysis_sub = nats.subscribe(NatsSubject::AnalysisEventsWildcard.to_string()).await?;
+    // Topic 1: Workflow Commands (Start)
+    // Note: NatsSubject::WorkflowCommandStart string representation is "workflow.commands.start"
+    let mut start_sub = nats.subscribe(NatsSubject::WorkflowCommandStart.to_string()).await?;
 
-    // Prepare components
-    let repo = DashMapWorkflowRepository::new(state.workflows.clone());
-    let broker = NatsMessageBroker::new(nats.clone());
+    // Topic 2: Workflow Task Events (Generic)
+    // Note: NatsSubject::WorkflowEventTaskCompleted string representation is "workflow.evt.task_completed"
+    let mut task_sub = nats.subscribe(NatsSubject::WorkflowEventTaskCompleted.to_string()).await?;
 
-    // ... (Task 1 & 2 remain same) ...
+    let engine = Arc::new(WorkflowEngine::new(state.clone(), nats.clone()));
 
-    // --- Task 3: Analysis Events (Report Gen updates) ---
-    let repo3 = repo.clone();
-    let broker3 = broker.clone();
+    // --- Task 1: Start Workflow ---
+    let engine1 = engine.clone();
     tokio::spawn(async move {
-        let engine = WorkflowEngine::new(Box::new(repo3), Box::new(broker3));
-        while let Some(msg) = analysis_sub.next().await {
-            match NatsSubject::try_from(msg.subject.as_str()) {
-                Ok(NatsSubject::AnalysisReportGenerated) => {
-                    if let Ok(evt) = serde_json::from_slice::<common_contracts::messages::ReportGeneratedEvent>(&msg.payload) {
-                        info!("Received ReportGenerated: {:?}", evt);
-                        if let Err(e) = engine.handle_report_generated(evt).await {
-                            error!("Failed to handle ReportGenerated: {}", e);
-                        }
-                    }
-                },
-                Ok(NatsSubject::AnalysisReportFailed) => {
-                    if let Ok(evt) = serde_json::from_slice::<common_contracts::messages::ReportFailedEvent>(&msg.payload) {
-                        info!("Received ReportFailed: {:?}", evt);
-                        if let Err(e) = engine.handle_report_failed(evt).await {
-                            error!("Failed to handle ReportFailed: {}", e);
-                        }
-                    }
-                },
-                _ => {
-                    // Ignore other subjects or log warning
+        while let Some(msg) = start_sub.next().await {
+            if let Ok(cmd) = serde_json::from_slice::<StartWorkflowCommand>(&msg.payload) {
+                info!("Received StartWorkflow: {:?}", cmd);
+                if let Err(e) = engine1.handle_start_workflow(cmd).await {
+                    error!("Failed to handle StartWorkflow: {}", e);
                 }
+            } else {
+                error!("Failed to parse StartWorkflowCommand");
             }
         }
     });
 
-    // --- Task 1: Workflow Commands (Start, Sync) ---
-    let repo1 = repo.clone();
-    let broker1 = broker.clone();
+    // --- Task 2: Task Completed Events ---
+    let engine2 = engine.clone();
     tokio::spawn(async move {
-        let engine = WorkflowEngine::new(Box::new(repo1), Box::new(broker1));
-        while let Some(msg) = workflow_sub.next().await {
-            match NatsSubject::try_from(msg.subject.as_str()) {
-                Ok(NatsSubject::WorkflowCommandStart) => {
-                    if let Ok(cmd) = serde_json::from_slice::<StartWorkflowCommand>(&msg.payload) {
-                        info!("Received StartWorkflow: {:?}", cmd);
-                        if let Err(e) = engine.handle_start_workflow(cmd).await {
-                            error!("Failed to handle StartWorkflow: {}", e);
-                        }
-                    } else {
-                        error!("Failed to parse StartWorkflowCommand");
-                    }
-                },
-                Ok(NatsSubject::WorkflowCommandSyncState) => {
-                    if let Ok(cmd) = serde_json::from_slice::<SyncStateCommand>(&msg.payload) {
-                        info!("Received SyncState: {:?}", cmd);
-                        // 关键修复：这里直接调用处理逻辑，如果还不行，我们需要检查 state 是否真的加载到了
-                        // 或者在这里加更多的 log
-                        match engine.handle_sync_state(cmd).await {
-                            Ok(_) => info!("Successfully processed SyncState"),
-                            Err(e) => error!("Failed to handle SyncState: {}", e),
-                        }
-                    }
-                },
-                _ => {}
-            }
-        }
-    });
-
-    // --- Task 2: Data Events (Progress updates) ---
-    let repo2 = repo.clone();
-    let broker2 = broker.clone();
-    tokio::spawn(async move {
-        let engine = WorkflowEngine::new(Box::new(repo2), Box::new(broker2));
-        while let Some(msg) = data_sub.next().await {
-            match NatsSubject::try_from(msg.subject.as_str()) {
-                Ok(NatsSubject::DataFinancialsPersisted) => {
-                    if let Ok(evt) = serde_json::from_slice::<FinancialsPersistedEvent>(&msg.payload) {
-                        info!("Received DataFetched: {:?}", evt);
-                        if let Err(e) = engine.handle_data_fetched(evt).await {
-                            error!("Failed to handle DataFetched: {}", e);
-                        }
-                    }
-                },
-                Ok(NatsSubject::DataFetchFailed) => {
-                    if let Ok(evt) = serde_json::from_slice::<DataFetchFailedEvent>(&msg.payload) {
-                        info!("Received DataFailed: {:?}", evt);
-                        if let Err(e) = engine.handle_data_failed(evt).await {
-                            error!("Failed to handle DataFailed: {}", e);
-                        }
-                    }
-                },
-                _ => {}
+        while let Some(msg) = task_sub.next().await {
+            if let Ok(evt) = serde_json::from_slice::<WorkflowTaskEvent>(&msg.payload) {
+                info!("Received TaskCompleted: task_id={}", evt.task_id);
+                if let Err(e) = engine2.handle_task_completed(evt).await {
+                    error!("Failed to handle TaskCompleted: {}", e);
+                }
+            } else {
+                error!("Failed to parse WorkflowTaskEvent");
             }
         }
     });

