fix(janitor): 禁止启动时因目录不存在而自动删除数据库任务记录

reconcile_tasks_on_startup() 中有一段逻辑：遍历所有任务，如果 output_dir
在磁盘上不存在，直接 DELETE FROM tasks。这导致重启时如果目录被移动/重命名/
被之前的 janitor 清理，所有任务记录静默丢失。

修改为仅输出 WARNING 日志，保留数据库记录。用户可通过 DELETE /task/<id> 手动清理。

build_exe.bat

@@ -5,19 +5,19 @@ REM This script builds a standalone executable using Waitress WSGI server
 echo Building GasFlux Web API executable...
 
 REM Check if PyInstaller is installed
-python -c "import PyInstaller" >nul 2>&1
+py -3 -c "import PyInstaller" >nul 2>&1
 if errorlevel 1 (
     echo Error: PyInstaller is not installed. Please run:
-    echo pip install pyinstaller waitress
+    echo   py -3 -m pip install pyinstaller waitress
     pause
     exit /b 1
 )
 
 REM Check if Waitress is installed
-python -c "import waitress" >nul 2>&1
+py -3 -c "import waitress" >nul 2>&1
 if errorlevel 1 (
     echo Error: Waitress is not installed. Please run:
-    echo pip install waitress
+    echo   py -3 -m pip install waitress
     pause
     exit /b 1
 )
@@ -41,8 +41,24 @@ pyinstaller --onefile ^
     --hidden-import skgstat ^
     --hidden-import skgstat.Variogram ^
     --hidden-import skgstat.OrdinaryKriging ^
-    --add-data "src\gasflux\gasflux_config.yaml;src\gasflux" ^
+    --hidden-import skgstat.DirectionalVariogram ^
-    --add-data "API_DOCUMENTATION.md;." ^
+    --hidden-import plotly ^
+    --hidden-import plotly.graph_objects ^
+    --hidden-import plotly.express ^
+    --hidden-import plotly.subplots ^
+    --hidden-import pybaselines ^
+    --hidden-import scipy ^
+    --hidden-import scipy.odr ^
+    --hidden-import scipy.signal ^
+    --hidden-import scipy.stats ^
+    --hidden-import scipy.integrate ^
+    --hidden-import scipy.optimize ^
+    --hidden-import molmass ^
+    --hidden-import openpyxl ^
+    --hidden-import pyproj ^
+    --hidden-import jinja2 ^
+    --hidden-import requests ^
+    --hidden-import urllib3 ^
     --hidden-import matplotlib ^
     --hidden-import matplotlib.pyplot ^
     --hidden-import matplotlib.backends ^
@@ -53,7 +69,19 @@ pyinstaller --onefile ^
     --hidden-import matplotlib.patches ^
     --hidden-import matplotlib.text ^
     --hidden-import matplotlib.transforms ^
+    --hidden-import tqdm ^
+    --hidden-import certifi ^
+    --hidden-import charset_normalizer ^
+    --hidden-import geopandas ^
+    --hidden-import shapely ^
+    --hidden-import fiona ^
+    --hidden-import simplekml ^
+    --hidden-import joblib ^
     --exclude-module tkinter ^
+    --add-data "src\gasflux\gasflux_config.yaml;src\gasflux" ^
+    --add-data "src\gasflux\templates\mass_balance_template.html;src\gasflux\templates" ^
+    --add-data "gasflux.ini;." ^
+    --add-data "gasflux.ini.example;." ^
     server_waitress.py
 
 if errorlevel 1 (
@@ -69,6 +97,8 @@ echo.
 echo To run the server:
 echo   GasFluxAPI.exe
 echo.
-echo The server will start on http://localhost:5000
+echo The server will start on http://localhost:5001
+echo Make sure gasflux.ini is in the same directory as the executable,
+echo or configure paths via environment variables.
 echo.
 pause

gasflux.ini

@@ -30,9 +30,9 @@ task_cleanup_interval = 30
 # 24 hours in seconds
 max_task_age = 86400
 # 1 minute in seconds for successful tasks
-successful_task_cleanup_age = 60
+successful_task_cleanup_age = 315360000
 # 1 minute in seconds for failed tasks
-failed_task_cleanup_age = 60
+failed_task_cleanup_age = 315360000
 janitor_dry_run = false
 
 [performance]

server_waitress.py

@@ -45,9 +45,17 @@ import sys
 from pathlib import Path
 
 # 设置控制台编码为 UTF-8（解决 Windows 中文乱码问题）
+# PyInstaller onefile 模式下 sys.stdout.buffer 可能不存在（无控制台或管道重定向），
+# 此时跳过编码包装，避免 AttributeError 导致启动崩溃
 import io
-sys.stdout = io.TextIOWrapper(sys.stdout.buffer, encoding='utf-8')
+try:
-sys.stderr = io.TextIOWrapper(sys.stderr.buffer, encoding='utf-8')
+    sys.stdout = io.TextIOWrapper(sys.stdout.buffer, encoding='utf-8')
+except (AttributeError, ValueError):
+    pass
+try:
+    sys.stderr = io.TextIOWrapper(sys.stderr.buffer, encoding='utf-8')
+except (AttributeError, ValueError):
+    pass
 
