Merge pull request #12 from SongY123/main

support electron build

Author: SongY123

.github/workflows/package-electron.yml

@@ -0,0 +1,51 @@
+name: Package Electron Desktop
+
+on:
+  push:
+    branches:
+      - main
+      - master
+  workflow_dispatch:
+
+jobs:
+  package-desktop:
+    runs-on: macos-latest
+
+    steps:
+      - name: Checkout backend repository
+        uses: actions/checkout@v4
+
+      - name: Set up Node.js
+        uses: actions/setup-node@v4
+        with:
+          node-version: "20"
+
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.11"
+          cache: "pip"
+          cache-dependency-path: requirements.txt
+
+      - name: Install backend dependencies
+        run: python -m pip install -r requirements.txt
+
+      - name: Package Electron app
+        env:
+          DATAFACTORY_FRONTEND_DIR: ${{ github.workspace }}/../DataFactory-Frontend
+          DATAFACTORY_FRONTEND_REPO: https://github.com/${{ github.repository_owner }}/DataFactory-Frontend.git
+          DATAFACTORY_FRONTEND_UPDATE: "0"
+        run: bash scripts/package_desktop_app.sh
+
+      - name: Prepare artifact directory
+        run: |
+          rm -rf desktop_artifacts
+          mkdir -p desktop_artifacts
+          cp -R ../DataFactory-Frontend/dist_electron/. desktop_artifacts/
+
+      - name: Upload desktop artifact
+        uses: actions/upload-artifact@v4
+        with:
+          name: datafactory-desktop-mac
+          path: desktop_artifacts/**
+          if-no-files-found: error

.gitignore

@@ -16,4 +16,5 @@ outputs/
 
 uploads/
 
-output/
+output/
+workspace/

README.md

@@ -1,14 +1,13 @@
 # DataFactory
 
-This project is a trimmed migration of the main `SpatialText2SQL` web architecture, preserving the layered structure: `api/service/dao/entity/resources`.
+This project preserves a layered web architecture with `api/service/dao/entity/resources`.
 
 Currently retained:
 - Local SQLite initialization (SQL migrations run automatically at startup)
 - User authentication and session flow (`/api/auth/login`, `/api/auth/logout`, `/api/auth/session`)
 - User management APIs (`/api/users`)
 
 Removed:
-- Map/spatial-related features
 - Multi-agent features
 - Other business features such as chat, admin, and database probing/execution
 

requirements.txt

@@ -12,3 +12,6 @@ ollama>=0.6,<1.0
 dashscope>=1.25,<2.0
 python-docx>=1.1,<2.0
 huggingface_hub>=0.34,<1.0
+statistics
+numpy
+pyarrow

runtime/sandbox_environments.json

@@ -4,16 +4,16 @@
     {
       "id": "env-default",
       "name": "Default Python",
-      "python_path": "C:\\Users\\86138\\.conda\\envs\\DataFactory-Backend\\python.exe",
+      "python_path": "/opt/anaconda3/envs/DataFactory/bin/python3.10",
       "created_at": "2026-03-22T01:02:59.073077+00:00",
-      "updated_at": "2026-03-22T01:02:59.073077+00:00"
+      "updated_at": "2026-03-24T15:46:56.188560+00:00"
     },
     {
       "id": "env-d3934326cd",
       "name": "DataFactory",
-      "python_path": "D:\\Anaconda\\envs\\DataFactory\\python.exe",
+      "python_path": "/opt/anaconda3/envs/DataFactory/bin/python3.10",
       "created_at": "2026-03-23T13:38:55.175962+00:00",
       "updated_at": "2026-03-23T13:38:55.175962+00:00"
     }
   ]
-}
+}

scripts/package_desktop_app.sh

@@ -0,0 +1,42 @@
+#!/usr/bin/env bash
+
+set -euo pipefail
+
+BACKEND_ROOT="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
+WORKSPACE_ROOT="$(cd "${BACKEND_ROOT}/.." && pwd)"
+FRONTEND_DIR="${DATAFACTORY_FRONTEND_DIR:-${WORKSPACE_ROOT}/DataFactory-Frontend}"
+FRONTEND_REPO="${DATAFACTORY_FRONTEND_REPO:-https://github.com/SongY123/DataFactory-Frontend.git}"
+FRONTEND_UPDATE="${DATAFACTORY_FRONTEND_UPDATE:-0}"
+
+echo "[package] backend root: ${BACKEND_ROOT}"
+echo "[package] frontend dir: ${FRONTEND_DIR}"
+
+if [[ -d "${FRONTEND_DIR}/.git" ]]; then
+  if [[ "${FRONTEND_UPDATE}" == "1" ]]; then
+    if ! git -C "${FRONTEND_DIR}" diff --quiet || ! git -C "${FRONTEND_DIR}" diff --cached --quiet; then
+      echo "[package] frontend repo has local changes; refusing to pull with DATAFACTORY_FRONTEND_UPDATE=1" >&2
+      exit 1
+    fi
+    echo "[package] updating existing frontend checkout"
+    git -C "${FRONTEND_DIR}" pull --ff-only
+  else
+    echo "[package] using existing frontend checkout without pulling"
+  fi
+else
+  echo "[package] cloning frontend repo: ${FRONTEND_REPO}"
+  git clone "${FRONTEND_REPO}" "${FRONTEND_DIR}"
+fi
+
+pushd "${FRONTEND_DIR}" >/dev/null
+
+echo "[package] installing frontend dependencies"
+npm ci
+
+echo "[package] packaging Electron desktop app"
+npm run desktop:package
+
+echo "[package] build outputs"
+find "${FRONTEND_DIR}/dist_electron" -maxdepth 3 -print | sed -n '1,120p'
+
+popd >/dev/null
+

