Design and add a MongoDB schema (Issue 75) - Revision (#101)

* (chore) added .DS_Store to .gitignore

* (chore) added requirements.txt file for common folder

* (feat): added the .env.example file for mongoDB

* (feat) added the basic mongoDB functions to run

* (feat) design for mongoDB schemas made and added

* (chore) removed debug statements and added credentials.json to gitignore

* (fix): fixed inconsistenicies in the code and removed debug statements

* Update frontend/streamlit_app.py

* fix checks


---------

Signed-off-by: Kannav02 <[email protected]>
Co-authored-by: Jack Luar <[email protected]>

Author: Kannav02

.gitignore

@@ -14,9 +14,11 @@ venv/
 .mypy_cache/
 .vscode/
 .venv
+.DS_Store
 
 # docs
 documents.txt
+credentials.json
 
 # virtualenv
 .venv

backend/src/agents/retriever_graph.py

@@ -196,7 +196,7 @@ def route(self, state: AgentState) -> list[str]:
             return ["retrieve_general"]
 
         if self.inbuit_tool_calling:
-            tool_names = [tool["name"] for tool in tools if "name" in tool]   # type: ignore
+            tool_names = [tool["name"] for tool in tools if "name" in tool]  # type: ignore
             return tool_names
         else:
             return tools

backend/src/api/main.py

@@ -1,5 +1,4 @@
 from fastapi import FastAPI
-
 from .routers import graphs, healthcheck
 
 app = FastAPI()

common/.env.example

@@ -0,0 +1 @@
+MONGO_DB_URI = YOUR_MONGO_DB_URI

common/mongoClient.py

@@ -0,0 +1,151 @@
+from pymongo import MongoClient
+from pymongo.database import Database
+from dotenv import load_dotenv
+import os
+
+load_dotenv()
+
+# HARDCODED collection names
+feedback_collection_name = "feedback"
+context_collection_name = "context"
+
+def get_mongo_db_client() -> Database:
+    """
+    Get the MongoDB client.
+
+    Returns:
+    - DatabaseClient: The Database client.
+
+    Note:
+    MongoDB doesn't create a collection or a database until it gets content, so no need to check if the data already exists or not.
+    """
+
+    uri = os.getenv("MONGO_DB_URI")
+    client = MongoClient(uri)
+    # this is the database that is returned by the client
+    return client["feedback_db"]
+
+def submit_feedback():
+    """
+    Submit feedback Record to the MongoDB database.
+
+    Args:
+    - question (str): The question for which feedback is being submitted.
+    - answer (str): The generated answer to the question.
+    - sources (list[str]): Source data used for the answer.
+    - context (list[str]): Additional context from the RAG.
+    - issue (str): Details about the issue.
+    - version (str): Version information.
+
+    Returns:
+    - None
+    """
+
+    feedback_db_client = get_mongo_client()
+
+    try:
+        if not check_collection_exists(feedback_collection_name,feedback_db_client,):
+            create_collection(feedback_collection_name,feedback_db_client,                validator={
+                    '$jsonSchema': {
+                        'bsonType': 'object',
+                        'required': ['question', 'answer', 'sources', 'context_ids', 'issue', 'version', 'timestamp'],
+                        'properties': {
+                            'question': {
+                                'bsonType': 'string',
+                                'description': 'must be a string and is required'
+                            },
+                            'answer': {
+                                'bsonType': 'string',
+                                'description': 'must be a string and is required'
+                            },
+                            'sources': {
+                                'bsonType': 'array',
+                                'items': {
+                                    'bsonType': 'objectId'
+                                },
+                                'description': 'must be an array of ObjectIds referencing the sources and is required'
+                            },
+                            'context': {
+                                'bsonType': 'array',
+                                'items': {
+                                    'bsonType': 'string'
+                                },
+                                'description': 'must be an array of strings and is required'
+                            },
+                            'issue': {
+                                'bsonType': 'string',
+                                'description': 'must be a string and is required'
+                            },
+                            'version': {
+                                'bsonType': 'string',
+                                'description': 'must be a string and is required'
+                            },
+                            'timestamp': {
+                                'bsonType': 'date',
+                                'description': 'must be a date and is required'
+                            },
+                            'status': {
+                                'enum': ['new', 'processing', 'resolved'],
+                                'description': 'can only be one of the enum values'
+                            }
+                        }
+                    }
+                })
+        if not check_collection_exists(context_collection_name,feedback_db_client):
+            create_collection(context_collection_name,feedback_db_client,{
+                        'bsonType': 'object',
+                        'required': ['source', 'timestamp'],
+                        'properties': {
+                            'source': {
+                                'bsonType': 'string',
+                                'description': 'must be a string and is required'
+                            },
+                            'metadata': {
+                                'bsonType': 'object',
+                                'description': 'additional metadata for the context'
+                            },
+                            'timestamp': {
+                                'bsonType': 'date',
+                                'description': 'must be a date and is required'
+                            }
+                        }
+                    }
+                )
+
+    except Exception as e:
+        print(f"Failed to submit feedback: {e}")
+        return None
+    
+def check_collection_exists(collection_name:str,client_database:Database)->bool:
+    """
+    Check if the collection exists in the database.
+
+    Args:
+    - collection_name (str): The name of the collection to check.
+    - client_database (Database): The database to check.
+
+    Returns:
+    - None
+    """
+    return collection_name in client_database.list_collection_names()
+
+def create_collection(collection_name:str,client_database:Database,validator:object)->None:
+    """
+    Create a collection in the database.
+
+    Args:
+    - collection_name (str): The name of the collection to create.
+    - client_database (Database): The database to create the collection in.
+
+    Returns:
+    - None
+    """
+    try:
+        client_database.create_collection(collection_name,validator=validator)
+        print("Collection created successfully")
+    except Exception as e:
+        print(f"Failed to create collection: {e}")
+        return None
+
+if __name__ == "__main__":
+    submit_feedback()

common/requirements.txt

@@ -0,0 +1,2 @@
+pymongo==4.6.2
+python-dotenv==1.0.1

frontend/utils/feedback.py

@@ -96,6 +96,7 @@ def submit_feedback_to_google_sheet(
     ]
 
     creds = Credentials.from_service_account_file(service_account_file, scopes=scope)
+    # this is the place where the client is coming from, need to change this for mongoDB client
     client = gspread.authorize(creds)
 
     sheet_id = os.getenv("FEEDBACK_SHEET_ID", "")
@@ -170,8 +171,13 @@ def show_feedback_form(
     feedback = st.sidebar.text_area("Please provide your feedback or report an issue:")
 
     if selected_question:
-        sources = [metadata[selected_question].get("sources", "N/A")]
-        context = [metadata[selected_question].get("context", "N/A")]
+        sources = metadata[selected_question].get("sources", ["N/A"])
+        if isinstance(sources, str):
+            sources = [sources]
+
+        context = metadata[selected_question].get("context", ["N/A"])
+        if isinstance(context, str):
+            context = [context]
 
         if st.sidebar.button("Submit"):
             selected_index = questions[selected_question]
@@ -180,8 +186,8 @@ def show_feedback_form(
             submit_feedback_to_google_sheet(
                 question=selected_question,
                 answer=gen_ans,
-                sources=sources,
-                context=context,
+                sources=sources,  # Now passing as list
+                context=context,  # Now passing as list
                 issue=feedback,
                 version=os.getenv("RAG_VERSION", get_git_commit_hash()),
             )