services/workflow-orchestrator-service/src/persistence.rs

@@ -1,72 +1,2 @@
-use async_trait::async_trait;
-use crate::workflow::{WorkflowRepository, WorkflowStateMachine, MessageBroker};
-use std::sync::Arc;
-use tokio::sync::Mutex;
-use dashmap::DashMap;
-use uuid::Uuid;
-use anyhow::Result;
-use async_nats::Client;
-use common_contracts::messages::WorkflowEvent;
-
-// --- Real Implementation of Repository (In-Memory for MVP) ---
-#[derive(Clone)]
-pub struct DashMapWorkflowRepository {
-    store: Arc<DashMap<Uuid, Arc<Mutex<WorkflowStateMachine>>>>
-}
-
-impl DashMapWorkflowRepository {
-    pub fn new(store: Arc<DashMap<Uuid, Arc<Mutex<WorkflowStateMachine>>>>) -> Self {
-        Self { store }
-    }
-}
-
-#[async_trait]
-impl WorkflowRepository for DashMapWorkflowRepository {
-    async fn save(&self, workflow: &WorkflowStateMachine) -> Result<()> {
-        if let Some(entry) = self.store.get(&workflow.request_id) {
-            let mut guard = entry.lock().await;
-            *guard = workflow.clone();
-        } else {
-            self.store.insert(workflow.request_id, Arc::new(Mutex::new(workflow.clone())));
-        }
-        Ok(())
-    }
-
-    async fn load(&self, id: Uuid) -> Result<Option<WorkflowStateMachine>> {
-        if let Some(entry) = self.store.get(&id) {
-            let guard = entry.lock().await;
-            return Ok(Some(guard.clone()));
-        }
-        // Add debug log to see what keys are actually in the store
-        // tracing::debug!("Keys in store: {:?}", self.store.iter().map(|r| *r.key()).collect::<Vec<_>>());
-        Ok(None)
-    }
-}
-
-// --- Real Implementation of Broker ---
-
-#[derive(Clone)]
-pub struct NatsMessageBroker {
-    client: Client
-}
-
-impl NatsMessageBroker {
-    pub fn new(client: Client) -> Self {
-        Self { client }
-    }
-}
-
-use common_contracts::subjects::NatsSubject;
-
-#[async_trait]
-impl MessageBroker for NatsMessageBroker {
-    async fn publish_event(&self, request_id: Uuid, event: WorkflowEvent) -> Result<()> {
-        let topic = NatsSubject::WorkflowProgress(request_id).to_string();
-        let payload = serde_json::to_vec(&event)?;
-        self.client.publish(topic, payload.into()).await.map_err(|e| anyhow::anyhow!(e))
-    }
-
-    async fn publish_command(&self, topic: &str, payload: Vec<u8>) -> Result<()> {
-        self.client.publish(topic.to_string(), payload.into()).await.map_err(|e| anyhow::anyhow!(e))
-    }
-}
+// Persistence module currently unused after refactor to DagScheduler.
+// Will be reintroduced when we implement DB persistence for the DAG.

services/workflow-orchestrator-service/src/state.rs

@@ -5,25 +5,33 @@ use common_contracts::persistence_client::PersistenceClient;
 use dashmap::DashMap;
 use uuid::Uuid;
 use tokio::sync::Mutex;
-use crate::workflow::WorkflowStateMachine;
+// use crate::workflow::WorkflowStateMachine; // Deprecated
+use crate::dag_scheduler::DagScheduler;
+use workflow_context::Vgcs;
 
 pub struct AppState {
+    #[allow(dead_code)]
     pub config: AppConfig,
+    #[allow(dead_code)]
     pub persistence_client: PersistenceClient,
+    
     // Key: request_id
     // We use Mutex here because state transitions need to be atomic per workflow
-    pub workflows: Arc<DashMap<Uuid, Arc<Mutex<WorkflowStateMachine>>>>,
+    pub workflows: Arc<DashMap<Uuid, Arc<Mutex<DagScheduler>>>>,
+    
+    pub vgcs: Arc<Vgcs>,
 }
 
 impl AppState {
     pub async fn new(config: AppConfig) -> Result<Self> {
         let persistence_client = PersistenceClient::new(config.data_persistence_service_url.clone());
+        let vgcs = Arc::new(Vgcs::new(&config.workflow_data_path));
         
