too many changes, hope this works

app.py

@@ -1,6 +1,6 @@
 from fastapi import FastAPI
-from routes import router  # routes.py must be in same folder
-from question import askMe
+from routes.category import router  # routes.py must be in same folder
+from routes.question import askMe
 from dotenv import load_dotenv
 from cache_init import fetch_and_cache_articles
 

models_initialization/bart_large_registry.py

@@ -0,0 +1,56 @@
+import os
+import requests
+from dotenv import load_dotenv
+
+load_dotenv()
+
+HF_TOKEN = os.getenv("HF_TOKEN")
+HF_BART_MODEL = "facebook/bart-large-mnli"
+HF_API_URL = f"https://api-inference.huggingface.co/models/{HF_BART_MODEL}"
+
+HEADERS = {
+    "Authorization": f"Bearer {HF_TOKEN}",
+    "Content-Type": "application/json"
+}
+
+def run_zero_shot_classification(prompt: str, candidate_labels: list[str]) -> str:
+    payload = {
+        "inputs": prompt,
+        "parameters": {
+            "candidate_labels": candidate_labels
+        }
+    }
+
+    try:
+        response = requests.post(HF_API_URL, headers=HEADERS, json=payload, timeout=20)
+        response.raise_for_status()
+        data = response.json()
+        if isinstance(data, dict) and "labels" in data and data["labels"]:
+            return data["labels"][0]  # Most confident label
+    except Exception as e:
+        print("[BART Zero-Shot Error]:", str(e))
+
+    return ""
+
+def run_entailment_check(premise: str, hypothesis: str) -> bool:
+    payload = {
+        "inputs": {
+            "premise": premise,
+            "hypothesis": hypothesis
+        }
+    }
+
+    try:
+        response = requests.post(HF_API_URL, headers=HEADERS, json=payload, timeout=20)
+        response.raise_for_status()
+        data = response.json()
+        if isinstance(data, dict) and "labels" in data:
+            labels = data["labels"]
+            scores = data["scores"]
+            if "entailment" in labels:
+                entailment_score = scores[labels.index("entailment")]
+                return entailment_score > 0.5
+    except Exception as e:
+        print("[BART Entailment Error]:", str(e))
+
+    return False

models_initialization/mistral_registry.py

@@ -0,0 +1,34 @@
+import os
+import json
+import requests
+from dotenv import load_dotenv
+
+load_dotenv()
+
+HF_TOKEN = os.getenv("HF_TOKEN")
+HF_API_URL = "https://api-inference.huggingface.co/models/mistralai/Mistral-7B-Instruct-v0.3"
+
+HEADERS = {
+    "Authorization": f"Bearer {HF_TOKEN}",
+    "Content-Type": "application/json"
+}
+
+def mistral_generate(prompt: str, max_new_tokens=128, temperature=0.7) -> str:
+    payload = {
+        "inputs": prompt,
+        "parameters": {
+            "max_new_tokens": max_new_tokens,
+            "temperature": temperature
+        }
+    }
+
+    try:
+        response = requests.post(HF_API_URL, headers=HEADERS, data=json.dumps(payload), timeout=30)
+        response.raise_for_status()
+        result = response.json()
+        if isinstance(result, list) and result:
+            return result[0].get("generated_text", "").strip()
+    except Exception as e:
+        print("Mistral API error:", e)
+
+    return ""

nuse_modules/classifier.py

@@ -1,12 +1,10 @@
 import os
 import re
-import requests
 from dotenv import load_dotenv
 
-load_dotenv()
+from models_initialization.bart_large_registry import run_zero_shot_classification
 
-HF_TOKEN = os.getenv("HF_TOKEN")
-HF_ZERO_SHOT_MODEL = "facebook/bart-large-mnli"
+load_dotenv()
 
 # Map readable categories to numeric IDs
 QUESTION_TYPES = {
@@ -60,39 +58,9 @@ def rule_based_classify(prompt: str) -> int:
     return -1
 
