final demo code

app.py

@@ -3,7 +3,7 @@ from llama_guard import moderate_chat, get_category_name
 import time
 from chat_agent import convo, main
 from chat_agent import choose_model1, delete_all_variables
-from recommendation_agent import recommend2, choose_model2, is_depressed
+from recommendation_agent import recommend2, choose_model2, is_depressed, start_recommend
 from functools import cached_property
 from streamlit_js_eval import streamlit_js_eval
 
@@ -19,11 +19,12 @@ st.title('BrighterDays Mentor')
 # Adjust sidebar width to take half the screen
 col1, col2 = st.columns([2, 3])
 
-model = st.sidebar.selectbox(label="Choose the LLM model", options=["Venilla Model", "Fine Tuned Model"])
+model = st.sidebar.selectbox(label="Choose the LLM model", options=["mistral-7b-base-model", "mental-health-mistral-7b-finetuned-model"])
 print("\n\nSelected LLM model from Dropdown",model)
 choose_model1(model)
 choose_model2(model)
 main()
+start_recommend()
 # Function to update recommendations in col1
 def update_recommendations(sum):
     # with col1:

chat_agent.py

@@ -23,12 +23,12 @@ HUGGINGFACEHUB_API_TOKEN  = os.environ["HUGGINGFACEHUB_API_TOKEN"]
 repo_id ="mistralai/Mistral-7B-Instruct-v0.2"
 def choose_model1(model):
     global repo_id
-    if model == "Venilla Model":
+    if model == "mistral-7b-base-model":
         repo_id="mistralai/Mistral-7B-Instruct-v0.2"
-        print("model chooosed from chat",repo_id)
+        print("model chosen from chat",repo_id)
     else:
         repo_id="GRMenon/mental-health-mistral-7b-instructv0.2-finetuned-V2"
-        print("model chooosed from chat",repo_id)
+        print("model chosen from chat",repo_id)
 
 query2 = " "
 def main():

nlp_models.py

@@ -0,0 +1,43 @@
+import torch
+from transformers import DistilBertForSequenceClassification
+import os
+# # Get the directory path of the current script
+# script_dir = os.path.dirname(os.path.abspath(__file__))
+# model = DistilBertForSequenceClassification.from_pretrained("model.safetensors")
+
+# Load model directly
+from transformers import AutoTokenizer, AutoModelForSequenceClassification
+
+tokenizer = AutoTokenizer.from_pretrained("lxs1/DistilBertForSequenceClassification_6h_768dim")
+model = AutoModelForSequenceClassification.from_pretrained("lxs1/DistilBertForSequenceClassification_6h_768dim")
+
+
+# from transformers import DistilBertTokenizerFast
+# tokenizer = DistilBertTokenizerFast.from_pretrained('distilbert-base-uncased')
+
+# Move the model to the GPU if available
+device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+model.to(device)
+
+def sentiment_class(summarized_text):    
+    '''
+    # 1 = non-depressed
+    # 0 = depressed
+    returns: example:- array([[0.00493283, 0.9950671 ]], dtype=float32)
+    '''
+    #inputs = tokenizer(summarized_text, padding = True, truncation = True, return_tensors='pt').to('cuda')
+    inputs = tokenizer(summarized_text, padding = True, truncation = True, return_tensors='pt').to(device)
+
+    outputs = model(**inputs)
+
+    predictions = torch.nn.functional.softmax(outputs.logits, dim=-1)
+    predictions = predictions.cpu().detach().numpy()
+    return predictions
+
+def pattern_classification():
+    result=""
+    return result
+
+def corelation_analysis():
+    result=""
+    return result

recommendation_agent.py

@@ -22,47 +22,53 @@ load_dotenv(find_dotenv())
 HUGGINGFACEHUB_API_TOKEN  = os.environ["HUGGINGFACEHUB_API_TOKEN"]
 
 repo_id ="mistralai/Mistral-7B-Instruct-v0.2"
+llm = None
+
 def choose_model2(model):
     global repo_id
-    if model == "Venilla Model":
+    global llm
+
+    if model == "mistral-7b-base-model":
         repo_id="mistralai/Mistral-7B-Instruct-v0.2"
         print("model chooosed from recomm",repo_id)
     else:
         repo_id="GRMenon/mental-health-mistral-7b-instructv0.2-finetuned-V2"
         print("model chooosed from recomm",repo_id)
 
-llm = HuggingFaceEndpoint(
-    repo_id=repo_id, max_length=512, temperature=0.5, token=HUGGINGFACEHUB_API_TOKEN
-)
+    llm = HuggingFaceEndpoint(
+        repo_id=repo_id, max_length=512, temperature=0.5, token=HUGGINGFACEHUB_API_TOKEN
+    )
 
-persist_directory="Data/chroma"
-#chroma_client = chromadb.PersistentClient(persist_directory=persist_directory)
-embedding_function = SentenceTransformerEmbeddings(model_name="all-MiniLM-L6-v2")
-vectors = Chroma(persist_directory = persist_directory, embedding_function = embedding_function, collection_name="split_parents")
-retriever = vectors.as_retriever() #(k=6)
-
-#prompt="you are a mental health therapist, talking to a person with who is facing some mental health issues. Following is the user feeling {question}"
-
-prompt = """You're a Mental Health Specialist. Support those with Depressive Disorder. Your task is to provide mental health advice.
-Listen compassionately, respond helpfully. For casual talk, be friendly. For facts, use context.
-Following is the user feeling {question}.
-If unsure, say, 'Out of my knowledge.' Always stay direct.
-If you cannot find the answer from the pieces of context, just say that you don't know, don't try to make up an answer.
-PLEASE GIVE THE RESPONSE IN THE FORM OF BULLET POINTS.
-----------------
-{context}"""
-
-prompt = PromptTemplate(input_variables=['question'],template=prompt)
-
-chain1 = LLMChain(llm=llm, prompt=prompt, verbose=True)
-doc_chain = load_qa_chain(llm, chain_type="stuff")
-
-chain = ConversationalRetrievalChain(
-       retriever=retriever,
-       question_generator=chain1,
-       combine_docs_chain=doc_chain,
-       verbose=True,
-)
+def start_recommend():
+    persist_directory="Data/chroma"
+    #chroma_client = chromadb.PersistentClient(persist_directory=persist_directory)
+    embedding_function = SentenceTransformerEmbeddings(model_name="all-MiniLM-L6-v2")
+    vectors = Chroma(persist_directory = persist_directory, embedding_function = embedding_function, collection_name="split_parents")
+    retriever = vectors.as_retriever() #(k=6)
+
+    #prompt="you are a mental health therapist, talking to a person with who is facing some mental health issues. Following is the user feeling {question}"
+
+    prompt = """You're a Mental Health Specialist. Support those with Depressive Disorder. Your task is to provide brief mental health advice. PLEASE KEEP IT BRIEF.
+    Listen compassionately, respond helpfully. For casual talk, be friendly. For facts, use context.
+    Following is the user feeling {question}.
+    If unsure, say, 'Out of my knowledge.' Always stay direct.
+    If you cannot find the answer from the pieces of context, just say that you don't know, don't try to make up an answer.
+    PLEASE GIVE THE BRIEF RESPONSE IN THE FORM OF BULLET POINTS.
+    ----------------
+    {context}"""
+
+    global chain
+    prompt = PromptTemplate(input_variables=['question'],template=prompt)
+
+    chain1 = LLMChain(llm=llm, prompt=prompt, verbose=True)
+    doc_chain = load_qa_chain(llm, chain_type="stuff")
+
+    chain = ConversationalRetrievalChain(
+        retriever=retriever,
+        question_generator=chain1,
+        combine_docs_chain=doc_chain,
+        verbose=True,
+    )
 
 
 def recommend2(query):
@@ -88,4 +94,21 @@ def is_depressed(human_inputs):
     # return status[0]["label"]
 
 
-# print(recommend2("i am feeling sad"))
+def is_depressed_from_nlp_model(human_inputs):
+    ''''
+    returns wether according to human inputs the person is depressed or not
+    '''
+    # Implement Classification
+    # all_user_inputs = ''.join(human_inputs)
+    from nlp_models import sentiment_class
+    predictions = sentiment_class(human_inputs)
+    # Assuming the predictions are shaped (1, 2) as in the example
+    probability_depressed, probability_not_depressed = predictions[0]
+    
+    if probability_depressed > probability_not_depressed:
+        return "Depressed", probability_depressed
+    else:
+        return "Not depressed", probability_not_depressed
+    #return 'Not so depressed' if status[0][1] > 0.5 else 'is_depressed'
+    #return 'Is depressed' if status[0]["label"] == "NEGATIVE" else 'Not Depressed'
+    # return status[0]["label"]