Refactor app.py to integrate threshold diagnosis and trend detection features, enhancing Gradio UI for improved user experience.

app.py

@@ -1,11 +1,13 @@
-# app.py
+# 統合版: 閾値診断 + 傾向検出 (CSV/Excel共通入力, MCP対応)
+
 import gradio as gr
 import pandas as pd
 import numpy as np
+from sklearn.linear_model import LinearRegression
 import json
 import os
 
-# --- 共通ユーティリティ ---
+# ===== 共通ユーティリティ =====
 def judge_status(value, ll, l, h, hh):
     if pd.notna(ll) and value < ll:
         return "LOW-LOW"
@@ -23,23 +25,8 @@ def convert_value(v):
         return v.item()
     return float(v) if isinstance(v, (np.floating, float)) else int(v) if isinstance(v, (np.integer, int)) else v
 
-# --- 閾値診断（本物処理） ---
-def diagnose_process_range(csv_file, excel_file, process_name, datetime_str, window_minutes):
-    try:
-        df = pd.read_csv(csv_file.name, header=[0, 1, 2])
-        timestamp_col = df.iloc[:, 0]
-        df = df.drop(df.columns[0], axis=1)
-        df.insert(0, "timestamp", timestamp_col)
-        df["timestamp"] = pd.to_datetime(df["timestamp"], errors="coerce")
-
-        thresholds_df = pd.read_excel(excel_file.name)
-        thresholds_df["Important"] = thresholds_df["Important"].astype(str).str.upper().map({"TRUE": True, "FALSE": False})
-        for col in ["LL", "L", "H", "HH"]:
-            if col in thresholds_df.columns:
-                thresholds_df[col] = pd.to_numeric(thresholds_df[col], errors="coerce")
-    except Exception as e:
-        return None, None, None, f"❌ 入力ファイルの読み込みに失敗しました: {e}", None
-
+# ===== 閾値診断処理 =====
+def diagnose_process_range(df, thresholds_df, process_name, datetime_str, window_minutes):
     try:
         target_time = pd.to_datetime(datetime_str)
     except Exception:
@@ -76,7 +63,11 @@ def diagnose_process_range(csv_file, excel_file, process_name, datetime_str, win
         ["LOW-LOW", "LOW", "OK", "HIGH", "HIGH-HIGH"], fill_value=0
     )
     status_ratio = (status_counts / total * 100).round(1)
-    result_df_all = pd.DataFrame({"状態": status_counts.index, "件数": status_counts.values, "割合(%)": status_ratio.values})
+    result_df_all = pd.DataFrame({
+        "状態": status_counts.index,
+        "件数": status_counts.values,
+        "割合(%)": status_ratio.values
+    })
 
     important_results = [r for r in all_results if r["重要項目"]]
     if important_results:
@@ -85,7 +76,11 @@ def diagnose_process_range(csv_file, excel_file, process_name, datetime_str, win
             ["LOW-LOW", "LOW", "OK", "HIGH", "HIGH-HIGH"], fill_value=0
         )
         status_ratio_imp = (status_counts_imp / total_imp * 100).round(1)
-        result_df_imp = pd.DataFrame({"状態": status_counts_imp.index, "件数": status_counts_imp.values, "割合(%)": status_ratio_imp.values})
+        result_df_imp = pd.DataFrame({
+            "状態": status_counts_imp.index,
+            "件数": status_counts_imp.values,
+            "割合(%)": status_ratio_imp.values
+        })
     else:
         result_df_imp = pd.DataFrame(columns=["状態", "件数", "割合(%)"])
         status_ratio_imp = pd.Series(dtype=float)
@@ -126,32 +121,122 @@ def diagnose_process_range(csv_file, excel_file, process_name, datetime_str, win
 
