added file highlighting

checker/api/views.py

@@ -36,7 +36,6 @@ class RetireDocxSerializer(APIView):
 
 
 class ListCreateWordDocxApiView(generics.ListCreateAPIView):
-    parser_classes = [FormParser, MultiPartParser]
     serializer_class = WordDocxSerializer
     queryset = WordDocx.objects.all()
 

checker/migrations/0001_initial.py

@@ -11,20 +11,4 @@ class Migration(migrations.Migration):
     dependencies = []
 
     operations = [
-        migrations.CreateModel(
-            name="Docx",
-            fields=[
-                (
-                    "uuid",
-                    models.UUIDField(
-                        default=uuid.uuid4,
-                        editable=False,
-                        primary_key=True,
-                        serialize=False,
-                        unique=True,
-                    ),
-                ),
-                ("file", models.FileField(upload_to="")),
-            ],
-        ),
     ]

checker/signals.py

@@ -2,13 +2,14 @@ from django.db.models.signals import post_save
 from django.dispatch import receiver
 
 from checker.models import Docx, WordDocx
-from checker.tasks import process_file, process_word
+from checker.tasks import process_file, process_word, highlight_file
 
 
 @receiver(post_save, sender=Docx)
 def create_docs(sender, instance, created, **kwargs):
     if created:
         process_file.apply_async(kwargs={"pk": instance.pk})
+        highlight_file.apply_async(kwargs={"pk": instance.pk})
         return
 
 

checker/tasks.py

@@ -1,6 +1,8 @@
 import docx2txt
 import requests
 from celery import shared_task
+from docx import Document
+from docx.enum.text import WD_COLOR_INDEX
 
 from checker.models import Paragraph, Docx, WordDocx, WordParagraph
 from checker.services.file import process_paragraphs, process_word_paragraphs
@@ -77,3 +79,37 @@ def process_word(pk: int):
 
     return f"ok, {pk}"
 
+
+@shared_task
+def highlight_file(pk: int):
+    c = 0
+    title = True
+    file = Docx.objects.get(pk=pk)
+    document = Document(file.file.path)
+
+    for paragraph in document.paragraphs:
+        if title:
+            if (
+                paragraph.text
+                and len(paragraph.text) > 2
+                and paragraph.text[:2] == "1."
+            ):
+                title = False
+        else:
+            if paragraph.text:
+                x = requests.post(
+                    "http://109.248.175.223:5000/api", json={1: paragraph.text}
+                )
+                if x.status_code == 200:
+                    el_id, dat = x.json()["1"]
+                    if dat < 50:
+                        text = paragraph.text
+                        paragraph.clear()
+                        run = paragraph.add_run()
+                        run.font.highlight_color = WD_COLOR_INDEX.RED
+                        run.add_text(text)
+                        c += 1
+                else:
+                    print("AI ERROR")
+    document.save(file.file.path)
+    return f"highlighted {c}, {pk}"

requirements/base.txt

@@ -15,4 +15,5 @@ psutil
 dj-database-url
 uuid
 docx2txt
+python-docx
 requests-async