Update app.py

app.py

@@ -0,0 +1,371 @@
+import gradio as gr
+import pandas as pd
+import pdfplumber
+import os
+import tempfile
+import re
+from datetime import datetime
+from reportlab.pdfgen import canvas
+from reportlab.lib.pagesizes import A4
+from reportlab.pdfbase import pdfmetrics
+from reportlab.pdfbase.ttfonts import TTFont
+from reportlab.lib import colors
+
+
+class KDChecker:
+    def __init__(self):
+        self.excel_db = pd.DataFrame()
+        self.cabinet_list = []
+        self.known_docs = ["Э3", "В4", "ПЭ3", "ВО", "ТЭ5", "СБ", "С5", "ОЛ", "Э1", "Э4", "Э7", "Д3", "Э6"]
+
+    def load_excel_db(self, excel_file):
+        if excel_file is None:
+            return "Файл не выбран", gr.update(choices=[], value=None)
+
+        all_data = []
+        sheets_log = []
+
+        try:
+            xls = pd.read_excel(excel_file.name, sheet_name=None, header=None)
+
+            for sheet_name, df_raw in xls.items():
+                header_row_index = -1
+                cab_col_idx = -1
+                rem_col_idx = -1
+
+                for i in range(min(20, len(df_raw))):
+                    row_values = [str(x).lower().strip() for x in df_raw.iloc[i].values]
+                    c_idx = -1
+                    r_idx = -1
+                    for idx, val in enumerate(row_values):
+                        if "шкаф" in val or "cabinet" in val: c_idx = idx
+                        if "примечание" in val or "remark" in val: r_idx = idx
+                    if c_idx != -1 and r_idx != -1:
+                        header_row_index = i
+                        cab_col_idx = c_idx
+                        rem_col_idx = r_idx
+                        break
+
+                if header_row_index != -1:
+                    df = pd.read_excel(excel_file.name, sheet_name=sheet_name, header=header_row_index)
+                    df_subset = df.iloc[:, [cab_col_idx, rem_col_idx]]
+                    df_subset.columns = ["Cabinet", "Remark"]
+
+                    df_subset["Cabinet"] = df_subset["Cabinet"].ffill()
+                    df_subset = df_subset.dropna(subset=["Remark"]).astype(str)
+
+                    df_subset["Cabinet_Clean"] = df_subset["Cabinet"].apply(
+                        lambda x: x.strip().replace(" ", "").replace("\n", "").replace("\r", "")
+                    )
+
+                    all_data.append(df_subset)
+                    sheets_log.append(f"Лист '{sheet_name}': {len(df_subset)} стр.")
+                else:
+                    sheets_log.append(f"Лист '{sheet_name}': заголовки не найдены")
+
+            if not all_data:
+                return "❌ Ошибка: Не найдены заголовки 'Шкаф' и 'Примечание'.", gr.update(choices=[], value=None)
+
+            self.excel_db = pd.concat(all_data, ignore_index=True)
+            self.cabinet_list = sorted(self.excel_db["Cabinet"].unique().tolist())
+
+            msg = f"✅ База знаний загружена успешно!\nВсего записей: {len(self.excel_db)}\nОбработаны листы: {', '.join(sheets_log)}"
+            return msg, gr.update(choices=self.cabinet_list, value=None, interactive=True)
+
+        except Exception as e:
+            return f"❌ Ошибка чтения Excel: {e}", gr.update(choices=[], value=None)
+
+    def extract_text(self, pdf_path):
+        try:
+            full_text = ""
+            with pdfplumber.open(pdf_path) as pdf:
+                for page in pdf.pages:
+                    full_text += (page.extract_text() or "") + "\n"
+            return full_text
+        except:
+            return ""
+
+    def find_all_decimal_numbers(self, text):
+        pattern = r"(РЛТ|ЛДАР|ВНАР|ШТМ)[\s\.]*\d{1}[\s\.]*\d{3}[\s\.]*[А-ЯA-Z]{1,4}[\s\.]*\d{3}(-[\d]+)?"
