Optimised file processing, added doc, odt -> docx converter, minor changes

app/checker/services/file.py

@@ -1,4 +1,7 @@
 import os
+import re
+import convertapi
+
 
 from checker.services.generators import generate_charset
 
@@ -28,5 +31,31 @@ def process_word_paragraphs(text):
     return _base_process(text)
 
 
+def doc_to_docx(file_path):
+    convertapi.api_secret = '0fp22XFRPwKmNJql'
+    result = convertapi.convert('docx', {'File': file_path}, from_format='doc')
+    result.file.save(file_path.split(".")[0] + ".docx")
+    return file_path.split(".")[0] + ".docx"
+
+
+def doc_to_odt(file_path):
+    convertapi.api_secret = '0fp22XFRPwKmNJql'
+    result = convertapi.convert('docx', {'File': file_path}, from_format='odt')
+    result.file.save(file_path.split(".")[0] + ".docx")
+    return file_path.split(".")[0] + ".docx"
+
+
 def media_upload_path(instance, filename):
     return os.path.join(f"uploads/{generate_charset(7)}/", filename)
+
+
+def split_text(text):
+    texts, groups = [], []
+    regt = re.findall(r"{(.*?)}(.*?){(.*?)}", text.replace('\n', ''))
+    for t in regt:
+        if t[0] == t[-1]:
+            texts.append(t[1])
+            groups.append(int(t[0]))
+        else:
+            print(t)
+    return texts, groups

app/checker/signals.py

@@ -1,15 +1,31 @@
+import magic
+
 from django.db.models.signals import post_save
+from django.core.files import File
 from django.dispatch import receiver
+from celery import chain
 
 from checker.models import Docx, WordDocx
+from checker.services.file import doc_to_docx, doc_to_odt
 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})
+        type = magic.from_file(instance.file.path, mime=True)
+        if type == "application/msword":
+            pth = doc_to_docx(instance.file.path)
+            with open(pth, 'rb') as f:
+                instance.file = File(f, name=pth.split("/")[-1])
+                instance.save(update_fields=["file"])
+        elif type == "application/vnd.oasis.opendocument.text":
+            pth = doc_to_odt(instance.file.path)
+            with open(pth, 'rb') as f:
+                instance.file = File(f, name=pth.split("/")[-1])
+                instance.save(update_fields=["file"])
+
+        chain(process_file.s(instance.pk), highlight_file.s()).apply_async()
         return
 
 

app/checker/tasks.py

@@ -3,49 +3,63 @@ import requests
 from celery import shared_task
 from docx import Document
 from docx.enum.text import WD_COLOR_INDEX
+from requests.exceptions import InvalidJSONError
 
 from checker.models import Paragraph, Docx, WordDocx, WordParagraph
-from checker.services.file import process_paragraphs, process_word_paragraphs
+from checker.services.file import process_paragraphs, process_word_paragraphs, split_text
 
 
 @shared_task()
-def process_file(pk: int):
+def process_file(pk: int, *args, **kwargs):
     file = Docx.objects.get(pk=pk)
     uuid = file.uuid
     document = docx2txt.process(file.file.path)
-    paragraphs = process_paragraphs(document)
+    # paragraphs = process_paragraphs(document)
+    paragraphs, groups = split_text(document)
 
     file.paragraphs_loaded = len(paragraphs)
     file.save(update_fields=["paragraphs_loaded"])
 
-    cut = 100
-    counter = 0
-    len_c = len(paragraphs) + 1
-    paragraphs = list(paragraphs.values())
-    for i in range(0, len(paragraphs) // cut + 1):
-        vals = paragraphs[i * cut : (i + 1) * cut + 1]
-        dct = {x: vals[x] for x in range(len(vals))}
-
-        x = requests.post("http://109.248.175.223:5000/api", json=dct)
-        if x.status_code == 200:
+    cut = 10
+    for i in range(len(paragraphs) // cut):
+        vals = [x for x in range(i * cut, (i+ 1) * cut)]
+        dct = {x: paragraphs[x] for x in vals}
+        x = requests.post("http://109.248.175.223:5000/api", json=dct, timeout=1)
         for el_id, dat in x.json().items():
             type_id, score = dat
             Paragraph.objects.create(
-                    type_id=type_id, docx=file, text=dct[int(el_id)], score=score
+                type_id=type_id, docx=file, text=str(groups[int(el_id)]) + dct[int(el_id)] + str(groups[int(el_id)]), score=score
             )
 
-            counter += len(vals)
-            print(f"processing {uuid}, {counter}/{len_c}")
-            file.paragraphs_processed = counter
-            file.save(update_fields=["paragraphs_processed"])
-        else:
-            print(f"AI server error, {x.status_code}")
+    #for i in range(0, len(paragraphs) // cut + 1):
+    #    vals = paragraphs[i * cut : (i + 1) * cut + 1]
+    #    dct = {x: vals[x] for x in range(len(vals))}
+    #
+    #    x = requests.post("http://109.248.175.223:5000/api", json=dct)
+    #    if x.status_code == 200:
+    #        try:
+    #            for el_id, dat in x.json().items():
+    #                type_id, score = dat
+    #                Paragraph.objects.create(
+    #                    type_id=type_id, docx=file, text=str(groups[g_c])+dct[int(el_id)]+str(groups[g_c]), score=score
+    #                )
+    #                g_c += 1
+    #
+    #            counter += len(vals)
+    #            print(f"processing {uuid}, {counter}/{len_c}")
+    #            file.paragraphs_processed = counter
+    #            file.save(update_fields=["paragraphs_processed"])
+    #        except InvalidJSONError:
+    #            print("json pars error")
+    #    else:
+    #        print(f"AI server error, {x.status_code}")
 
-    return f"ok, {pk}"
+
+    return pk
 
 
 @shared_task()
-def process_word(pk: int):
+def process_word(pk: int, *args, **kwargs):
     file = WordDocx.objects.get(pk=pk)
     uuid = file.uuid
     paragraphs = process_word_paragraphs(file.text.tobytes().decode())
@@ -53,7 +67,7 @@ def process_word(pk: int):
     file.paragraphs_loaded = len(paragraphs)
     file.save(update_fields=["paragraphs_loaded"])
 
-    cut = 150
+    cut = 10
     len_c = len(paragraphs) + 1
     paragraphs = list(paragraphs.values())
     counter = 0
@@ -63,6 +77,7 @@ def process_word(pk: int):
 
         x = requests.post("http://109.248.175.223:5000/api", json=dct)
         if x.status_code == 200:
+            try:
                 for el_id, dat in x.json().items():
                     type_id, score = dat
                     WordParagraph.objects.create(
@@ -73,21 +88,23 @@ def process_word(pk: int):
                 print(f"processing {uuid}, {counter}/{len_c}")
                 file.paragraphs_processed = counter
                 file.save(update_fields=["paragraphs_processed"])
+            except InvalidJSONError:
+                print("json pars error")
         else:
             print(f"AI server error, {x.status_code}")
 
-    return f"ok, {pk}"
+    return pk
 
 
 @shared_task
-def highlight_file(pk: int):
+def highlight_file(pk: int, *args, **kwargs):
     c = 0
     lim = 0
     file = Docx.objects.get(pk=pk)
     document = Document(file.file.path)
 
     paragraphs = document.paragraphs
-    cut = 100
+    cut = 10
 
     for paragraph in paragraphs:
         if paragraph.text and len(paragraph.text) > 2 and paragraph.text[:2] == "1.":
@@ -101,6 +118,7 @@ def highlight_file(pk: int):
             if dat:
                 n_dct[el] = dat
         x = requests.post("http://109.248.175.223:5000/api", json=n_dct)
+        try:
             jsn = x.json()
             if x.status_code == 200:
                 for j in range(len(paragraphs_sliced)):
@@ -116,5 +134,7 @@ def highlight_file(pk: int):
                             c += 1
             else:
                 print("AI server error")
+        except InvalidJSONError:
+            print("json pars error")
     document.save(file.file.path)
-    return f"highlighted {c}, {pk}"
+    return pk