 # Add the project root and src directory to PYTHONPATH
 project_root = Path(__file__).parent.absolute()

src/gasflux/blueprints/download.py

@@ -3,6 +3,7 @@ Download Blueprint
 Handles file download endpoints.
 """
 
+from datetime import datetime
 from pathlib import Path
 from flask import Blueprint, send_file, current_app
 
@@ -10,40 +11,6 @@ from ..shared import _format_response, log_performance, logger
 from ..auth import require_api_key
 
 
-def _mark_task_downloaded(task_id):
-    """Mark task as downloaded and schedule deletion in database."""
-    import sqlite3
-    from pathlib import Path
-
-    # Use independent database connection (not from flask.g which may be closed)
-    from ..db import get_db_path as get_config_db_path
-    db_path = get_config_db_path(current_app)
-
-    # Get cleanup age for successful tasks from config (in seconds)
-    successful_task_cleanup_age = current_app.config.get('SUCCESSFUL_TASK_CLEANUP_AGE', 3600)
-
-    try:
-        conn = sqlite3.connect(str(db_path), check_same_thread=False)
-        conn.execute("PRAGMA foreign_keys=ON")
-        conn.execute("PRAGMA busy_timeout=3000")
-
-        # Update downloaded timestamp and set deletion time based on config
-        conn.execute("""
-            UPDATE tasks
-            SET downloaded_at = datetime('now', '+8 hours'),
-                delete_after_at = datetime('now', '+8 hours', '+' || ? || ' seconds')
-            WHERE task_id = ?
-        """, (successful_task_cleanup_age, task_id))
-
-        conn.commit()
-        logger.info(f"Task {task_id} marked as downloaded, scheduled for deletion in {successful_task_cleanup_age} seconds")
-
-    except Exception as e:
-        logger.error(f"Failed to mark task {task_id} as downloaded: {str(e)}", exc_info=True)
-    finally:
-        if 'conn' in locals():
-            conn.close()
-
 # Create blueprint
 download_bp = Blueprint('download', __name__, url_prefix='/download')
 
@@ -124,12 +91,19 @@ def download_file(filename):
         file_size = file_path.stat().st_size
         logger.info(f"Serving file: {filename} ({file_size} bytes)")
 
-        # Mark download immediately before sending file
+        # 记录下载时间（仅时间戳，不设置自动删除）
         if task_id:
             try:
-                _mark_task_downloaded(task_id)
+                from ..db import get_db
+                db = get_db()
+                db.execute(
+                    "UPDATE tasks SET downloaded_at = datetime('now', '+8 hours') WHERE task_id = ?",
+                    (task_id,)
+                )
+                db.commit()
+                logger.info(f"Task {task_id} downloaded at {datetime.now()}")
             except Exception as e:
-                logger.error(f"Failed to mark download for task {task_id}: {str(e)}")
+                logger.error(f"Failed to record download for task {task_id}: {str(e)}")
 
         response = send_file(file_path)
         return response

src/gasflux/interpolation.py

@@ -99,11 +99,6 @@ def ordinary_kriging(
     # np.nan_to_num(error_1s, copy=False, nan=0)
     volume_error = simpsonintegrate(error_1s, x_cell_size, y_cell_size)
 
-    # Plots disabled
-    contour_plot = None
-    grid_plot = None
-    semivariogram_plot = None
-
     output_text = (
         f"The emissions flux of {gas.upper()} is {volume:.3f}kgh⁻¹; "
         f"the cut and fill volumes of the grid are {volumepos:.3f} and {volumeneg:.3f}kgh⁻¹. "
@@ -123,6 +118,20 @@ def ordinary_kriging(
         "volume_error": volume_error,
     }
 
+    # Generate plots (must be after krig_variables is defined)
+    try:
+        contour_plot = plotting._contour_krig_wrapper(krig_variables)
+    except Exception:
+        contour_plot = None
+    try:
+        grid_plot = plotting._heatmap_krig_wrapper(krig_variables)
+    except Exception:
+        grid_plot = None
+    try:
+        semivariogram_plot = plotting._semivariogram_plot(semivariogram, gas)
+    except Exception:
+        semivariogram_plot = None
+
     return krig_variables, output_text, contour_plot, grid_plot, semivariogram_plot
 
