add curation.models.TranslatedContent

Author: darjeeling

pythonkr_backend/curation/migrations/0013_translatedcontent.py

@@ -0,0 +1,33 @@
+# Generated by Django 5.2.1 on 2025-06-14 05:26
+
+import curation.models
+import django.db.models.deletion
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('curation', '0012_llmusage_total_tokens'),
+    ]
+
+    operations = [
+        migrations.CreateModel(
+            name='TranslatedContent',
+            fields=[
+                ('id', models.BigAutoField(auto_created=True, primary_key=True, serialize=False, verbose_name='ID')),
+                ('title', models.CharField(help_text='제목', max_length=512)),
+                ('slug', models.SlugField(help_text='slug', max_length=512)),
+                ('description', models.TextField(help_text='설명')),
+                ('tags', models.JSONField(default=list, help_text='태그')),
+                ('written_date', models.DateField(blank=True, help_text='작성 일자', null=True)),
+                ('author', models.CharField(blank=True, help_text='작성자', max_length=100, null=True)),
+                ('content', models.FileField(blank=True, help_text='번역된 마크다운 콘텐츠 파일', null=True, upload_to=curation.models.translated_item_upload_path)),
+                ('model_name', models.CharField(help_text='사용된 모델명', max_length=100)),
+                ('source_url', models.URLField(help_text='원본 URL')),
+                ('created_at', models.DateTimeField(auto_now_add=True)),
+                ('updated_at', models.DateTimeField(auto_now=True)),
+                ('source_rss_item', models.ForeignKey(help_text='원본 RSS 아이템', null=True, on_delete=django.db.models.deletion.CASCADE, related_name='translated_contents', to='curation.rssitem')),
+            ],
+        ),
+    ]

pythonkr_backend/curation/models.py

@@ -10,7 +10,6 @@
 
 def rss_item_upload_path(instance, filename):
     """Generate upload path for RSS item crawled content"""
-    from datetime import datetime
     now = datetime.now()
     return f"rssitem-crawling/{now.year}/{now.month:02d}/{instance.id}-crawl.md"
 
@@ -604,3 +603,62 @@ class Meta:
         verbose_name = "LLM Usage"
         verbose_name_plural = "LLM Usage"
         ordering = ['-date', '-created_at']
+
+def translated_item_upload_path(instance, filename):
+    """Generate upload path for RSS item translated content"""
+    now = datetime.now()
+    return f"tr/{now.year}/{now.month:02d}/{instance.id}-ko.md"
+
+
+class TranslatedContent(models.Model):
+    title = models.CharField(
+        max_length=512,
+        help_text="제목"
+    )
+    slug = models.SlugField(
+        max_length=512,
+        help_text="slug"
+    )
+    description = models.TextField(
+        help_text="설명"
+    )
+    tags = models.JSONField(
+        default=list,
+        help_text="태그"
+    )
+    written_date = models.DateField(
+        help_text="작성 일자",
+        blank=True,
+        null=True,
+    )
+    author = models.CharField(
+        max_length=100,
+        blank=True,
+        null=True,
+        help_text="작성자"
+    )
+    content = models.FileField(
+        upload_to=translated_item_upload_path,
+        blank=True,
+        null=True,
+        help_text="번역된 마크다운 콘텐츠 파일"
+    )
+
+    model_name = models.CharField(
+        max_length=100,
+        help_text="사용된 모델명"
+    )
+
+    source_rss_item = models.ForeignKey(
+        RSSItem,
+        on_delete=models.CASCADE,
+        related_name="translated_contents",
+        null=True,
+        help_text="원본 RSS 아이템"
+    )
+    source_url = models.URLField(
+        help_text="원본 URL"
+    )
+
+    created_at = models.DateTimeField(auto_now_add=True)
+    updated_at = models.DateTimeField(auto_now=True)

pythonkr_backend/curation/utils.py

@@ -1,6 +1,8 @@
 import httpx
 import llm
-from pydantic import BaseModel
+from pydantic import BaseModel, Field
+from pydantic_ai import Agent
+from datetime import date, datetime, time, timedelta
 
 import os
 
@@ -75,3 +77,18 @@ def categorize_summary(summary: str, categories: list[str]):
 
     return response.text()
 
+
+class TranslatedResult(BaseModel):
+    title: str = Field(description="The title of the translated article")
+    slug: str = Field(
+        description="The URL slug. Do not include the language code. Make it similar to the original URL."
+    )
+    description: str = Field(
+        description="The description of the translated article. Don't mention that it's translated."
+    )
+    author: str = Field(description="The author of the translated article")
+    tags: list[str] = Field(
+        description="List of Python-related tags inferred from the document."
+    )
+    written_date: date = Field(description="The written date of the translated article")
+    content: str = Field(description="The content of the translated article")