Merge pull request #48 from monkey-patch-sdk/demo/only-food

Demo/only food

apps/app.py

@@ -6,7 +6,6 @@
 
 # Configure sys path to include src dir for monkey patching
 sys.path.append("../src")
-from monkey import Monkey
 
 from dotenv import load_dotenv
 
@@ -30,22 +29,8 @@
     allow_headers=["*"],
 )
 
-
 app.include_router(router)
 
-# origins = [CLIENT_URL]
-# if APP_ENV != "production":
-#     origins.append("https://demo.paperplane.ai")
-
-# app.add_middleware(
-#     CORSMiddleware,
-#     allow_origins=origins,
-#     allow_credentials=True,
-#     allow_methods=["*"],
-#     allow_headers=["authorization", "x-app-version"],
-# )
-
-
 if __name__ == "__main__":
     uvicorn.run("app:app", host="0.0.0.0", port=8000, reload=True)
     # uvicorn.run("app:app", host="0.0.0.0", port=8000, reload=True, log_config=log_config)

apps/controllers/food_controller.py

@@ -1,8 +1,13 @@
 from typing import List, Dict
-from services.food_service import specific_ratings, recommended_dishes
+from services.food_service import (
+    specific_ratings,
+    recommended_dishes,
+    test_specific_ratings,
+)
 
 
 def get_ratings(reviews: List[str]) -> Dict[str, int]:
+    test_specific_ratings()
     rating = specific_ratings(reviews)
     return rating.model_dump()
 

apps/extenders/yelp_scraper.py

@@ -0,0 +1,32 @@
+import json
+import os
+
+from apify_client import ApifyClient
+
+# Initialize the ApifyClient with your API token
+APIFY_KEY = os.getenv("APIFY_API_KEY")
+client = ApifyClient(APIFY_KEY)
+
+
+def yelp_scraper(url: str):
+    # Prepare the Actor input
+    run_input = {
+        "debugLog": False,
+        "directUrls": [url],
+        "maxImages": 0,
+        "proxy": {"useApifyProxy": True, "apifyProxyGroups": ["RESIDENTIAL"]},
+        "reviewLimit": 20,
+        "reviewsLanguage": "ALL",
+        "scrapeReviewerName": False,
+        "scrapeReviewerUrl": False,
+        "searchLimit": 1,
+    }
+
+    # Run the Actor and wait for it to finish
+    run = client.actor("yin/yelp-scraper").call(run_input=run_input)
+
+    # Fetch and print Actor results from the run's dataset (if there are any)
+    items = client.dataset(run["defaultDatasetId"]).get_items_as_bytes()
+    items = json.loads(items)
+
+    return items

apps/routers/food_router.py

@@ -17,10 +17,18 @@
 @router.get("/")
 async def analyze_reviews(url: str):
     reviews = get_yelp_reviews(url)
+    max_reviews = 20
+    if len(reviews) == 0:
+        return {
+            "message": "error",
+        }
+    if len(reviews) > max_reviews:
+        reviews = reviews[:max_reviews]
+
     ratings = get_ratings(reviews)
-    best_dishes = get_best_dishes(reviews)
+    # best_dishes = get_best_dishes(reviews)
     return {
         "message": "success",
         "ratings": ratings,
-        "best_dishes": best_dishes,
+        # "best_dishes": best_dishes,
     }

apps/services/food_service.py

@@ -1,50 +1,105 @@
 from pydantic import Field, BaseModel
 from typing import List, Annotated, Dict
-from monkey import Monkey
+
+from monkey_patch.monkey import Monkey
+
 import openai
 import requests
 from bs4 import BeautifulSoup
 
 # from dotenv import load_dotenv
 import os
 
+from extenders.yelp_scraper import yelp_scraper
 
 # load_dotenv()
 openai.api_key = os.getenv("OPENAI_API_KEY")
 
 
+from zenrows import ZenRowsClient
+
+client = ZenRowsClient(os.getenv("ZENROWS_API_KEY"))
+
+
+class RatingItem(BaseModel):
+    rating: int = Field(int, ge=1, le=10)
+    confidence: int = Field(
+        int, ge=1, le=10, description="How confident are you in your answer?"
+    )
+
+
 class RatingModel(BaseModel):
-    food: int = Annotated[
-        Field(..., ge=1, le=10), "The food-only rating based on the provided reviews"
-    ]
-    service: int = Annotated[
-        Field(..., ge=1, le=10), "The service-only rating based on the provided reviews"
-    ]
-    atmosphere: int = Annotated[
-        Field(..., ge=1, le=10),
-        "The atmosphere-only rating based on the provided reviews",
-    ]
-    location: int = Annotated[
-        Field(..., ge=1, le=10),
-        "The location-only rating based on the provided reviews",
-    ]
+    food: RatingItem = Field(..., description="The food-only rating")
+    service: RatingItem = Field(..., description="The service-only rating")
+    atmosphere: RatingItem = Field(..., description="The atmosphere-only rating")
+    location: RatingItem = Field(..., description="The location-only rating")
 
 
 @Monkey.patch
 def specific_ratings(reviews: List[str]) -> RatingModel:
     """
     based on the reviews, separately rate (from 1 to 10) the following aspects:
     - food
+        rating: 1-10
+        confidence: 1-10
     - service
+        rating: 1-10
+        confidence: 1-10
     - atmosphere
+        rating: 1-10
+        confidence: 1-10
     - location
-    - overall
+        rating: 1-10
+        confidence: 1-10
     """
 
 
 @Monkey.align
 def test_specific_ratings():
     """We can test the function as normal using Pytest or Unittest"""
+    assert specific_ratings(["The food was great"]) == RatingModel(
+        food=RatingItem(rating=10, confidence=10),
+        service=RatingItem(rating=5, confidence=1),
+        atmosphere=RatingItem(rating=5, confidence=1),
+        location=RatingItem(rating=5, confidence=1),
+    )
+    assert specific_ratings(["The service was great"]) == RatingModel(
+        food=RatingItem(rating=5, confidence=1),
+        service=RatingItem(rating=10, confidence=10),
+        atmosphere=RatingItem(rating=5, confidence=1),
+        location=RatingItem(rating=5, confidence=1),
+    )
+    assert specific_ratings(
+        [
+            "The atmosphere was kinda average, the location was really nice, and the service was shitty"
+        ]
+    ) == RatingModel(
+        food=RatingItem(rating=5, confidence=1),
+        service=RatingItem(rating=1, confidence=10),
+        atmosphere=RatingItem(rating=5, confidence=10),
+        location=RatingItem(rating=10, confidence=10),
+    )
+    assert specific_ratings(
+        [
+            "The food was the worst I've ever had, the service is a little slow but not bad, the atmosphere was really nice but kinda loud, and the location was perfect except that it was a little hard to find parking"
+        ]
+    ) == RatingModel(
+        food=RatingItem(rating=1, confidence=10),
+        service=RatingItem(rating=4, confidence=10),
+        atmosphere=RatingItem(rating=7, confidence=10),
+        location=RatingItem(rating=9, confidence=10),
+    )
+    assert specific_ratings(
+        [
+            "The food was the best I've ever had",
+            "The food was the worst I've ever had",
+        ]
+    ) == RatingModel(
+        food=RatingItem(rating=5, confidence=5),
+        service=RatingItem(rating=5, confidence=1),
+        atmosphere=RatingItem(rating=5, confidence=1),
+        location=RatingItem(rating=5, confidence=1),
+    )
 
 
 @Monkey.patch
@@ -59,9 +114,13 @@ def test_recommended_dishes():
     """We can test the function as normal using Pytest or Unittest"""
 
 
-def get_yelp_reviews(yelp_url: str) -> List[str]:
-    response = requests.get(yelp_url)
-    soup = BeautifulSoup(response.text, "html.parser")
-    # get all "p" tags with class beginning with comment__
-    reviews = soup.find_all("p", class_=lambda x: x and x.startswith("comment__"))
-    return [review.text for review in reviews]
+def get_yelp_reviews(business_id_or_alias: str) -> List[str]:
+    url = f"https://www.yelp.com/biz/{business_id_or_alias}"
+
+    params = {"autoparse": "true"}
+    response = client.get(url, params=params)
+    if "review" not in response.json():
+        return []
+    res_json = response.json()["review"]
+    reviews = [review["description"] for review in res_json]
+    return reviews

requirements.txt

@@ -12,3 +12,6 @@ flake8==6.1.0
 black==23.10.0
 mypy==1.6.1
 beautifulsoup4==4.12.2
+apify_client
+zenrows
+monkey-patch.py