src/utils/auth_guard.py

@@ -20,9 +20,9 @@ def _get_session_service():
     return get_global_session_service()
 
 
-def _is_auto_login_as_admin_enabled() -> bool:
+def _is_auto_login_as_user_enabled() -> bool:
     try:
-        raw = get_config("auth.auto_login_as_admin", False)
+        raw = get_config("auth.auto_login_as_user", False)
     except Exception:
         return False
 
@@ -73,7 +73,7 @@ def _resolve_auto_login_payload() -> Dict[str, Any]:
         # Delay import to avoid unnecessary dependency initialization at module import time.
         from web.service.user_service import UserService
 
-        login_data = UserService().login(username="admin", password="admin")
+        login_data = UserService().login(username="user", password="user")
         session_id = str(login_data.get("session_id") or "").strip()
         if not session_id:
             raise RuntimeError("auto login failed to create a session id")
@@ -95,7 +95,7 @@ def _get_auto_login_user() -> Dict[str, Any]:
     except Exception as exc:
         raise HTTPException(
             status_code=500,
-            detail=f"Auto login as admin failed: {exc}",
+            detail=f"Auto login as user failed: {exc}",
         )
 
 
@@ -107,7 +107,7 @@ def get_login_user(request: Request) -> Dict[str, Any]:
         if payload:
             return _normalize_login_user(payload, session_id)
 
-    if _is_auto_login_as_admin_enabled():
+    if _is_auto_login_as_user_enabled():
         return _get_auto_login_user()
 
     raise HTTPException(status_code=401, detail="Not logged in.")

src/web/api/__init__.py

@@ -5,6 +5,8 @@
 from .agentic_synthesis_api import router as agentic_synthesis_router
 from .reasoning_distillation_api import router as reasoning_distillation_router
 from .sandbox_environment_api import router as sandbox_environment_router
+from .user_preference_api import router as user_preference_router
+from .workflow_assistant_api import router as workflow_assistant_router
 
 __all__ = [
     "user_router",
@@ -15,4 +17,6 @@
     "agentic_synthesis_router",
     "reasoning_distillation_router",
     "sandbox_environment_router",
+    "user_preference_router",
+    "workflow_assistant_router",
 ]

src/web/api/chat_api.py

@@ -372,6 +372,25 @@ def delete_file(request: Request, path: str = Query(...)):
     }
 
 
+@router.get("/files/preview")
+def preview_file(
+    request: Request,
+    path: str = Query(...),
+    page: int = Query(1, ge=1),
+    page_size: int = Query(200, ge=1, le=200),
+):
+    user = get_login_user(request)
+    try:
+        payload = _asset_service.preview_file_page(int(user["id"]), path, page=page, page_size=page_size)
+    except ValueError as exc:
+        raise HTTPException(status_code=400, detail=str(exc)) from exc
+
+    return {
+        "ok": True,
+        **payload,
+    }
+
+
 @router.post("/upload")
 async def upload_file(request: Request, file: UploadFile = File(...), folder_path: str = Form("")):
     user = get_login_user(request)

src/web/api/reasoning_distillation_api.py

@@ -29,7 +29,13 @@ def start_reasoning_distillation_task(request: Request, body: ReasoningDistillat
             source_type=body.source_type,
             source_dataset_id=body.source_dataset_id,
             source_task_id=body.source_task_id,
+            selected_file_paths=body.selected_file_paths,
+            file_mappings=[item.model_dump() for item in body.file_mappings],
+            prompt_field=body.prompt_field,
+            completion_field=body.completion_field,
             prompt=body.prompt,
+            evaluation_enabled=body.evaluation_enabled,
+            evaluation_prompt=body.evaluation_prompt,
             strategy=body.strategy,
             target_max_tokens=body.target_max_tokens,
             compression_ratio=body.compression_ratio,