         Ok(Self {
             config,
             persistence_client,
             workflows: Arc::new(DashMap::new()),
+            vgcs,
         })
     }
 }
-

services/workflow-orchestrator-service/src/workflow.rs

@@ -1,680 +1,181 @@
-use std::collections::HashMap;
-use uuid::Uuid;
-use common_contracts::messages::{
-    WorkflowEvent, WorkflowDag, TaskStatus, TaskType, TaskNode, TaskDependency,
-    StartWorkflowCommand, FinancialsPersistedEvent, DataFetchFailedEvent, 
-    SyncStateCommand, FetchCompanyDataCommand, GenerateReportCommand
+use std::sync::Arc;
+use common_contracts::workflow_types::{
+    WorkflowTaskCommand, WorkflowTaskEvent, TaskStatus, StorageConfig
+};
+use common_contracts::messages::{
+    StartWorkflowCommand, TaskType
 };
-use common_contracts::symbol_utils::CanonicalSymbol;
-use tracing::{info, warn};
-use async_trait::async_trait;
-use anyhow::Result;
-
-#[derive(Debug, Clone)]
-pub struct WorkflowStateMachine {
-    pub request_id: Uuid,
-    pub symbol: CanonicalSymbol,
-    pub market: String,
-    pub template_id: String,
-    pub dag: WorkflowDag,
-    
-    // Runtime state: TaskID -> Status
-    pub task_status: HashMap<String, TaskStatus>,
-    // Fast lookup for dependencies: TaskID -> List of dependencies (upstream)
-    pub reverse_deps: HashMap<String, Vec<String>>,
-    // Fast lookup for dependents: TaskID -> List of tasks that depend on it (downstream)
-    pub forward_deps: HashMap<String, Vec<String>>,
-}
-
-impl WorkflowStateMachine {
-    pub fn new(request_id: Uuid, symbol: CanonicalSymbol, market: String, template_id: String) -> Self {
-        let (dag, reverse_deps, forward_deps) = Self::build_dag(&template_id);
-        
-        let mut task_status = HashMap::new();
-        for node in &dag.nodes {
-            task_status.insert(node.id.clone(), node.initial_status);
-        }
-
-        Self {
-            request_id,
-            symbol,
-            market,
-            template_id,
-            dag,
-            task_status,
-            reverse_deps,
-            forward_deps,
-        }
-    }
-
-    // Determines which tasks are ready to run immediately (no dependencies)
-    pub fn get_initial_tasks(&self) -> Vec<String> {
-        self.dag.nodes.iter()
-            .filter(|n| n.initial_status == TaskStatus::Pending && self.reverse_deps.get(&n.id).map_or(true, |deps| deps.is_empty()))
-            .map(|n| n.id.clone())
-            .collect()
-    }
-
-    // Core logic for state transition
-    // Returns a list of tasks that should be TRIGGERED (transitioned from Pending -> Scheduled)
-    pub fn update_task_status(&mut self, task_id: &str, new_status: TaskStatus) -> Vec<String> {
-        if let Some(current) = self.task_status.get_mut(task_id) {
-            if *current == new_status {
-                return vec![]; // No change
-            }
-            
-            info!("Workflow {}: Task {} transition {:?} -> {:?}", self.request_id, task_id, current, new_status);
-            *current = new_status;
-        } else {
-            warn!("Workflow {}: Unknown task id {}", self.request_id, task_id);
-            return vec![];
-        }
-
-        // If task completed, check downstream dependents
-        if new_status == TaskStatus::Completed {
-            return self.check_dependents(task_id);
-        } 
-        // If task failed, we might need to skip downstream tasks (Policy decision)
-        else if new_status == TaskStatus::Failed {
-            // For now, simplistic policy: propagate failure as Skipped
-            self.propagate_skip(task_id);
-        }
-
-        vec![]
-    }
-
-    fn check_dependents(&self, completed_task_id: &str) -> Vec<String> {
-        let mut ready_tasks = Vec::new();
-
-        if let Some(downstream) = self.forward_deps.get(completed_task_id) {
-            for dependent_id in downstream {
-                if self.is_task_ready(dependent_id) {
-                    ready_tasks.push(dependent_id.clone());
-                }
-            }
-        }
-        
-        ready_tasks
-    }
-
-    fn is_task_ready(&self, task_id: &str) -> bool {