 
-# ---------- Step 2: HF Zero-Shot Fallback ----------
-def zero_shot_classify(prompt: str) -> int:
-    candidate_labels = list(QUESTION_TYPES.keys())
-    payload = {
-        "inputs": prompt,
-        "parameters": {
-            "candidate_labels": candidate_labels
-        }
-    }
-
-    headers = {
-        "Authorization": f"Bearer {HF_TOKEN}",
-        "Content-Type": "application/json"
-    }
-
-    url = f"https://api-inference.huggingface.co/models/{HF_ZERO_SHOT_MODEL}"
-
-    try:
-        res = requests.post(url, headers=headers, json=payload, timeout=20)
-        res.raise_for_status()
-        data = res.json()
-        if isinstance(data, dict) and "labels" in data:
-            top_label = data["labels"][0]
-            return QUESTION_TYPES.get(top_label, -1)
-    except Exception as e:
-        print("[HF Classifier Error]", str(e))
-    
-    return -1
-
-
 # ---------- Public Hybrid Classifier ----------
 def classify_question(prompt: str) -> int:
     rule_result = rule_based_classify(prompt)
     if rule_result != -1:
         return rule_result
-    return zero_shot_classify(prompt)
+    return run_zero_shot_classification(prompt, candidate_labels=list(QUESTION_TYPES.keys()))

nuse_modules/keyword_extracter.py

@@ -4,34 +4,7 @@ import os
 import requests
 import json
 
-HF_TOKEN = os.getenv("HF_TOKEN")
-
-HF_API_URL = "https://api-inference.huggingface.co/models/mistralai/Mistral-7B-Instruct-v0.3"
-HEADERS = {
-    "Authorization": f"Bearer {HF_TOKEN}",
-    "Content-Type": "application/json"
-}
-
-
-def mistral_generate(prompt: str, max_new_tokens=128) -> str:
-    payload = {
-        "inputs": prompt,
-        "parameters": {
-            "max_new_tokens": max_new_tokens,
-            "temperature": 0.7
-        }
-    }
-    try:
-        response = requests.post(HF_API_URL, headers=HEADERS, data=json.dumps(payload), timeout=30)
-        response.raise_for_status()
-        result = response.json()
-        if isinstance(result, list) and len(result) > 0:
-            return result[0].get("generated_text", "").strip()
-    except Exception as e:
-        print("[mistral_generate error]", str(e))
-
-    return ""
-
+from models_initialization.mistral_registry import mistral_generate
 
 def extract_last_keywords(raw: str, max_keywords: int = 8) -> list[str]:
     segments = raw.strip().split("\n")

routes.py → routes/category.py

@@ -1,7 +1,7 @@
 import os
 from fastapi import APIRouter
 from dotenv import load_dotenv
-from redis_client import redis_client as r
+from clients.redis_client import redis_client as r
 
 load_dotenv()
 

question.py → routes/question.py

@@ -3,24 +3,16 @@ import requests
 import json
 from fastapi import APIRouter
 from pydantic import BaseModel
-from typing import List
-from redis_client import redis_client as r
+from clients.redis_client import redis_client as r
 from dotenv import load_dotenv
-from urllib.parse import quote
 
+from models_initialization.mistral_registry import mistral_generate
 from nuse_modules.classifier import classify_question, REVERSE_MAP
 from nuse_modules.keyword_extracter import keywords_extractor
 from nuse_modules.google_search import search_google_news
 
 load_dotenv()
 
-HF_TOKEN = os.getenv("HF_TOKEN")
-HF_API_URL = "https://api-inference.huggingface.co/models/mistralai/Mistral-7B-Instruct-v0.3"
-HEADERS = {
-    "Authorization": f"Bearer {HF_TOKEN}",
-    "Content-Type": "application/json"
-}
-
 askMe = APIRouter()
 
 class QuestionInput(BaseModel):
@@ -41,26 +33,6 @@ def extract_answer_after_label(text: str) -> str:
     return text.strip()
 
 
-def mistral_generate(prompt: str, max_new_tokens=128):
-    payload = {
-        "inputs": prompt,
-        "parameters": {
-            "max_new_tokens": max_new_tokens,
-            "temperature": 0.7
-        }
-    }
-    try:
-        response = requests.post(HF_API_URL, headers=HEADERS, data=json.dumps(payload), timeout=30)
-        response.raise_for_status()
-        result = response.json()
-        if isinstance(result, list) and len(result) > 0:
-            return result[0].get("generated_text", "").strip()
-        else:
-            return ""
-    except Exception:
-        return ""
-
-
 @askMe.post("/ask")
 async def ask_question(input: QuestionInput):
     question = input.question