 

src/gasflux/janitor.py

@@ -310,8 +310,7 @@ def reconcile_tasks_on_startup():
                 except Exception as e:
                     current_app.logger.error(f"Failed to delete failed task {task_id}: {str(e)}", exc_info=True)
 
-        # Check for tasks with output directories that no longer exist
+        # 检查 output_dir 不存在的任务（仅警告，不自动删除）
-        # This helps clean up database entries for manually deleted directories
         rows = conn.execute("""
             SELECT task_id, output_dir
             FROM tasks
@@ -321,18 +320,21 @@ def reconcile_tasks_on_startup():
 
         orphaned_count = 0
         for row in rows:
-            task_id_from_db = row[0]  # task_id是第一个字段
+            task_id_from_db = row[0]
-            output_dir_from_db = row[1]  # output_dir是第二个字段
+            output_dir_from_db = row[1]
 
             if not Path(output_dir_from_db).exists():
-                conn.execute(
+                current_app.logger.warning(
-                    "DELETE FROM tasks WHERE task_id = ?",
+                    f"Startup reconciliation: Task {task_id_from_db} output directory "
-                    (task_id_from_db,)
+                    f"not found on disk: {output_dir_from_db}. Database record preserved."
                 )
                 orphaned_count += 1
 
         if orphaned_count > 0:
-            current_app.logger.info(f"Startup reconciliation: Marked {orphaned_count} tasks as deleted (directories not found)")
+            current_app.logger.info(
+                f"Startup reconciliation: Found {orphaned_count} tasks with missing directories "
+                f"(records preserved, use DELETE /task/<id> to manually clean up)"
+            )
 
         # Backfill output_dir for rows with NULL, using OUTPUT_FOLDER/<task_id>
         output_base = Path(current_app.config.get('OUTPUT_FOLDER') or '')

src/gasflux/pre_processing.py

@@ -129,7 +129,11 @@ def remove_outliers(df: pd.DataFrame, column: str, name: str):
     fence_low = q1 - 3 * iqr
     fence_high = q3 + 3 * iqr
 
-    fig = None  # plotting disabled
+    try:
+        from . import plotting
+        fig = plotting._outliers_wrapper(df, column, name)
+    except Exception:
+        fig = None
 
     outliers = df.loc[(df[column] < fence_low) | (df[column] > fence_high)]
     if len(outliers) > 0:

src/gasflux/processing_pipelines.py

@@ -5,6 +5,7 @@ from scipy import stats
 
 import pandas as pd
 import yaml
+import plotly.graph_objects as go
 
 from src.gasflux import background,plotting,processing,reporting,interpolation,pre_processing,gas
 
@@ -123,9 +124,22 @@ class InSituSensorStrategy(SensorStrategy):
     def process(self):
         logger.info("Processing in-situ (point) data")
         for gas in self.data_processor.gases:
-            self.data_processor.figs["scatter_3d"][gas] = None
+            # Create 3D scatter plot of flight path with gas concentration
-        self.data_processor.figs["windrose"] = None
+            try:
-        self.data_processor.figs["wind_timeseries"] = None
+                self.data_processor.figs["scatter_3d"][gas] = plotting._scatter_3d_wrapper(
+                    self.data_processor.df, gas
+                )
+            except Exception:
+                self.data_processor.figs["scatter_3d"][gas] = plotting.blank_figure()
+        # Create wind rose and time series plots
+        try:
+            self.data_processor.figs["windrose"] = plotting._windrose_wrapper(self.data_processor.df)
+        except Exception:
+            self.data_processor.figs["windrose"] = plotting.blank_figure()
+        try:
+            self.data_processor.figs["wind_timeseries"] = plotting._time_series_wrapper(self.data_processor.df)
+        except Exception:
+            self.data_processor.figs["wind_timeseries"] = plotting.blank_figure()
 
 
 class SpatialProcessingStrategy(ABC):

src/gasflux/reporting.py

@@ -27,13 +27,22 @@ def mass_balance_report(
     """Generate a mass balance report (plots disabled)."""
     template_path = Path(__file__).parent / "templates" / "mass_balance_template.html"
 
-    # Plots disabled -> use empty strings
+    # Convert figures to HTML strings
+    def _fig_to_html(fig) -> str:
+        """Safely convert a plotly figure to HTML string."""
+        if fig is None:
+            return ""
+        try:
+            return fig.to_html(full_html=False, include_plotlyjs=True)
+        except Exception:
+            return ""
+
     plot_htmls = {
-        "3D": "",
+        "3D": _fig_to_html(threed_fig),
-        "krig": "",
+        "krig": _fig_to_html(krig_fig),
-        "windrose": "",
+        "windrose": _fig_to_html(windrose_fig),
-        "wind": "",
+        "wind": _fig_to_html(wind_fig),
-        "background": "",
+        "background": _fig_to_html(background_fig),
     }
 
     summary_data = {