     return result_df_all, result_df_imp, result_df_imp_items, summary, result_json
 
+# ===== 傾向検出処理 =====
+def detect_trends_with_forecast(df, thresholds_df, process_name, datetime_str, window_minutes, forecast_minutes):
+    target_time = pd.to_datetime(datetime_str)
+    start_time = target_time - pd.Timedelta(minutes=window_minutes)
+    end_time = target_time
+    df_window = df[(df["timestamp"] >= start_time) & (df["timestamp"] <= end_time)]
+    if df_window.empty:
+        return None, "⚠ 指定時間幅にデータなし", None
+
+    interval = df_window["timestamp"].diff().median()
+    if pd.isna(interval):
+        return None, "⚠ サンプリング間隔を検出できません", None
+    interval_minutes = interval.total_seconds() / 60
+
+    proc_thresholds = thresholds_df[(thresholds_df["ProcessNo_ProcessName"] == process_name) &
+                                    (thresholds_df["Important"] == True)]
+    if proc_thresholds.empty:
+        return None, f"⚠ プロセス {process_name} の重要項目なし", None
+
+    results = []
+    for _, thr in proc_thresholds.iterrows():
+        col_tuple = (thr["ColumnID"], thr["ItemName"], thr["ProcessNo_ProcessName"])
+        if col_tuple not in df.columns:
+            continue
+        series = df_window[col_tuple].dropna()
+        if len(series) < 3:
+            continue
+
+        x = np.arange(len(series)).reshape(-1, 1)
+        y = series.values.reshape(-1, 1)
+        model = LinearRegression().fit(x, y)
+        slope = model.coef_[0][0]
+
+        last_val = series.iloc[-1]
+        n = len(series)
+
+        forecast_steps = int(forecast_minutes / interval_minutes)
+        forecast_index = n + forecast_steps
+        forecast_val = model.predict([[forecast_index]])[0][0]
+        forecast_time = target_time + pd.Timedelta(minutes=forecast_minutes)
+
+        l, ll, h, hh = thr.get("L"), thr.get("LL"), thr.get("H"), thr.get("HH")
+
+        status = "安定"
+        if slope < 0 and pd.notna(ll):
+            if last_val > ll:
+                status = "LL接近下降傾向"
+            elif last_val <= ll:
+                status = "LL逸脱下降傾向"
+        if slope > 0 and pd.notna(hh):
+            if last_val < hh:
+                status = "HH接近上昇傾向"
+            elif last_val >= hh:
+                status = "HH逸脱上昇傾向"
+
+        forecast_status = "安定"
+        if pd.notna(ll) and forecast_val <= ll:
+            forecast_status = "LL逸脱予測"
+        elif pd.notna(hh) and forecast_val >= hh:
+            forecast_status = "HH逸脱予測"
+
+        results.append({
+            "ItemName": thr["ItemName"],
+            "傾向": status,
+            "傾き": round(slope, 4),
+            "最終値": round(float(last_val), 3) if pd.notna(last_val) else None,
+            "予測値": round(float(forecast_val), 3),
+            "予測時刻": str(forecast_time),
+            "予測傾向": forecast_status,
+            "サンプリング間隔(分)": interval_minutes,
+            "LL": ll, "L": l, "H": h, "HH": hh
+        })
 
-# --- 傾向検出（ダミー） ---
-def dummy_trend(csv_file, excel_file, process_name, datetime_str, window_minutes, forecast_minutes):
-    return pd.DataFrame([{"ItemName": "ダミー項目", "傾向": "安定"}]), "✅ ダミー傾向", {"dummy": True}
+    result_df = pd.DataFrame(results)
+    result_json = json.dumps(results, ensure_ascii=False, indent=2)
+    return result_df, "✅ 傾向検出＋未来予測完了", result_json
 
-# --- 予兆解析（ダミー） ---
-def dummy_forecast(csv_file, excel_file, process_name, datetime_str, forecast_minutes):
-    return pd.DataFrame([{"ItemName": "ダミー項目", "予測": "OK"}]), "✅ ダミー予兆", {"dummy": True}
+# ===== Gradio UI =====
+with gr.Blocks(css=".gradio-container {overflow:auto !important}") as demo:
+    gr.Markdown("## 統合版アプリ (閾値診断 + 傾向検出)")
 
+    with gr.Row():
+        csv_input = gr.File(label="CSVファイルをアップロード", file_types=[".csv"], type="filepath")
+        excel_input = gr.File(label="Excel閾値ファイルをアップロード", file_types=[".xlsx"], type="filepath")
 
-# --- Gradio UI (統合フレーム Step 2) ---
-with gr.Blocks(css="body {overflow-y: auto !important;}") as demo:
-    gr.Markdown("## 統合解析ツール (Step 2: 閾値診断のみ本物処理)")
+    # ファイルを保持するステート
+    state_df = gr.State()
+    state_thresholds = gr.State()
+
+    def load_files(csv_file, excel_file):
+        try:
+            df = pd.read_csv(csv_file.name, header=[0,1,2])
+            timestamp_col = df.iloc[:,0]
+            df = df.drop(df.columns[0], axis=1)
+            df.insert(0,"timestamp", timestamp_col)
+            df["timestamp"] = pd.to_datetime(df["timestamp"], errors="coerce")
+            thresholds_df = pd.read_excel(excel_file.name)
+            thresholds_df["Important"] = thresholds_df["Important"].astype(str).str.upper().map({"TRUE":True,"FALSE":False})
+            for col in ["LL","L","H","HH"]:
+                if col in thresholds_df.columns:
+                    thresholds_df[col] = pd.to_numeric(thresholds_df[col], errors="coerce")
+            return df, thresholds_df, "✅ ファイルを読み込みました"
+        except Exception as e:
+            return None, None, f"❌ 読み込み失敗: {e}"
+
+    load_btn = gr.Button("ファイルを読み込み")
+    load_status = gr.Textbox(label="ロード状態")
+
+    load_btn.click(load_files, inputs=[csv_input, excel_input], outputs=[state_df, state_thresholds, load_status])
 
     with gr.Tabs():
-        # Tab1: 閾値診断
         with gr.Tab("閾値診断"):
-            with gr.Row():
-                csv_input = gr.File(label="CSVファイル", file_types=[".csv"], type="filepath")
-                excel_input = gr.File(label="Excel閾値ファイル", file_types=[".xlsx"], type="filepath")
-
             process_name = gr.Textbox(label="プロセス名", value="E018-A012_除害RO")
             datetime_str = gr.Textbox(label="診断基準日時", value="2025/8/1 1:05")
             window_minutes = gr.Number(label="さかのぼる時間幅（分）", value=60)
             run_btn = gr.Button("診断を実行")
-
             result_df_all = gr.Dataframe(label="全項目の状態集計結果")
             result_df_imp = gr.Dataframe(label="重要項目全体の状態集計結果")
             result_df_imp_items = gr.Dataframe(label="重要項目ごとの状態集計結果")
@@ -160,21 +245,28 @@ with gr.Blocks(css="body {overflow-y: auto !important;}") as demo:
 
             run_btn.click(
                 diagnose_process_range,
-                inputs=[csv_input, excel_input, process_name, datetime_str, window_minutes],
+                inputs=[state_df, state_thresholds, process_name, datetime_str, window_minutes],
                 outputs=[result_df_all, result_df_imp, result_df_imp_items, summary_output, json_output]
             )
 
-        # Tab2: 傾向検出
         with gr.Tab("傾向検出"):
-            gr.Markdown("ここはまだダミーです。")
+            process_name2 = gr.Textbox(label="プロセス名", value="E018-A012_除害RO")
+            datetime_str2 = gr.Textbox(label="基準日時", value="2025/8/1 1:05")
+            window_minutes2 = gr.Number(label="さかのぼる時間幅（分）", value=60)
+            forecast_minutes2 = gr.Number(label="未来予測時間幅（分）", value=60)
+            run_btn2 = gr.Button("傾向検出を実行")
+            result_df2 = gr.Dataframe(label="傾向＋未来予測結果")
+            summary_output2 = gr.Textbox(label="サマリー")
+            json_output2 = gr.Json(label="JSON結果")
 
-        # Tab3: 予兆解析
-        with gr.Tab("予兆解析"):
-            gr.Markdown("ここはまだダミーです。")
+            run_btn2.click(
+                detect_trends_with_forecast,
+                inputs=[state_df, state_thresholds, process_name2, datetime_str2, window_minutes2, forecast_minutes2],
+                outputs=[result_df2, summary_output2, json_output2]
+            )
 
-# --- 実行モード選択 ---
 if __name__ == "__main__":
-    use_mcp = os.getenv("USE_MCP", "0") == "1"
+    use_mcp = os.getenv("USE_MCP","0") == "1"
     if use_mcp:
         demo.launch(mcp_server=True)
     else: