Merge pull request #387 from AI4Bharat/translation_bugs

Bug fix for monolingual docx
AI4Bharat · Jul 18, 2023 · 29480d3 · 29480d3
2 parents 1129bf5 + b9dbadb
commit 29480d3
Show file tree

Hide file tree

Showing 2 changed files with 25 additions and 5 deletions.
diff --git a/backend/translation/utils.py b/backend/translation/utils.py
@@ -84,6 +84,28 @@ def convert_to_paragraph(lines):
     return content
 
 
+def convert_to_paragraph_monolingual(payload):
+    lines = []
+    content = ""
+    translated_content = ""
+    sentences_count = 0
+    number_of_paragraphs = math.ceil(len(payload) / 5)
+    count_paragraphs = 0
+    for index, segment in enumerate(payload):
+        if "text" in segment.keys():
+            lines.append(segment["target_text"])
+            translated_content = translated_content + " " + segment["target_text"]
+            sentences_count += 1
+            if sentences_count % 5 == 0:
+                count_paragraphs += 1
+                content = content + translated_content + "\n" + "\n"
+                translated_content = ""
+
+    if count_paragraphs < number_of_paragraphs:
+        content = content + translated_content + "\n" + "\n"
+    return content
+
+
 def convert_to_paragraph_bilingual(payload):
     lines = []
     transcripted_lines = []
@@ -100,7 +122,7 @@ def convert_to_paragraph_bilingual(payload):
             transcripted_content = (
                 transcripted_content + " " + segment["text"].replace("\n", " ")
             )
-            translated_content = translated_content + segment["target_text"]
+            translated_content = translated_content + " " + segment["target_text"]
             sentences_count += 1
             if sentences_count % 5 == 0:
                 count_paragraphs += 1

diff --git a/backend/translation/views.py b/backend/translation/views.py
@@ -42,6 +42,7 @@
     get_batch_translations_using_indictrans_nmt_api,
     convert_to_docx,
     convert_to_paragraph,
+    convert_to_paragraph_monolingual,
     convert_to_paragraph_bilingual,
     generate_translation_payload,
 )
@@ -188,11 +189,8 @@ def export_translation(request):
         filename = "translation.txt"
         content = convert_to_paragraph(lines)
     elif export_type == "docx":
-        for index, segment in enumerate(payload):
-            if "text" in segment.keys():
-                lines.append(segment["target_text"])
         filename = "translation.docx"
-        content = convert_to_paragraph(lines)
+        content = convert_to_paragraph_monolingual(payload)
         return convert_to_docx(content)
     elif export_type == "docx-bilingual":
         filename = "translation.docx"