Release 202506101742

pyworker/pyworker/database.py

@@ -332,7 +332,7 @@ def remove_service_attribute_by_slug(service_id: int, attribute_slug: str) -> bo
     return remove_service_attribute(service_id, attribute_id)
 
 
-def save_tos_review(service_id: int, review: TosReviewType):
+def save_tos_review(service_id: int, review: Optional[TosReviewType]):
     """
     Save a TOS review for a specific service.
 
@@ -341,8 +341,8 @@ def save_tos_review(service_id: int, review: TosReviewType):
         review: A TypedDict containing the review data.
     """
     try:
-        # Serialize the dictionary to a JSON string for the database
-        review_json = json.dumps(review)
+        # Only serialize to JSON if review is not None
+        review_json = json.dumps(review) if review is not None else None
         with get_db_connection() as conn:
             with conn.cursor(row_factory=dict_row) as cursor:
                 cursor.execute(

pyworker/pyworker/tasks/tos_review.py

@@ -3,7 +3,7 @@ Task for retrieving Terms of Service (TOS) text.
 """
 
 import hashlib
-from typing import Any, Dict, Optional
+from typing import Any, Dict, Optional, Literal
 
 from pyworker.database import TosReviewType, save_tos_review, update_kyc_level
 from pyworker.tasks.base import Task
@@ -52,65 +52,71 @@ class TosReviewTask(Task):
         )
         self.logger.info(f"TOS URLs: {tos_urls}")
 
+        review = self.get_tos_review(tos_urls, service.get("tosReview"))
+        save_tos_review(service_id, review)
+
+        # Update the KYC level based on the review
+        if "kycLevel" in review:
+            kyc_level = review["kycLevel"]
+            self.logger.info(f"Updating KYC level to {kyc_level} for service {service_name}")
+            update_kyc_level(service_id, kyc_level)
+
+        return review
+
+    def get_tos_review(
+        self, tos_urls: list[str], current_review: Optional[TosReviewType]
+    ) -> Optional[TosReviewType]:
+        """
+        Get TOS review from a list of URLs.
+
+        Args:
+            tos_urls: List of TOS URLs to check
+            current_review: Current review data from the database
+
+        Returns:
+            Dict containing:
+            - status: Literal["skipped", "failed", "success"]
+            - review: Optional[TosReviewType] - The review data if successful
+        """
+        all_skipped = True
+
         for tos_url in tos_urls:
             api_url = f"{tos_url}"
             self.logger.info(f"Fetching TOS from URL: {api_url}")
 
-            # Sleep for 1 second to avoid rate limiting
             content = fetch_markdown(api_url)
 
-            if content:
-                # Hash the content to avoid repeating the same content
-                content_hash = hashlib.sha256(content.encode()).hexdigest()
-                self.logger.info(f"Content hash: {content_hash}")
+            if not content:
+                self.logger.warning(f"Failed to retrieve TOS content for URL: {tos_url}")
+                all_skipped = False
+                continue
 
-                # service.get("tosReview") can be None if the DB field is NULL.
-                # Default to an empty dict to prevent AttributeError on .get()
-                tos_review_data_from_service: Optional[Dict[str, Any]] = service.get(
-                    "tosReview"
-                )
-                tos_review: Dict[str, Any] = (
-                    tos_review_data_from_service
-                    if tos_review_data_from_service is not None
-                    else {}
+            # Hash the content to avoid repeating the same content
+            content_hash = hashlib.sha256(content.encode()).hexdigest()
+            self.logger.info(f"Content hash: {content_hash}")
+
+            # Skip processing if we've seen this content before
+            if current_review and current_review.get("contentHash") == content_hash:
+                self.logger.info(
+                    f"Skipping already processed TOS content with hash: {content_hash}"
                 )
+                continue
 
-                stored_hash = tos_review.get("contentHash")
+            all_skipped = False
 
-                # Skip processing if we've seen this content before
-                if stored_hash == content_hash:
-                    self.logger.info(
-                        f"Skipping already processed TOS content with hash: {content_hash}"
-                    )
-                    continue
+            # Skip incomplete TOS content
+            check = prompt_check_tos_review(content)
+            if not check or not check["isComplete"]:
+                continue
 
-                # Skip incomplete TOS content
-                check = prompt_check_tos_review(content)
-                if not check:
-                    continue
-                elif not check["isComplete"]:
-                    continue
-
-                # Query OpenAI to summarize the content
-                review = prompt_tos_review(content)
-
-                if review:
-                    review["contentHash"] = content_hash
-                    # Save the review to the database
-                    save_tos_review(service_id, review)
-
-                    # Update the KYC level based on the review
-                    if "kycLevel" in review:
-                        kyc_level = review["kycLevel"]
-                        self.logger.info(
-                            f"Updating KYC level to {kyc_level} for service {service_name}"
-                        )
-                        update_kyc_level(service_id, kyc_level)
-                    # no need to check other TOS URLs
-                    break
+            # Query OpenAI to summarize the content
+            review = prompt_tos_review(content)
 
+            if review:
+                review["contentHash"] = content_hash
                 return review
-            else:
-                self.logger.warning(
-                    f"Failed to retrieve TOS content for URL: {tos_url}"
-                )
+
+        if all_skipped:
+            return current_review
+
+        return None

pyworker/pyworker/utils/ai.py

@@ -178,7 +178,7 @@ type TosReview = {
 
 The rating is a number between 0 and 2, where 0 is informative, 1 is warning, and 2 is critical.
 
-Be concise but thorough, and make sure your output is properly formatted JSON.
+Do not provide more than 8 highlights. Focus on the most important information for the user. Be concise but thorough, and make sure your output is properly formatted JSON.
 """
 
 PROMPT_COMMENT_SENTIMENT_SUMMARY = """