Update data/validation_schemas.py

data/validation_schemas.py

@@ -1,7 +1,7 @@
-# File: data/validation_schemas.py (NEW FILE)
+# File: data/validation_schemas.py
 # Comprehensive Pydantic validation schemas for data quality assurance
 
-from pydantic import BaseModel, Field, validator, root_validator
+from pydantic import BaseModel, Field, field_validator, model_validator
 from typing import List, Optional, Dict, Any, Union
 from datetime import datetime
 import re
@@ -53,7 +53,8 @@ class TextContentSchema(BaseModel):
         description="The news article text content"
     )
     
-    @validator('text')
+    @field_validator('text')
+    @classmethod
     def validate_text_content(cls, v):
         """Comprehensive text content validation"""
         if not v or not isinstance(v, str):
@@ -155,7 +156,8 @@ class LabelSchema(BaseModel):
         description="Source reliability score (0-1)"
     )
     
-    @validator('label')
+    @field_validator('label')
+    @classmethod
     def validate_label(cls, v):
         """Validate label value"""
         if v not in [0, 1]:
@@ -188,7 +190,8 @@ class DataSourceSchema(BaseModel):
         description="Batch identifier for grouped data"
     )
     
-    @validator('url')
+    @field_validator('url')
+    @classmethod
     def validate_url(cls, v):
         """Validate URL format"""
         if v is not None:
@@ -244,7 +247,8 @@ class NewsArticleSchema(BaseModel):
         description="Overall quality score (0-1)"
     )
     
-    @validator('title')
+    @field_validator('title')
+    @classmethod
     def validate_title(cls, v):
         """Validate article title"""
         if v is not None:
@@ -259,7 +263,8 @@ class NewsArticleSchema(BaseModel):
         
         return v
     
-    @validator('language')
+    @field_validator('language')
+    @classmethod
     def validate_language(cls, v):
         """Validate language code"""
         valid_languages = ['en', 'es', 'fr', 'de', 'it', 'pt', 'nl', 'ru', 'zh', 'ja', 'ko']
@@ -267,11 +272,11 @@ class NewsArticleSchema(BaseModel):
             raise ValueError(f"Unsupported language code: {v}")
         return v
     
-    @root_validator
-    def validate_article_consistency(cls, values):
+    @model_validator(mode='after')
+    def validate_article_consistency(self):
         """Cross-field validation"""
-        text_content = values.get('text_content')
-        title = values.get('title')
+        text_content = self.text_content
+        title = self.title
         
         if text_content and title:
             # Check if title and content are suspiciously similar
@@ -282,7 +287,7 @@ class NewsArticleSchema(BaseModel):
                 # This is fine, just noting high similarity
                 pass
         
-        return values
+        return self
     
     @property
     def text_quality_level(self) -> TextQualityLevel:
@@ -338,8 +343,8 @@ class BatchValidationSchema(BaseModel):
     
     articles: List[NewsArticleSchema] = Field(
         ...,
-        min_items=1,
-        max_items=10000,
+        min_length=1,
+        max_length=10000,
         description="List of articles to validate"
     )
     
@@ -360,7 +365,8 @@ class BatchValidationSchema(BaseModel):
         description="Minimum quality score threshold"
     )
     
-    @validator('articles')
+    @field_validator('articles')
+    @classmethod
     def validate_article_list(cls, v):
         """Validate article list"""
         if not v:
@@ -509,26 +515,27 @@ class BatchValidationResultSchema(BaseModel):
         description="Validation summary statistics"
     )
     
-    @validator('valid_articles', 'invalid_articles')
-    def validate_article_counts(cls, v, values):
+    @field_validator('valid_articles', 'invalid_articles')
+    @classmethod
+    def validate_article_counts(cls, v, info):
         """Validate article count consistency"""
-        if 'total_articles' in values:
-            total = values['total_articles']
+        if 'total_articles' in info.data:
+            total = info.data['total_articles']
             if v > total:
                 raise ValueError("Article count cannot exceed total")
         return v
     
-    @root_validator
-    def validate_counts_consistency(cls, values):
+    @model_validator(mode='after')
+    def validate_counts_consistency(self):
         """Validate count consistency"""
-        total = values.get('total_articles', 0)
-        valid = values.get('valid_articles', 0)
-        invalid = values.get('invalid_articles', 0)
+        total = self.total_articles
+        valid = self.valid_articles
+        invalid = self.invalid_articles
         
         if valid + invalid != total:
             raise ValueError("Valid + invalid articles must equal total articles")
         
-        return values
+        return self
     
     @property
     def success_rate(self) -> float: