1. update the limit of recommendations, 2. chroma HttpClient using do…

…cker

src/__init__.py

@@ -10,11 +10,14 @@
     User,
     GithubUser,
     RepositoryRecommendation,
-    get_user_collection
+    get_user_collection,
+    # check_daily_limit
 )
 from .db import (
     recommend,
     get_topic_based_recommendations,
     get_chromadb_collection,
     upsert_to_chroma_db
-)
+)
+
+from .settings import *

src/api.py

@@ -161,9 +161,16 @@ async def preflight_verify_token():
 async def get_recommendations(request: Request, current_user: dict = Depends(get_current_user)) -> dict:
     try:
         body = await request.json()
+        username = body.get("username")
         extra_topics = body.get("extra_topics", [])
         languages = body.get("languages", [])
 
+        # TODO: v2
+        # try: 
+        #     check_daily_limit(username)
+        # except ValueError as e:
+        #     raise HTTPException(status_code=403, detail=str(e))
+
         urls = []
         user = User(username=current_user["username"], access_token=current_user["access_token"],extra_topics=extra_topics, languages=languages)
 
@@ -181,16 +188,17 @@ async def get_recommendations(request: Request, current_user: dict = Depends(get
         #     raise ValueError("Error fetching repositories")
 
         try:
+            print('Recommending\n\n')
             urls = await recommend(user_details=user_details, languages_topics=language_topics)
         except Exception as e:
             logger.error(f"Error generating recommendations: {str(e)}")
-            logger.info("Generating topic-based recommendations")
-            return get_topic_based_recommendations(user)
+            print("Error: Generating topic-based recommendations")
+            return await get_topic_based_recommendations(user)
 
-        if urls and len(urls) < 10:
+        if urls and len(urls) < 5:
             logger.info("Fewer than 10 recommendations found, fetching more repositories based on topics")
             fetched_repos = await main(language_topics, access_token=user.access_token, extra_topics=extra_topics, extra_languages=languages)
-            urls = recommend(user_details, language_topics)
+            urls = await recommend(user_details=user_details, languages_topics=language_topics)
 
         seen_full_names = set()
         unique_recommendations = []
@@ -204,6 +212,7 @@ async def get_recommendations(request: Request, current_user: dict = Depends(get
         if not unique_recommendations:
             return {'recommendations': [], 'message': 'No recommendations found'}
 
+        # update_daily_limit(username) # updates the daily limit of the user.
         return {'recommendations': unique_recommendations[::-1]}
     except Exception as e:
         logger.error(f"Error: {str(e)}")

src/db.py

@@ -3,7 +3,9 @@
 import random
 import os
 import logging
-from typing import List
+import asyncio
+import settings
+from typing import List, Optional
 from datetime import datetime, timezone
 from src.models import RepositoryRecommendation
 from src.oai import generate_embeddings
@@ -39,7 +41,7 @@ async def recommend(user_details=None,
             if results['metadatas'][0]:
                 metadatas = results["metadatas"][0]
 
-                for metadata in metadatas:
+                for metadata in metadatas[:4]: # considering only the top 4 recommendations
                     repo_name = metadata.get("full_name")
                     print('----------------\n Repo Name:', repo_name)
                     print('\n------------------')
@@ -59,13 +61,12 @@ async def recommend(user_details=None,
                                 "topics": metadata.get("topics", [])
                             })
                             recommended_repos.add(repo_url)
-                            print('\n--------------\nRecommendations:', recommendations)
-                    if len(recommendations) >= max_recommendations:
-                        break
+                    # if len(recommendations) >= max_recommendations:
+                    #     break
             else:
                 logger.info(f"No recommendations found for project {user_proj['project_name']}")
 
-    if topics:
+    if topics and not user_details:
         logger.info(f"Querying ChromaDB for topics: {topics}")
         embeddings = [generate_embeddings(topic) for topic in topics]
 
@@ -97,8 +98,8 @@ async def recommend(user_details=None,
                                 "topics": metadata.get("topics", [])
                             })
                             recommended_repos.add(repo_url)
-                            if len(recommendations) >= max_recommendations:
-                                break
+                    if len(recommendations) >= max_recommendations:
+                        break
         else:
             logger.info(f"No recommendations found for project {user_proj['project_name']}")
 
@@ -109,7 +110,7 @@ async def recommend(user_details=None,
 async def get_topic_based_recommendations(user):
     all_topics = user.languages + user.extra_topics
     if not all_topics:
-        raise ValueError("Please provide at least one language or topic")
+        raise ValueError("Please provide at least one language or topic, no projects found on your profile")
 
     # Get recommendations based on topics
     try:
@@ -126,15 +127,19 @@ async def get_topic_based_recommendations(user):
 
 def get_chromadb_collection():
     try:
-        # client = chromadb.PersistentClient(path=r"C:\Users\hrush\OneDrive - Student Ambassadors\Desktop\Open-Source-Recommender\chroma")
-        project_dir = os.path.abspath(os.path.join(os.path.dirname(__file__), ".."))
-        db_path = os.path.join(project_dir, "chroma")
-
-        client = chromadb.PersistentClient(path=db_path)
-        collection = client.get_or_create_collection("projects")
+        if settings.DEBUG:
+            print('Using local chromadb')
+            project_dir = os.path.abspath(os.path.join(os.path.dirname(__file__), ".."))
+            db_path = os.path.join(project_dir, "chroma")
+            client = chromadb.PersistentClient(path=db_path)
+        else:
+            client = chromadb.HttpClient(host=os.getenv('CHROMA_HOST'), port=8000)
+
+        collection = client.get_or_create_collection(name="projects",
+                                                     metadata={"hnsw:space": "cosine"})
         return collection
     except DatabaseError as e:
-        raise Exception(f"Error in getting collection: {e}")
+        raise DatabaseError(f"Error in getting collection: {e}")
 
 
 
@@ -206,7 +211,7 @@ def upsert_to_chroma_db(collection, unique_repos):
 
     except Exception as e:
         # Catch any unexpected errors during upsert
-        raise Exception(f"Unexpected error upserting data to ChromaDB: {e}")
+        raise ValueError(f"Unexpected error upserting data to ChromaDB: {e}")
 
 
 def convert_to_readable_format(time_str):
@@ -219,7 +224,7 @@ def convert_to_readable_format(time_str):
     return readable_time_str
 
 
-if __name__ == "__main__":
+async def main():
     collection = get_chromadb_collection()
     print(collection)
 
@@ -233,10 +238,14 @@ def convert_to_readable_format(time_str):
         'languages': ['Python', 'typescript'],
         'topics': ['agentic-ai', 'openai', "GPT", "llm"]
     }
+    topics = ["docker", "kubernetes", "devops"]
     try:
-        recommendations = recommend(user_details, languages_topics)
+        recommendations = await recommend(user_details=user_details, languages_topics=languages_topics, topics=topics)
         # logger.info(recommendations)
         print('--------')
         print(recommendations)
     except Exception as e:
-        print(f"Error Recommending data: {e}")
+        print(f"Error Recommending data: {e}")
+
+if __name__ == "__main__":
+    asyncio.run(main())

src/models.py

@@ -75,3 +75,27 @@ class RepositoryRecommendation(BaseModel):
     language: str
     updated_at: str
     topics: str
+
+# TODO: v2
+# def update_daily_limit_to_all_users():
+#     try:
+#         user_collection = db['users']
+#         user_collection.update_many({}, {"$set": {"daily_limit": 2}})
+#     except Exception as e:
+#         logger.error(f"Failed to update daily limit: {str(e)}")
+#         raise ValueError("Failed to update daily limit")
+
+
+# def check_and_update_daily_limit(username: str):
+#     user_collection = db['users']
+#     result = user_collection.find_one_and_update(
+#         {"username": username, "daily_limit": {"$gt": 0}},
+#         {"$inc": {"daily_limit": -1}},
+#         return_document=True
+#     )
+
+#     if not result:
+#         logger.warning(f"Daily limit exceeded for user: {username}")
+#         raise ValueError("Daily limit exceeded")
+
+#     return result

src/settings.py

@@ -0,0 +1,5 @@
+import os
+from datetime import timedelta
+
+
+DEBUG = False

src/user_data.py

@@ -11,7 +11,6 @@
 
 logger = logging.getLogger(__name__)
 
-GPAT = os.getenv('GPAT')
 
 async def get_repos(user):
     """
@@ -71,15 +70,15 @@ async def get_repos(user):
 
     return user_details, language_topics
 
-if __name__ == '__main__':
-    from .models import User
-    username = 'Hk669'
-    user = User(username=username, access_token=GPAT)
-
-    loop = asyncio.get_event_loop()
-    user_details, language_topics = loop.run_until_complete(get_repos(user))
-    loop.close()
-    print('-----')
-    print(user_details)
-    print('-----')
-    print(language_topics)
+# if __name__ == '__main__':
+#     from .models import User
+#     username = 'Hk669'
+#     user = User(username=username, access_token=GPAT)
+
+#     loop = asyncio.get_event_loop()
+#     user_details, language_topics = loop.run_until_complete(get_repos(user))
+#     loop.close()
+#     print('-----')
+#     print(user_details)
+#     print('-----')
+#     print(language_topics)