+        matches = []
+        for match in re.finditer(pattern, text):
+            clean_num = match.group(0).replace(" ", "").replace("\n", "")
+            if clean_num not in matches:
+                matches.append(clean_num)
+        return matches
+
+    def determine_doc_type(self, filename):
+        fname = filename.upper()
+        if "С2" in fname: return "С2"
+        if "ПЭ3" in fname or "ПЕРЕЧЕНЬ" in fname: return "ПЭ3"
+        if "Э3" in fname or "СХЕМА ЭЛЕКТРИЧЕСКАЯ" in fname: return "Э3"
+        if "Э4" in fname: return "Э4"
+        if "В4" in fname or "СПЕЦИФИКАЦИЯ" in fname: return "В4"
+        if "ВО" in fname or "Э7" in fname or "ГАБАРИТ" in fname: return "ВО"
+        if "ТЭ5" in fname or "ТАБЛИЦА" in fname: return "ТЭ5"
+        if "СБ" in fname: return "СБ"
+        if "С5" in fname: return "С5"
+        if "ОЛ" in fname: return "ОЛ"
+        if "Э1" in fname: return "Э1"
+        if "Э6" in fname or "ЗАЗЕМЛЕНИЯ" in fname: return "Э6"
+        if "Д3" in fname or "МОНТАЖ" in fname: return "Д3"
+        return "UNKNOWN"
+
+    def get_remarks(self, cabinet_key, is_clean_key=True):
+        if self.excel_db.empty: return {}
+
+        if is_clean_key:
+            target = cabinet_key.replace(" ", "")
+            mask = self.excel_db['Cabinet_Clean'].str.contains(re.escape(target), case=False, na=False)
+        else:
+            mask = self.excel_db['Cabinet'] == cabinet_key
+
+        rows = self.excel_db[mask]
+        if rows.empty: return {}
+
+        parsed = {}
+
+        for remark_cell in rows['Remark']:
+            cell_text = str(remark_cell)
+            cell_text = re.sub(r'(\d+)\.([А-ЯA-Z])', r'\1. \2', cell_text)
+
+            items = re.split(r'(?:^|\n)\s*(?=\d+[\.\)])', cell_text)
+
+            for item in items:
+                if len(item) < 3: continue
+                clean_item = item.strip()
+                clean_item_no_num = re.sub(r'^\d+[\.\)]\s*', '', clean_item)
+
+                doc_pattern = r'^(?:Документ\s+|В\s+)?([А-ЯA-Z0-9\s,\(\)\-]+?)(?:[\.\:\-]|\s+)(.*)'
+                match = re.match(doc_pattern, clean_item_no_num, re.IGNORECASE | re.DOTALL)
+
+                detected_docs = []
+                final_text = clean_item
+
+                if match:
+                    potential_docs_str = match.group(1).upper()
+                    cleaned_codes = potential_docs_str.replace("(", " ").replace(")", " ").replace(",", " ")
+                    parts = cleaned_codes.split()
+
+                    valid_parts = [p for p in parts if p in self.known_docs]
+                    if valid_parts:
+                        detected_docs = valid_parts
+                        final_text = match.group(2).strip()
+
+                if not detected_docs:
+                    detected_docs = ["ALL"]
+
+                for doc in detected_docs:
+                    if doc not in parsed: parsed[doc] = []
+                    parsed[doc].append(final_text)
+
+        return parsed
+
+    def check_files(self, files, manual_cabinet):
+        if not files: return "Файлы не загружены", None
+        if self.excel_db.empty: return "Сначала загрузите Excel базу!", None
+
+        checklist = {}
+        detected_cabinet = "Не определен"
+        found_by_method = ""
+        is_manual = False
+
+        # 1. Ручной выбор
+        if manual_cabinet and manual_cabinet.strip():
+            detected_cabinet = manual_cabinet
+            found_by_method = "manual"
+            is_manual = True
+
+        # 2. Автопоиск
+        else:
+            all_pdf_text = ""
+            for file in files:
+                all_pdf_text += self.extract_text(file.name) + "\n"
+
+            # А. По номеру
+            pdf_numbers = self.find_all_decimal_numbers(all_pdf_text)
+            db_clean_keys = set(self.excel_db["Cabinet_Clean"].tolist())
+
+            for cand in pdf_numbers:
+                if cand in db_clean_keys:
+                    detected_cabinet = cand
+                    found_by_method = "number"
+                    break
+
+            # Б. По имени
+            if detected_cabinet == "Не определен":
+                unique_cabinets = self.excel_db["Cabinet"].unique()
+                for cab_name in unique_cabinets:
+                    sub_names = [n.strip() for n in cab_name.split(',')]
+                    for sub_name in sub_names:
+                        if len(sub_name) < 5: continue
+                        if sub_name.lower() in all_pdf_text.lower():
+                            detected_cabinet = cab_name
+                            found_by_method = "name"
+                            break
+                    if found_by_method == "name": break
+
+        if detected_cabinet == "Не определен":
+            examples = ", ".join(self.excel_db["Cabinet"].head(3).tolist())
+            return f"⚠️ Шкаф не опознан автоматически.\nСовет: Выберите похожий шкаф из выпадающего списка вручную.", None
+
+        is_clean_search = (found_by_method == "number")
+        remarks = self.get_remarks(detected_cabinet, is_clean_key=is_clean_search)
+
+        if not remarks:
+            return f"⚠️ Шкаф '{detected_cabinet}' выбран, но в базе нет замечаний для него.", None
+
+        processed_count = 0
+        for file in files:
+            fname = os.path.basename(file.name)
+            dtype = self.determine_doc_type(fname)
+
+            tasks = []
+            if dtype in remarks: tasks.extend(remarks[dtype])
+            if "ALL" in remarks and dtype != "С2": tasks.extend(remarks["ALL"])
+
+            if tasks:
+                checklist[fname] = list(dict.fromkeys(tasks))
+            processed_count += 1
+
+        pdf_title = detected_cabinet
+        if is_manual: pdf_title += " (Выбор вручную)"
+
+        pdf = self.create_pdf(pdf_title, checklist)
+        total = sum(len(v) for v in checklist.values())
+
+        method_str = "Ручной выбор" if is_manual else (
+            "По децимальному н��меру" if is_clean_search else "По наименованию")
+
+        return f"✅ Готово!\n\n📂 Шкаф: {detected_cabinet}\n🔍 Метод: {method_str}\n📄 Обработано файлов: {processed_count}\n🚩 Всего замечаний: {total}", pdf
+
+    def create_pdf(self, cabinet, data):
+        fname = f"CheckList_Result.pdf"
+        path = os.path.join(tempfile.gettempdir(), fname)
+        c = canvas.Canvas(path, pagesize=A4)
+        form = c.acroForm
+        width, height = A4
+
+        # Подключение шрифтов (попытка найти Arial, иначе стандартный)
+        font_name = 'Helvetica'
+        font_path = "arial.ttf"  # Ожидаем файл рядом со скриптом на сервере
+
+        # Проверяем системные пути Windows и локальную папку
+        possible_paths = ["C:\\Windows\\Fonts\\arial.ttf", "arial.ttf",
+                          "/usr/share/fonts/truetype/msttcorefonts/Arial.ttf"]
+
+        for p in possible_paths:
+            if os.path.exists(p):
+                try:
+                    pdfmetrics.registerFont(TTFont('Arial', p))
+                    font_name = 'Arial'
+                    break
+                except:
+                    continue
+
+        y = height - 50
+        c.setFont(font_name, 16)
+        c.drawString(50, y, f"ЧЕК-ЛИСТ ПРОВЕРКИ КД")
+        y -= 25
+        c.setFont(font_name, 12)
+        disp_cab = cabinet[:60] + "..." if len(cabinet) > 60 else cabinet
+        c.drawString(50, y, f"Шкаф: {disp_cab}")
+        c.drawString(400, y, f"Дата: {datetime.now().strftime('%d.%m.%Y')}")
+        y -= 20
+        c.line(50, y, width - 50, y)
+        y -= 30
+
+        if not data:
+            c.drawString(50, y, "Нет замечаний для загруженных файлов.")
+            c.save()
+            return path
+
+        cb_id = 0
+        for filename, tasks in data.items():
+            if y < 100: c.showPage(); y = height - 50; c.setFont(font_name, 12)
+
+            c.setFillColor(colors.darkblue)
+            c.setFont(font_name, 11)
+            c.drawString(50, y, f"Файл: {filename}")
+            c.setFillColor(colors.black)
+            y -= 15
+            c.setFont(font_name, 10)
+
+            for task in tasks:
+                paragraphs = task.split('\n')
+
+                if y < 80:
+                    c.showPage();
+                    y = height - 50;
+                    c.setFont(font_name, 10)
+
+                c.rect(50, y - 10, 10, 10)
+                form.checkbox(name=f"cb_{cb_id}", x=50, y=y - 10, size=10, buttonStyle='check', forceBorder=True)
+                cb_id += 1
+
+                text_start_y = y - 2
+
+                for paragraph in paragraphs:
+                    max_len = 95
+                    lines = []
+                    words = paragraph.split(' ')
+                    cur_line = ""
+                    for w in words:
+                        if len(cur_line) + len(w) + 1 <= max_len:
+                            cur_line += w + " "
+                        else:
+                            lines.append(cur_line); cur_line = w + " "
+                    lines.append(cur_line)
+
+                    for l in lines:
+                        if text_start_y < 40:
+                            c.showPage();
+                            text_start_y = height - 50;
+                            c.setFont(font_name, 10)
+                        c.drawString(65, text_start_y, l.strip())
+                        text_start_y -= 12
+
+                y = text_start_y - 8
+
+            y -= 10
+            c.setStrokeColor(colors.lightgrey)
+            c.line(50, y, width - 50, y)
+            c.setStrokeColor(colors.black)
+            y -= 20
+
+        c.save()
+        return path
+
+
+def create_app():
+    checker = KDChecker()
+    with gr.Blocks(title="Генератор чек-листов КД") as app:
+        gr.Markdown("# ✅ Генератор чек-листов КД")
+        gr.Markdown("Автоматическая проверка конструкторской документации по базе знаний Excel.")
+
+        with gr.Row():
+            with gr.Column():
+                gr.Markdown("### 1. База знаний")
+                db_in = gr.File(label="Загрузить Excel (.xlsx)")
+                manual_cab = gr.Dropdown(label="Или выберите шкаф-аналог вручную", choices=[], interactive=True)
+                db_out = gr.Textbox(label="Статус загрузки", lines=8, max_lines=30)
+
+                db_in.upload(checker.load_excel_db, inputs=[db_in], outputs=[db_out, manual_cab])
+
+            with gr.Column():
+                gr.Markdown("### 2. Документация (PDF)")
+                files_in = gr.File(label="Загрузить чертежи", file_count="multiple")
+                btn = gr.Button("Сформировать чек-лист", variant="primary")
+
+        with gr.Row():
+            res_txt = gr.Textbox(label="Результат проверки", lines=8, max_lines=30)
+            res_pdf = gr.File(label="Скачать PDF чек-лист")
+
+        btn.click(checker.check_files, inputs=[files_in, manual_cab], outputs=[res_txt, res_pdf])
+    return app
+
+
+app = create_app()
+
+if __name__ == "__main__":
+    app.launch()