phase3/quangdt: Tiny changes

.gitignore

@@ -5,3 +5,4 @@ credentials/
 data/CV
 data/JD
 data/QUESTION
+!.gitkeep

app/modules/__init__.py

@@ -6,6 +6,7 @@ from app.modules.matching_cv import cvmatching_router
 from app.modules.crud_question_test import crud_question_tests_router
 from app.modules.crud_cvs import crud_cvs_router
 from app.modules.crud_jds import crud_jds_router
+from app.modules.question_rag import quiz_gen_router
 
 modules_router = APIRouter(prefix="/modules", tags=["modules"])
 modules_router.include_router(qtretrieval_router)
@@ -13,6 +14,7 @@ modules_router.include_router(cvmatching_router)
 modules_router.include_router(crud_question_tests_router)
 modules_router.include_router(crud_cvs_router)
 modules_router.include_router(crud_jds_router)
+modules_router.include_router(quiz_gen_router)
 
 @modules_router.get("/")
 async def index():

app/modules/question_rag/__init__.py

@@ -0,0 +1,26 @@
+from fastapi import APIRouter, UploadFile, File
+from typing import Annotated
+
+from app.modules.question_rag.models.question_rag_logic import question_rag
+from app.modules.question_tests_retrieval.models.jd2text import jobdes2text
+
+quiz_gen_router = APIRouter(prefix="/quiz_gen", tags=["quiz_gen"])
+
+@quiz_gen_router.get("/")
+async def index():
+    return {"message": "Welcome to quiz generator page"}
+
+@quiz_gen_router.post("/quiz_gen")
+# only upload .txt file
+async def quiz_gen(txt_file: Annotated[UploadFile, File(..., description="The JD file (only .txt file)", media_type=["text/plain"])]):
+    try:
+        # read the txt file with format
+        jobdes = txt_file.file.read().decode("utf-8")
+        sumaryjd_text = jobdes2text(jobdes)
+        if question_rag(sumaryjd_text):
+            return {"message": "Generate quiz success",
+                    "quiz": question_rag(sumaryjd_text)}
+        else:
+            return {"message": "Please upload only .txt file", "error": str(e)}
+    except Exception as e:
+        return {"message": "Please upload only .txt file", "error": str(e)}

app/modules/question_rag/models/question_rag_logic.py

@@ -0,0 +1,57 @@
+''' 
+Input : str of job description
+Output : str of quiz
+'''
+
+from langchain_google_genai import ChatGoogleGenerativeAI
+from langchain_community.document_loaders import Docx2txtLoader
+from langchain_google_genai import GoogleGenerativeAIEmbeddings
+from langchain_community.vectorstores import FAISS
+from langchain_text_splitters import RecursiveCharacterTextSplitter
+from langchain_core.prompts import ChatPromptTemplate
+from langchain.chains.combine_documents import create_stuff_documents_chain
+from langchain.chains import create_retrieval_chain
+
+import os
+from dotenv import load_dotenv
+
+# load the environment variables
+load_dotenv()
+
+# Define the google api key
+os.environ['GOOGLE_API_KEY'] = os.getenv('GOOGLE_API_KEY')
+GOOGLE_API_KEY = os.environ.get("GOOGLE_API_KEY")
+
+# JOB_TEXT = "Job Title is Senior Python Software Engineer, Level is Senior, and Brief summary of required skills is 5+ years of professional Python development experience, Expertise in Python and its frameworks."
+
+def question_rag(jobtext):
+    llm = ChatGoogleGenerativeAI(model="gemini-pro")
+
+    loader = Docx2txtLoader("data/w3school_data.docx")
+
+    docs = loader.load()
+
+    embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001")
+
+    text_splitter = RecursiveCharacterTextSplitter()
+    documents = text_splitter.split_documents(docs)
+    vector = FAISS.from_documents(documents, embeddings)
+
+    prompt = ChatPromptTemplate.from_template("""Answer the question based only on the following context:                                
+    <context>
+    {context}
+    </context>
+                                                                            
+    Generate a 10 quiz suitable for the given job description "{input}". Do not include "All of the above" answers.
+    Output format is JSON:
+    ("count": 10, "data": ( "id": "", "question": "", "choices": [ "A. ", "B. ", "C.", "D. " ], "explanation": "", "answer": "", "level": "", "domain": "" )).
+    About level help me three levels: "Fresher, Junior, Senior".
+                                            """)
+
+    document_chain = create_stuff_documents_chain(llm, prompt)
+
+    retriever = vector.as_retriever()
+    retrieval_chain = create_retrieval_chain(retriever, document_chain)
+    response = retrieval_chain.invoke({"input": jobtext})
+
+    return response["answer"]

app/modules/question_tests_retrieval/models/jd2text.py

@@ -15,7 +15,7 @@ GOOGLE_API_KEY = os.environ.get("GOOGLE_API_KEY")
 # define the parser object
 parser = JsonOutputParser()
 
-def jobdes2text(jobdes):
+def jobdes2text(jobdes: str) -> str:
     # setup the gemini pro
     llm = ChatGoogleGenerativeAI(model="gemini-pro", temperature=0.3, convert_system_message_to_human=True, api_key=GOOGLE_API_KEY)
 

requirements.txt

@@ -13,6 +13,7 @@ colorama==0.4.6
 cryptography==42.0.5
 dataclasses-json==0.6.4
 dnspython==2.6.1
+docx2txt==0.8
 email_validator==2.1.1
 fastapi==0.110.0
 firebase-admin==6.4.0
@@ -67,7 +68,7 @@ protobuf==4.25.3
 pyasn1==0.5.1
 pyasn1-modules==0.3.0
 pycparser==2.21
-pydantic==2.6.3 
+pydantic==2.6.3
 pydantic-extra-types==2.6.0
 pydantic-settings==2.2.1
 pydantic_core==2.16.3
@@ -76,6 +77,7 @@ pyparsing==3.1.2
 python-docx==1.1.0
 python-dotenv==1.0.1
 python-multipart==0.0.9
+pywin32==306
 PyYAML==6.0.1
 qdrant-client==1.8.0
 requests==2.31.0