-        if let Some(status) = self.task_status.get(task_id) {
-            if *status != TaskStatus::Pending {
-                return false; 
-            }
-        } else {
-            return false;
-        }
-
-        if let Some(deps) = self.reverse_deps.get(task_id) {
-            for dep_id in deps {
-                match self.task_status.get(dep_id) {
-                    Some(TaskStatus::Completed) => continue, // Good
-                    _ => return false, // Any other status means not ready
-                }
-            }
-        }
-        
-        true
-    }
-
-    fn propagate_skip(&mut self, failed_task_id: &str) {
-         if let Some(downstream) = self.forward_deps.get(failed_task_id) {
-            for dependent_id in downstream {
-                if let Some(status) = self.task_status.get_mut(dependent_id) {
-                    if *status == TaskStatus::Pending {
-                        *status = TaskStatus::Skipped;
-                        // For MVP we do 1 level. Real impl needs full graph traversal.
-                    }
-                }
-            }
-        }
-    }
-
-    fn build_dag(template_id: &str) -> (WorkflowDag, HashMap<String, Vec<String>>, HashMap<String, Vec<String>>) {
-        let mut nodes = vec![];
-        let mut edges = vec![];
-        
-        // Simplistic Hack for E2E Testing Phase 4
-        if template_id == "simple_test_analysis" {
-            // Simple DAG: Fetch(yfinance) -> Analysis
-            nodes.push(TaskNode {
-                id: "fetch:yfinance".to_string(),
-                name: "Fetch yfinance".to_string(),
-                r#type: TaskType::DataFetch,
-                initial_status: TaskStatus::Pending,
-            });
-             nodes.push(TaskNode {
-                id: "analysis:report".to_string(),
-                name: "Generate Report".to_string(),
-                r#type: TaskType::Analysis,
-                initial_status: TaskStatus::Pending,
-            });
-             edges.push(TaskDependency { from: "fetch:yfinance".to_string(), to: "analysis:report".to_string() });
-
-        } else {
-            // Default Hardcoded DAG
-            // 1. Data Fetch Nodes (Roots)
-            let providers = vec!["alphavantage", "tushare", "finnhub", "yfinance"];
-            for p in &providers {
-                nodes.push(TaskNode {
-                    id: format!("fetch:{}", p),
-                    name: format!("Fetch {}", p),
-                    r#type: TaskType::DataFetch,
-                    initial_status: TaskStatus::Pending,
-                });
-            }
-
-            // 2. Analysis Node
-            nodes.push(TaskNode {
-                id: "analysis:report".to_string(),
-                name: "Generate Report".to_string(),
-                r#type: TaskType::Analysis,
-                initial_status: TaskStatus::Pending,
-            });
-
-            // Edges: All fetchers -> Analysis
-            for p in &providers {
-                edges.push(TaskDependency { from: format!("fetch:{}", p), to: "analysis:report".to_string() });
-            }
-        }
-
-        let mut reverse_deps: HashMap<String, Vec<String>> = HashMap::new();
-        let mut forward_deps: HashMap<String, Vec<String>> = HashMap::new();
-
-        for edge in &edges {
-            reverse_deps.entry(edge.to.clone()).or_default().push(edge.from.clone());
-            forward_deps.entry(edge.from.clone()).or_default().push(edge.to.clone());
-        }
-
-        (WorkflowDag { nodes, edges }, reverse_deps, forward_deps)
-    }
-    
-    pub fn get_snapshot_event(&self) -> WorkflowEvent {
-        WorkflowEvent::WorkflowStateSnapshot {
-            timestamp: chrono::Utc::now().timestamp_millis(),
-            task_graph: self.dag.clone(),
-            tasks_status: self.task_status.clone(),
-            tasks_output: HashMap::new(),
-        }
-    }
-}
-
-// --- Traits & Engine ---
-
 use common_contracts::subjects::SubjectMessage;
+use common_contracts::symbol_utils::CanonicalSymbol;
+use tracing::{info, warn, error};
+use anyhow::Result;
+use serde_json::json;
+use tokio::sync::Mutex;
 
