Merge branch 'dev' to master (sync)

Author: HadronCollider

Dockerfile

@@ -1,4 +1,4 @@
-FROM node:20-alpine as frontend_build
+FROM node:20-alpine AS frontend_build
 
 WORKDIR /app
 ADD package.json webpack.config.js ./
@@ -7,10 +7,11 @@ RUN npm install && npm install webpack
 ADD ./assets ./assets
 RUN npm run build
 
-FROM dvivanov/dis-base:v0.4
+FROM dvivanov/dis-base:v0.5
 
 LABEL project='dis'
-LABEL version='0.4'
+LABEL version='0.5'
+ENV PYTHONPATH="${PYTHONPATH}:/usr/src/project/app"
 
 WORKDIR /usr/src/project
 
@@ -19,6 +20,4 @@ ADD ./db_versioning ./db_versioning/
 ADD ./app ./app/
 COPY --from=frontend_build /app/src ./src/
 
-ENV PYTHONPATH "${PYTHONPATH}:/usr/src/project/app"
-
-CMD ./scripts/local_start.sh
+CMD ["./scripts/local_start.sh"]

Dockerfile_base

@@ -1,14 +1,13 @@
-FROM python:3.10-slim-bullseye
+FROM python:3.12-slim-bullseye
 
 LABEL project='dis'
-LABEL version='0.4-base'
+LABEL version='0.5-base'
 
-ENV LANG en_US.UTF-8
+ENV LANG=en_US.UTF-8
 ENV TZ=Europe/Moscow
 
 RUN ln -snf /usr/share/zoneinfo/$TZ /etc/localtime && echo $TZ > /etc/timezone
-
-RUN apt update && apt install -y libreoffice-writer libreoffice-impress default-jre
+RUN apt update && apt install -y g++ gcc libreoffice-writer libreoffice-impress default-jre
 
 ADD requirements.txt .
 RUN python3 -m pip install -r requirements.txt --no-cache-dir

app/db/clear_users.py

@@ -4,5 +4,6 @@
 
 logger = logging.getLogger('root_logger')
 
-client.drop_database('dis-db')
-logger.info("Вся информация очищена!")
+
+def drop_database():
+    client.drop_database('pres-parser-db')

app/main/check_packs/base_criterion_pack.py

@@ -1,10 +1,11 @@
 import logging
+import traceback
 from .utils import init_criterions
 
 logger = logging.getLogger('root_logger')
 
 PRIORITY_CHECK_FAILED_MSG = "<b>Данный критерий является обязательным для прохождения.<br>Результат всей проверки обнулен, но вы можете ознакомиться с результатами каждого критерия.</b><br>"
-UNEXPECTED_CHECK_FAIL_MSG = "<b>Во время проверки произошла ошибка, попробуйте позже или обратитесь к администратору системы.<b>"
+UNEXPECTED_CHECK_FAIL_MSG = "<b>Во время проверки произошла ошибка: проверьте соответствия оформления файла шаблона (в том числе разделов и уровней заголовков), попробуйте позже или обратитесь к администратору системы.<b>"
 
 class BaseCriterionPack:
 
@@ -26,8 +27,10 @@ def check(self):
             try:
                 criterion_check_result = criterion.check()
             except Exception as e:
-                err_msg = f'{criterion.id}: oшибка во время проверки: {e}'
+                trace_msg = traceback.format_exc()
+                err_msg = f'{criterion.id}: oшибка во время проверки: {e} ({trace_msg[len(trace_msg)//2:]})'
                 logger.error(err_msg)
+                logger.error(trace_msg)
                 criterion_check_result = {'score': 0, 'verdict': [UNEXPECTED_CHECK_FAIL_MSG, f"Информация об ошибке для администратора: {err_msg}"]}
             if criterion.priority and not criterion_check_result['score']:
                 failed_priority_check = True
@@ -72,5 +75,5 @@ def get_proportion(result):
         score = 0.
         for check in result:
             score += float(check['score'])
-        return score, len(result)
+        return round(score, 2), len(result)
 

app/main/checks/base_check.py

@@ -1,6 +1,6 @@
-import pymorphy2
+import pymorphy3
 
-morph = pymorphy2.MorphAnalyzer()
+morph = pymorphy3.MorphAnalyzer()
 
 
 def answer(mod, *args):

app/main/checks/presentation_checks/find_theme_in_pres.py

@@ -7,7 +7,7 @@
 import nltk
 from nltk.tokenize import word_tokenize, sent_tokenize
 from nltk.corpus import stopwords
-from pymorphy2 import MorphAnalyzer
+from pymorphy3 import MorphAnalyzer
 
 
 MORPH_ANALYZER = MorphAnalyzer()

app/main/checks/report_checks/find_theme_in_report.py

@@ -6,7 +6,7 @@
 import  string
 from nltk.tokenize import word_tokenize, sent_tokenize
 from nltk.corpus import stopwords
-from pymorphy2 import MorphAnalyzer
+from pymorphy3 import MorphAnalyzer
 
 
 MORPH_ANALYZER = MorphAnalyzer()

app/main/checks/report_checks/sw_keywords_check.py

@@ -3,7 +3,7 @@
 
 from nltk.tokenize import word_tokenize
 from nltk.corpus import stopwords
-from pymorphy2 import MorphAnalyzer
+from pymorphy3 import MorphAnalyzer
 from ..base_check import BaseReportCriterion, answer
 
 

app/main/reports/md_uploader/md_uploader.py

@@ -176,7 +176,7 @@ def build_chapter_tree(self, chapters):
             while len(stack) > level:
                 stack.pop()
 
-            parent = stack[-1]
+            parent = stack[-1] if stack else []
             new_chapter = {
                 'name': chapter['name'],
                 'text': chapter['text'],

app/main/reports/pdf_document/pdf_document_manager.py

@@ -1,6 +1,6 @@
 
 # import pdfplumber
-import fitz
+import pymupdf
 
 
 from app.utils import convert_to
@@ -9,10 +9,10 @@ class PdfDocumentManager:
     def __init__(self, path_to_file, pdf_filepath):
         if not pdf_filepath:
             # self.pdf_file = pdfplumber.open(convert_to(path_to_file, target_format='pdf'))
-            self.pdf_file = fitz.open(convert_to(path_to_file, target_format='pdf'))
+            self.pdf_file = pymupdf.open(convert_to(path_to_file, target_format='pdf'))
         else:
             # self.pdf_file = pdfplumber.open(pdf_filepath)
-            self.pdf_file = fitz.open(pdf_filepath)
+            self.pdf_file = pymupdf.open(pdf_filepath)
         self.pages = [self.pdf_file.load_page(page_num) for page_num in range(self.pdf_file.page_count)]
         self.page_count_all = self.pdf_file.page_count
         # self.page_count = len(self.pages)
@@ -34,7 +34,7 @@ def page_images(self, page_without_pril):
         total_height = 0
         for page_num in range(page_without_pril):
             page = self.pdf_file[page_num]
-            images = self.pdf_file.get_page_images(page)
+            images = self.pdf_file.get_page_images(page_num)
             for image in images:
                 image_coord = page.get_image_bbox(image[7], transform=0)    # might be [1.0, 1.0, -1.0, -1.0]
                 image_height = image_coord[3] - image_coord[1]