Making restaurant data changes

Author: pamelafox

convert_csv_json.py

@@ -0,0 +1,51 @@
+import ast
+import csv
+import json
+
+# Read CSV file - Using the correct dialect to handle quotes properly
+with open("data.csv", encoding="utf-8") as csv_file:
+    # Use the csv.reader with proper quoting parameters
+    csv_reader = csv.reader(csv_file, quoting=csv.QUOTE_ALL, doublequote=True, escapechar="\\")
+    header = next(csv_reader)  # Get the header row
+    data = list(csv_reader)  # Get all data rows
+
+# Convert to JSON format
+json_data = []
+for row in data:
+    item = {}
+    for i in range(len(header)):
+        if i < len(row):  # Ensure we don't go out of bounds
+            value = row[i].strip()
+            # Check if the value looks like a JSON array
+            if value.startswith("[") and value.endswith("]"):
+                try:
+                    # Parse the JSON-like string into a Python object
+                    value = json.loads(value.replace("'", '"'))
+                except (ValueError, SyntaxError):
+                    try:
+                        # Try with ast as a fallback
+                        value = ast.literal_eval(value)
+                    except (ValueError, SyntaxError):
+                        # If parsing fails, keep it as a string
+                        pass
+            # Convert boolean strings
+            elif value.lower() == "true":
+                value = True
+            elif value.lower() == "false":
+                value = False
+            # Try to convert numbers
+            elif value.isdigit():
+                value = int(value)
+            elif value.replace(".", "", 1).isdigit() and value.count(".") <= 1:
+                value = float(value)
+
+            item[header[i]] = value
+    # remove is_open column
+    del item["is_open"]
+    json_data.append(item)
+
+# Write to JSON file
+with open("data.json", "w", encoding="utf-8") as f:
+    json.dump(json_data, f, indent=4, ensure_ascii=False)
+
+print(f"Successfully converted CSV data to JSON format with {len(json_data)} records")

data.csv

@@ -12,7 +12,7 @@
 from sqlalchemy import create_engine, select
 from sqlalchemy.orm import Session
 
-from fastapi_app.postgres_models import Item
+from fastapi_app.postgres_models import Restaurant
 
 logger = logging.getLogger("ragapp")
 

data.json

@@ -71,14 +71,21 @@ class RetrievalResponseDelta(BaseModel):
 
 class ItemPublic(BaseModel):
     id: int
-    type: str
-    brand: str
     name: str
+    location: str
+    cuisine: str
+    rating: int
+    price_level: int
+    review_count: int
+    hours: int
+    tags: str
     description: str
-    price: float
+    menu_summary: str
+    top_reviews: str
+    vibe: str
 
 
-class ItemWithDistance(ItemPublic):
+class ItemPublicWithDistance(RestaurantPublic):
     distance: float
 
     def __init__(self, **data):

evals/generate_ground_truth.py

@@ -13,11 +13,19 @@ class Base(DeclarativeBase):
 class Item(Base):
     __tablename__ = "items"
     id: Mapped[int] = mapped_column(primary_key=True, autoincrement=True)
-    type: Mapped[str] = mapped_column()
-    brand: Mapped[str] = mapped_column()
     name: Mapped[str] = mapped_column()
+    location: Mapped[str] = mapped_column()
+    cuisine: Mapped[str] = mapped_column()
+    rating: Mapped[int] = mapped_column()
+    price_level: Mapped[int] = mapped_column()
+    review_count: Mapped[int] = mapped_column()
+    hours: Mapped[str] = mapped_column()
+    tags: Mapped[str] = mapped_column()  # JSON array
     description: Mapped[str] = mapped_column()
-    price: Mapped[float] = mapped_column()
+    menu_summary: Mapped[str] = mapped_column()
+    top_reviews: Mapped[str] = mapped_column()
+    vibe: Mapped[str] = mapped_column()
+
     # Embeddings for different models:
     embedding_3l: Mapped[Vector] = mapped_column(Vector(1024), nullable=True)  # text-embedding-3-large
     embedding_nomic: Mapped[Vector] = mapped_column(Vector(768), nullable=True)  # nomic-embed-text
@@ -33,11 +41,10 @@ def to_dict(self, include_embedding: bool = False):
         return model_dict
 
     def to_str_for_rag(self):
-        return f"Name:{self.name} Description:{self.description} Price:{self.price} Brand:{self.brand} Type:{self.type}"
-
+        return f"Name:{self.name} Description:{self.description} Location:{self.location} Cuisine:{self.cuisine} Rating:{self.rating} Price Level:{self.price_level} Review Count:{self.review_count} Hours:{self.hours} Tags:{self.tags} Menu Summary:{self.menu_summary} Top Reviews:{self.top_reviews} Vibe:{self.vibe}"
+    
     def to_str_for_embedding(self):
-        return f"Name: {self.name} Description: {self.description} Type: {self.type}"
-
+        return f"Name: {self.name} Description: {self.description} Cuisine: {self.cuisine} Tags: {self.tags} Menu Summary: {self.menu_summary} Top Reviews: {self.top_reviews} Vibe: {self.vibe}"
 
 """
 **Define HNSW index to support vector similarity search**

src/backend/fastapi_app/api_models.py

@@ -68,8 +68,9 @@ export const Answer = ({
                             return (
                                 <li key={rowId}>
                                     <h4>{citation.name}</h4>
-                                    <p className={styles.referenceMetadata}>Brand: {citation.brand}</p>
-                                    <p className={styles.referenceMetadata}>Price: {citation.price}</p>
+                                    <p className={styles.referenceMetadata}>Location: {citation.location}</p>
+                                    <p className={styles.referenceMetadata}>Price level: {citation.price_level}</p>
+                                    <p className={styles.referenceMetadata}>Rating: {citation.rating}</p>
                                     <p>{citation.description}</p>
                                 </li>
                             );