-#[async_trait]
-pub trait WorkflowRepository: Send + Sync {
-    async fn save(&self, workflow: &WorkflowStateMachine) -> Result<()>;
-    async fn load(&self, id: Uuid) -> Result<Option<WorkflowStateMachine>>;
-}
-
-#[async_trait]
-pub trait MessageBroker: Send + Sync {
-    async fn publish_event(&self, request_id: Uuid, event: WorkflowEvent) -> Result<()>;
-    async fn publish_command(&self, topic: &str, payload: Vec<u8>) -> Result<()>;
-}
+use crate::dag_scheduler::DagScheduler;
+use crate::state::AppState;
+use workflow_context::{Vgcs, ContextStore}; // Added ContextStore
 
 pub struct WorkflowEngine {
-    repo: Box<dyn WorkflowRepository>,
-    broker: Box<dyn MessageBroker>,
+    state: Arc<AppState>,
+    nats: async_nats::Client,
 }
 
 impl WorkflowEngine {
-    pub fn new(repo: Box<dyn WorkflowRepository>, broker: Box<dyn MessageBroker>) -> Self {
-        Self { repo, broker }
+    pub fn new(state: Arc<AppState>, nats: async_nats::Client) -> Self {
+        Self { state, nats }
     }
 
     pub async fn handle_start_workflow(&self, cmd: StartWorkflowCommand) -> Result<()> {
-        let mut machine = WorkflowStateMachine::new(
-            cmd.request_id, 
-            cmd.symbol.clone(), 
-            cmd.market.clone(),
-            cmd.template_id.clone()
+        let req_id = cmd.request_id;
+        info!("Starting workflow {}", req_id);
+
+        // 1. Init VGCS Repo
+        self.state.vgcs.init_repo(&req_id.to_string())?;
+
+        // 2. Create Scheduler
+        // Initial commit is empty for a fresh workflow
+        let mut dag = DagScheduler::new(req_id, String::new());
+
+        // 3. Build DAG (Simplified Hardcoded logic for now, matching old build_dag)
+        // In a real scenario, we fetch Template from DB/Service using cmd.template_id
+        self.build_dag(&mut dag, &cmd.template_id, &cmd.market, &cmd.symbol);
+
+        // 4. Save State
+        self.state.workflows.insert(req_id, Arc::new(Mutex::new(dag.clone())));
+
+        // 5. Trigger Initial Tasks
+        let initial_tasks = dag.get_initial_tasks();
+        
+        // Lock the DAG for updates
+        let dag_arc = self.state.workflows.get(&req_id).unwrap().clone();
+        let mut dag_guard = dag_arc.lock().await;
+
+        for task_id in initial_tasks {
+            self.dispatch_task(&mut dag_guard, &task_id, &self.state.vgcs).await?;
+        }
+
+        Ok(())
+    }
+
+    pub async fn handle_task_completed(&self, evt: WorkflowTaskEvent) -> Result<()> {
+        let req_id = evt.request_id;
+        
+        let dag_arc = match self.state.workflows.get(&req_id) {
+            Some(d) => d.clone(),
+            None => {
+                warn!("Received event for unknown workflow {}", req_id);
+                return Ok(());
+            }
+        };
+
+        let mut dag = dag_arc.lock().await;
+
+        // 1. Update Status & Record Commit
+        dag.update_status(&evt.task_id, evt.status);
+        
+        if let Some(result) = evt.result {
+            if let Some(commit) = result.new_commit {
+                info!("Task {} produced commit {}", evt.task_id, commit);
+                dag.record_result(&evt.task_id, Some(commit));
+            }
+            if let Some(err) = result.error {
+                warn!("Task {} failed with error: {}", evt.task_id, err);
+            }
+        }
+
+        // 2. Check for downstream tasks
+        if evt.status == TaskStatus::Completed {
+            let ready_tasks = dag.get_ready_downstream_tasks(&evt.task_id);
+            for task_id in ready_tasks {
+                if let Err(e) = self.dispatch_task(&mut dag, &task_id, &self.state.vgcs).await {
+                    error!("Failed to dispatch task {}: {}", task_id, e);
+                }
+            }
+        } else if evt.status == TaskStatus::Failed {
+            // Handle failure propagation (skip downstream?)
+            // For now, just log.
+            error!("Task {} failed. Workflow might stall.", evt.task_id);
+        }
+
+        // 3. Check Workflow Completion (TODO)
+
+        Ok(())
+    }
+
+    async fn dispatch_task(&self, dag: &mut DagScheduler, task_id: &str, vgcs: &Vgcs) -> Result<()> {
+        // 1. Resolve Context (Merge if needed)
+        let context = dag.resolve_context(task_id, vgcs)?;
+        
+        // 2. Update Status
+        dag.update_status(task_id, TaskStatus::Scheduled);
+
+        // 3. Construct Command
+        let node = dag.nodes.get(task_id).ok_or_else(|| anyhow::anyhow!("Node not found"))?;
+        
+        let cmd = WorkflowTaskCommand {
+            request_id: dag.request_id,
+            task_id: task_id.to_string(),
+            routing_key: node.routing_key.clone(),
+            config: node.config.clone(),
+            context,
+            storage: StorageConfig {
+                root_path: self.state.config.workflow_data_path.clone(),
+            },
+        };
+
+        // 4. Publish
+        let subject = cmd.subject().to_string(); // This uses the routing_key
+        let payload = serde_json::to_vec(&cmd)?;
+        
+        info!("Dispatching task {} to subject {}", task_id, subject);
+        self.nats.publish(subject, payload.into()).await?;
+
+        Ok(())
+    }
+
+    // Helper to build DAG (Migrated from old workflow.rs)
+    fn build_dag(&self, dag: &mut DagScheduler, template_id: &str, market: &str, symbol: &CanonicalSymbol) {
+        // Logic copied/adapted from old WorkflowStateMachine::build_dag
+        
+        let mut providers = Vec::new();
+        match market {
+            "CN" => {
+                providers.push("tushare");
+                // providers.push("yfinance"); 
+            },
+            "US" => providers.push("yfinance"),
+            _ => providers.push("yfinance"),
+        }
+
+        // 1. Data Fetch Nodes
+        for p in &providers {
+            let task_id = format!("fetch:{}", p);
+            dag.add_node(
+                task_id.clone(), 
+                TaskType::DataFetch, 
+                format!("provider.{}", p), // routing_key: workflow.cmd.provider.tushare
+                json!({
+                    "symbol": symbol.as_str(), // Simplification
+                    "market": market
+                })
+            );
+        }
+
+        // 2. Analysis Node (Simplified)
+        let report_task_id = "analysis:report";
+        dag.add_node(
+            report_task_id.to_string(),
+            TaskType::Analysis,
+            "analysis.report".to_string(), // routing_key: workflow.cmd.analysis.report
+            json!({
+                "template_id": template_id
+            })
         );
 
-        self.repo.save(&machine).await?;
-
-        // Broadcast Started
-        let start_event = WorkflowEvent::WorkflowStarted {
-            timestamp: chrono::Utc::now().timestamp_millis(),
-            task_graph: machine.dag.clone()
-        };
-        self.broker.publish_event(cmd.request_id, start_event).await?;
-
-        // Initial Tasks
-        let initial_tasks = machine.get_initial_tasks();
-        for task_id in initial_tasks {
-            let _ = machine.update_task_status(&task_id, TaskStatus::Scheduled);
-            
-            self.broadcast_task_state(cmd.request_id, &task_id, TaskType::DataFetch, TaskStatus::Scheduled, None, None).await?;
-            
-            if task_id.starts_with("fetch:") {
-                let fetch_cmd = FetchCompanyDataCommand {
-                    request_id: cmd.request_id,
-                    symbol: cmd.symbol.clone(),
-                    market: cmd.market.clone(),
-                    template_id: Some(cmd.template_id.clone()),
-                };
-                // Extract provider from task_id (e.g., "fetch:yfinance" -> "yfinance")
-                // We only publish if the task_id matches the fetch provider we intend to trigger.
-                // But wait, 'FetchCompanyDataCommand' is generic.
-                // If we send it, all listening providers will pick it up.
-                // We rely on the fact that unconfigured providers (Tushare etc) are "paused" or won't act if no key.
-                // However, if we have "simple_test_analysis" which ONLY has "fetch:yfinance",
-                // we still broadcast "data_fetch_commands".
-                // Tushare service WILL receive it. If it doesn't have a key, it warns and ignores or retries.
-                // Ideally, the command should carry the target provider ID.
-                // But FetchCompanyDataCommand structure is:
-                // pub struct FetchCompanyDataCommand { request_id, symbol, market, template_id }
-                // It lacks provider_id.
-                
-                // For now, we just broadcast. The key is that the DAG only waits for "fetch:yfinance".
-                // So even if Tushare fails or never returns, it doesn't matter because "fetch:tushare" is NOT in the DAG.
-                
-                let payload = serde_json::to_vec(&fetch_cmd)?;
-                self.broker.publish_command(fetch_cmd.subject().to_string().as_str(), payload).await?;
-            }
-        }
-        
-        self.repo.save(&machine).await?;
-        Ok(())
-    }
-
-    pub async fn handle_data_fetched(&self, evt: FinancialsPersistedEvent) -> Result<()> {
-        if let Some(mut machine) = self.repo.load(evt.request_id).await? {
-             let provider_id = evt.provider_id.as_deref().unwrap_or("unknown");
-             let task_id = format!("fetch:{}", provider_id);
-             
-             let next_tasks = machine.update_task_status(&task_id, TaskStatus::Completed);
-             
-             let summary = evt.data_summary.clone().unwrap_or_else(|| "Data fetched".to_string());
-             self.broker.publish_event(evt.request_id, WorkflowEvent::TaskStreamUpdate {
-                 task_id: task_id.clone(),
-                 content_delta: summary,
-                 index: 0
-             }).await?;
-             
-             self.broadcast_task_state(evt.request_id, &task_id, TaskType::DataFetch, TaskStatus::Completed, Some("Data persisted".into()), None).await?;
-             
-             self.trigger_next_tasks(&mut machine, next_tasks).await?;
-             
-             self.repo.save(&machine).await?;
-        }
-        Ok(())
-    }
-
-    pub async fn handle_data_failed(&self, evt: DataFetchFailedEvent) -> Result<()> {
-        if let Some(mut machine) = self.repo.load(evt.request_id).await? {
-            let provider_id = evt.provider_id.as_deref().unwrap_or("unknown");
-            let task_id = format!("fetch:{}", provider_id);
-            
-            let _ = machine.update_task_status(&task_id, TaskStatus::Failed);
-            self.broadcast_task_state(evt.request_id, &task_id, TaskType::DataFetch, TaskStatus::Failed, Some(evt.error), None).await?;
-            self.repo.save(&machine).await?;
-        }
-        Ok(())
-    }
-
-    pub async fn handle_sync_state(&self, cmd: SyncStateCommand) -> Result<()> {
-        info!("Handling SyncState for request_id: {}", cmd.request_id);
-        if let Some(machine) = self.repo.load(cmd.request_id).await? {
-            info!("Found workflow machine for {}, publishing snapshot", cmd.request_id);
-            let snapshot = machine.get_snapshot_event();
-            self.broker.publish_event(cmd.request_id, snapshot).await?;
-        } else {
-            warn!("No workflow found for request_id: {}", cmd.request_id);
-        }
-        Ok(())
-    }
-
-    pub async fn handle_report_generated(&self, evt: common_contracts::messages::ReportGeneratedEvent) -> Result<()> {
-        if let Some(mut machine) = self.repo.load(evt.request_id).await? {
-            let task_id = format!("analysis:{}", evt.module_id);
-            // The module_id from ReportGen might be "report" or "step2_analyze" or whatever the template used.
-            // In "simple_test_analysis", the task ID is "analysis:report" (hardcoded in build_dag)
-            // But the template module key was "step2_analyze".
-            
-            // WAIT: In build_dag:
-            // nodes.push(TaskNode { id: "analysis:report".to_string(), ... });
-            // But in GenerateReportCommand sent to report-gen:
-            // template_id: machine.template_id
-            
-            // Report Generator uses the template modules.
-            // It executes "step2_analyze".
-            // It sends ReportGeneratedEvent with module_id = "step2_analyze".
-            
-            // Orchestrator's DAG has "analysis:report".
-            // MISMATCH!
-            
-            // The hardcoded build_dag uses "analysis:report".
-            // The actual template execution uses "step2_analyze".
-            
-            // This mismatch is why the E2E test fails even if Report Gen succeeds.
-            // Report Gen says: "step2_analyze" done.
-            // Orchestrator waits for: "analysis:report".
-            
-            // FIX:
-            // We need to map the incoming module_id to the task_id in our DAG.
-            // OR update build_dag to use the correct ID.
-            
-            // For simple_test_analysis, we can hack it here or fix build_dag.
-            // Since I cannot change build_dag dynamically easily without loading template,
-            // I will try to handle the mapping here.
-            
-            let target_task_id = if machine.template_id == "simple_test_analysis" && evt.module_id == "step2_analyze" {
-                "analysis:report".to_string()
-            } else {
-                // Default assumption: task_id = "analysis:{module_id}"
-                // But our hardcoded DAG uses "analysis:report".
-                // If module_id is "report", then it matches.
-                format!("analysis:{}", evt.module_id)
-            };
-            
-            // Fallback: if target_task_id not in DAG, try "analysis:report" if it exists and is Running/Scheduled
-            let final_task_id = if machine.task_status.contains_key(&target_task_id) {
-                target_task_id
-            } else if machine.task_status.contains_key("analysis:report") {
-                "analysis:report".to_string()
-            } else {
-                target_task_id
-            };
-
-            let next_tasks = machine.update_task_status(&final_task_id, TaskStatus::Completed);
-            
-            self.broadcast_task_state(evt.request_id, &final_task_id, TaskType::Analysis, TaskStatus::Completed, Some("Report generated".into()), None).await?;
-            
-            // If this was the last node, the workflow is complete?
-            // check_dependents returns next ready tasks. If empty, and no running tasks, we are done.
-            // We should check if workflow is complete.
-            
-            self.check_workflow_completion(&machine).await?;
-            
-            self.repo.save(&machine).await?;
-        }
-        Ok(())
-    }
-
-    pub async fn handle_report_failed(&self, evt: common_contracts::messages::ReportFailedEvent) -> Result<()> {
-        if let Some(mut machine) = self.repo.load(evt.request_id).await? {
-             let task_id = if machine.template_id == "simple_test_analysis" && evt.module_id == "step2_analyze" {
-                "analysis:report".to_string()
-            } else {
-                format!("analysis:{}", evt.module_id)
-            };
-            
-            let final_task_id = if machine.task_status.contains_key(&task_id) {
-                task_id
-            } else {
-                "analysis:report".to_string()
-            };
-
-            let _ = machine.update_task_status(&final_task_id, TaskStatus::Failed);
-            self.broadcast_task_state(evt.request_id, &final_task_id, TaskType::Analysis, TaskStatus::Failed, Some(evt.error.clone()), None).await?;
-            
-            // Propagate failure?
-            self.broker.publish_event(evt.request_id, WorkflowEvent::WorkflowFailed {
-                reason: format!("Analysis task failed: {}", evt.error),
-                is_fatal: true,
-                end_timestamp: chrono::Utc::now().timestamp_millis(),
-            }).await?;
-            
-            self.repo.save(&machine).await?;
-        }
-        Ok(())
-    }
-
-    async fn check_workflow_completion(&self, machine: &WorkflowStateMachine) -> Result<()> {
-        // If all tasks are Completed or Skipped, workflow is done.
-        let all_done = machine.task_status.values().all(|s| *s == TaskStatus::Completed || *s == TaskStatus::Skipped || *s == TaskStatus::Failed);
-        
-        if all_done {
-            self.broker.publish_event(machine.request_id, WorkflowEvent::WorkflowCompleted {
-                result_summary: serde_json::json!({"status": "success"}),
-                end_timestamp: chrono::Utc::now().timestamp_millis(),
-            }).await?;
-        }
-        Ok(())
-    }
-
-    async fn trigger_next_tasks(&self, machine: &mut WorkflowStateMachine, tasks: Vec<String>) -> Result<()> {
-        for task_id in tasks {
-            machine.update_task_status(&task_id, TaskStatus::Scheduled);
-            self.broadcast_task_state(machine.request_id, &task_id, TaskType::Analysis, TaskStatus::Scheduled, None, None).await?;
-
-            if task_id == "analysis:report" {
-                 let cmd = GenerateReportCommand {
-                    request_id: machine.request_id,
-                    symbol: machine.symbol.clone(),
-                    template_id: machine.template_id.clone(),
-                };
-                let payload = serde_json::to_vec(&cmd)?;
-                self.broker.publish_command(cmd.subject().to_string().as_str(), payload).await?;
-            }
-        }
-        Ok(())
-    }
-
-    async fn broadcast_task_state(&self, request_id: Uuid, task_id: &str, ttype: TaskType, status: TaskStatus, msg: Option<String>, progress: Option<u8>) -> Result<()> {
-        let event = WorkflowEvent::TaskStateChanged {
-            task_id: task_id.to_string(),
-            task_type: ttype,
-            status,
-            message: msg,
-            timestamp: chrono::Utc::now().timestamp_millis(),
-            progress, // Passed through
-        };
-        self.broker.publish_event(request_id, event).await
-    }
-}
-
-// --- Tests ---
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-    use tokio::sync::RwLock;
-    use std::sync::Arc;
-    use common_contracts::symbol_utils::Market;
-    use common_contracts::subjects::NatsSubject;
-
-    // --- Fakes ---
-
-    struct InMemoryRepo {
-        data: Arc<RwLock<HashMap<Uuid, WorkflowStateMachine>>>
-    }
-    
-    #[async_trait]
-    impl WorkflowRepository for InMemoryRepo {
-        async fn save(&self, workflow: &WorkflowStateMachine) -> Result<()> {
-            self.data.write().await.insert(workflow.request_id, workflow.clone());
-            Ok(())
-        }
-        async fn load(&self, id: Uuid) -> Result<Option<WorkflowStateMachine>> {
-            Ok(self.data.read().await.get(&id).cloned())
-        }
-    }
-
-    struct FakeBroker {
-        pub events: Arc<RwLock<Vec<(Uuid, WorkflowEvent)>>>,
-        pub commands: Arc<RwLock<Vec<(String, Vec<u8>)>>>,
-    }
-
-    #[async_trait]
-    impl MessageBroker for FakeBroker {
-        async fn publish_event(&self, request_id: Uuid, event: WorkflowEvent) -> Result<()> {
-            self.events.write().await.push((request_id, event));
-            Ok(())
-        }
-        async fn publish_command(&self, topic: &str, payload: Vec<u8>) -> Result<()> {
-            self.commands.write().await.push((topic.to_string(), payload));
-            Ok(())
-        }
-    }
-
-    // --- Test Cases ---
-
-    #[tokio::test]
-    async fn test_dag_execution_flow() {
-        // 1. Setup
