Upload 2 files

17th Dec Updated App and req

app.py

@@ -0,0 +1,134 @@
+from fastapi import FastAPI
+from pydantic import BaseModel
+from typing import List, Dict, Any
+from pymongo import MongoClient
+from transformers import pipeline
+import spacy
+
+# FastAPI app setup
+app = FastAPI()
+
+# ==========================
+# MongoDB Connection Setup
+# ==========================
+connection_string = "mongodb+srv://clician:[email protected]/?retryWrites=true&w=majority&appName=Hutterdev"
+client = MongoClient(connection_string)
+db = client["test"]  # Replace with your database name
+products_collection = db["products"]  # Replace with your collection name
+
+# ==========================
+# Transformers Pipeline Setup
+# ==========================
+# Load the Question-Answering pipeline
+qa_pipeline = pipeline("question-answering", model="bert-large-uncased-whole-word-masking-finetuned-squad")
+
+# ==========================
+# Static Context Message
+# ==========================
+context_msg = (
+    "Hutter Products GmbH provides a wide array of services to help businesses create high-quality, sustainable products. "
+    "Their offerings include comprehensive product design, ensuring items are both visually appealing and functional, and product consulting, "
+    "which provides expert advice on features, materials, and design elements. They also offer sustainability consulting to integrate eco-friendly practices, "
+    "such as using recycled materials and Ocean Bound Plastic. Additionally, they manage customized production to ensure products meet the highest standards "
+    "and offer product animation services, creating realistic rendered images and animations to enhance online engagement. These services collectively enable "
+    "businesses to develop products that are sustainable, market-responsive, and aligned with their brand identity."
+)
+
+# ==========================
+# spaCy NER Setup
+# ==========================
+nlp = spacy.load("en_core_web_sm")
+
+# ==========================
+# Pydantic Models
+# ==========================
+class PromptRequest(BaseModel):
+    input_text: str
+
+class CombinedResponse(BaseModel):
+    ner: Dict[str, Any]
+    qa: Dict[str, Any]
+    products_matched: List[Dict[str, Any]]
+
+# ==========================
+# Helper Functions
+# ==========================
+def extract_keywords(text: str) -> List[str]:
+    """
+    Extract keywords (nouns and proper nouns) using spaCy.
+    """
+    doc = nlp(text)
+    keywords = [token.text for token in doc if token.pos_ in ["NOUN", "PROPN"]]
+    return list(set(keywords))
+
+def search_products_by_keywords(keywords: List[str]) -> List[Dict[str, Any]]:
+    """
+    Search MongoDB for products that match any of the extracted keywords.
+    """
+    regex_patterns = [{"name": {"$regex": keyword, "$options": "i"}} for keyword in keywords]
+    query = {"$or": regex_patterns}
+
+    matched_products = []
+    cursor = products_collection.find(query)
+    for product in cursor:
+        matched_products.append({
+            "id": str(product.get("_id", "")),
+            "name": product.get("name", ""),
+            "description": product.get("description", ""),
+            "skuNumber": product.get("skuNumber", ""),
+            "baseModel": product.get("baseModel", ""),
+        })
+
+    return matched_products
+
+def get_combined_context(products: List[Dict]) -> str:
+    """
+    Combine the static context with product descriptions fetched from MongoDB.
+    """
+    product_descriptions = " ".join([p["description"] for p in products if "description" in p and p["description"]])
+    combined_context = f"{product_descriptions} {context_msg}"
+    return combined_context
+
+# ==========================
+# FastAPI Endpoints
+# ==========================
+@app.get("/")
+async def root():
+    return {"message": "Welcome to the NER and QA API!"}
+
+@app.post("/process/", response_model=CombinedResponse)
+async def process_prompt(request: PromptRequest):
+    input_text = request.input_text
+
+    # Step 1: Extract keywords using spaCy NER
+    keywords = extract_keywords(input_text)
+    ner_response = {"extracted_keywords": keywords}
+
+    # Step 2: Search MongoDB for matching products
+    products = search_products_by_keywords(keywords)
+
+    # Step 3: Generate Combined Context
+    combined_context = get_combined_context(products)
+
+    # Step 4: Use Q&A Model
+    if combined_context.strip():  # Ensure the combined context is not empty
+        qa_input = {"question": input_text, "context": combined_context}
+        qa_output = qa_pipeline(qa_input)
+        qa_response = {
+            "question": input_text,
+            "answer": qa_output["answer"],
+            "score": qa_output["score"]
+        }
+    else:
+        qa_response = {
+            "question": input_text,
+            "answer": "No relevant context available.",
+            "score": 0.0
+        }
+
+    # Step 5: Return Combined Response    
+    return {
+        "ner": ner_response,
+        "qa": qa_response,
+        "products_matched": products
+    }

requirements.txt

@@ -0,0 +1,8 @@
+fastapi==0.98.0
+uvicorn[standard]==0.23.2
+transformers==4.34.0
+torch==2.0.1
+pydantic==1.10.9
+pymongo==4.9.2
+spacy==3.8.3
+numpy<2.0  # Compatibility with PyTorch and Transformers