-        let repo = Box::new(InMemoryRepo { data: Arc::new(RwLock::new(HashMap::new())) });
-        
-        let events = Arc::new(RwLock::new(Vec::new()));
-        let commands = Arc::new(RwLock::new(Vec::new()));
-        let broker = Box::new(FakeBroker { events: events.clone(), commands: commands.clone() });
-        
-        let engine = WorkflowEngine::new(repo, broker);
-        
-        let req_id = Uuid::new_v4();
-        let symbol = CanonicalSymbol::new("AAPL", &Market::US);
-
-        // 2. Start Workflow
-        let cmd = StartWorkflowCommand {
-            request_id: req_id,
-            symbol: symbol.clone(),
-            market: "US".to_string(),
-            template_id: "default".to_string(),
-        };
-        engine.handle_start_workflow(cmd).await.unwrap();
-
-        // Assert: Started Event & Fetch Commands
-        {
-            let evs = events.read().await;
-            assert!(matches!(evs[0].1, WorkflowEvent::WorkflowStarted { .. }));
-            
-            let cmds = commands.read().await;
-            assert_eq!(cmds.len(), 4); // 4 fetchers
-            assert_eq!(cmds[0].0, NatsSubject::DataFetchCommands.to_string());
-        }
-
-        // 3. Simulate Completion (Alphavantage)
-        engine.handle_data_fetched(FinancialsPersistedEvent {
-            request_id: req_id,
-            symbol: symbol.clone(),
-            years_updated: vec![],
-            template_id: Some("default".to_string()),
-            provider_id: Some("alphavantage".to_string()),
-            data_summary: None,
-        }).await.unwrap();
-
-        // Assert: Task Completed
-        {
-            let evs = events.read().await;
-            // Started(1) + 4 Scheduled(4) + StreamUpdate(1) + Completed(1) = 7
-            // Wait, order depends.
-            // Just check last event
-            if let WorkflowEvent::TaskStateChanged { task_id, status, .. } = &evs.last().unwrap().1 {
-                assert_eq!(task_id, "fetch:alphavantage");
-                assert_eq!(*status, TaskStatus::Completed);
-            } else {
-                panic!("Last event should be state change");
-            }
-        }
-
-        // 4. Simulate Completion of ALL fetchers to trigger Analysis
-        let providers = vec!["tushare", "finnhub", "yfinance"];
-        for p in providers {
-            engine.handle_data_fetched(FinancialsPersistedEvent {
-                request_id: req_id,
-                symbol: symbol.clone(),
-                years_updated: vec![],
-                template_id: Some("default".to_string()),
-                provider_id: Some(p.to_string()),
-                data_summary: None,
-            }).await.unwrap();
-        }
-
-        // Assert: Analysis Triggered
-        {
-            let cmds = commands.read().await;
-            // 4 fetch commands + 1 analysis command
-            assert_eq!(cmds.len(), 5);
-            assert_eq!(cmds.last().unwrap().0, NatsSubject::AnalysisCommandGenerateReport.to_string());
-        }
-    }
-
-    #[tokio::test]
-    async fn test_simple_analysis_flow() {
-        // 1. Setup
-        let repo = Box::new(InMemoryRepo { data: Arc::new(RwLock::new(HashMap::new())) });
-        let events = Arc::new(RwLock::new(Vec::new()));
-        let commands = Arc::new(RwLock::new(Vec::new()));
-        let broker = Box::new(FakeBroker { events: events.clone(), commands: commands.clone() });
-        let engine = WorkflowEngine::new(repo, broker);
-        
-        let req_id = Uuid::new_v4();
-        let symbol = CanonicalSymbol::new("AAPL", &Market::US);
-
-        // 2. Start Workflow with "simple_test_analysis"
-        let cmd = StartWorkflowCommand {
-            request_id: req_id,
-            symbol: symbol.clone(),
-            market: "US".to_string(),
-            template_id: "simple_test_analysis".to_string(),
-        };
-        engine.handle_start_workflow(cmd).await.unwrap();
-
-        // Assert: DAG Structure (Fetch yfinance -> Analysis)
-        // Check events for WorkflowStarted
-        {
-            let evs = events.read().await;
-            if let WorkflowEvent::WorkflowStarted { task_graph, .. } = &evs[0].1 {
-                assert_eq!(task_graph.nodes.len(), 2);
-                assert!(task_graph.nodes.iter().any(|n| n.id == "fetch:yfinance"));
-                assert!(task_graph.nodes.iter().any(|n| n.id == "analysis:report"));
-            } else {
-                panic!("First event must be WorkflowStarted");
-            }
-        }
-
-        // 3. Simulate Fetch Completion
-        engine.handle_data_fetched(FinancialsPersistedEvent {
-            request_id: req_id,
-            symbol: symbol.clone(),
-            years_updated: vec![],
-            template_id: Some("simple_test_analysis".to_string()),
-            provider_id: Some("yfinance".to_string()),
-            data_summary: None,
-        }).await.unwrap();
-
-        // Assert: Analysis Scheduled
-        {
-            let evs = events.read().await;
-            let last_event = &evs.last().unwrap().1;
-            if let WorkflowEvent::TaskStateChanged { task_id, status, .. } = last_event {
-                assert_eq!(task_id, "analysis:report");
-                assert_eq!(*status, TaskStatus::Scheduled);
-            } else {
-                panic!("Expected TaskStateChanged to Scheduled");
-            }
-        }
-
-        // 4. Simulate Report Generated (Analysis Completion)
-        // Using the mismatch ID "step2_analyze" to verify our mapping logic
-        engine.handle_report_generated(common_contracts::messages::ReportGeneratedEvent {
-            request_id: req_id,
-            symbol: symbol.clone(),
-            module_id: "step2_analyze".to_string(),
-            content_snapshot: None,
-            model_id: None,
-        }).await.unwrap();
-
-        // Assert: Workflow Completed
-        {
-            let evs = events.read().await;
-            let last_event = &evs.last().unwrap().1;
-            if let WorkflowEvent::WorkflowCompleted { .. } = last_event {
-                // Success!
-            } else {
-                // Check if we have task completion before workflow completion
-                // The very last event should be WorkflowCompleted
-                // Let's print all events
-                for (_, e) in evs.iter() {
-                    println!("{:?}", e);
-                }
-                panic!("Workflow did not complete. Last event: {:?}", last_event);
-            }
+        // 3. Edges
+        for p in &providers {
+            dag.add_dependency(&format!("fetch:{}", p), report_task_id);
         }
     }
 }