Auth integration

BACKEND/app/auth.py

@@ -0,0 +1,68 @@
+from fastapi import  HTTPException, status, Request
+from fastapi.security import OAuth2PasswordBearer
+from functools import wraps
+import requests
+
+oauth2_scheme = OAuth2PasswordBearer(tokenUrl="token")
+
+GITEA_API_URL = "http://localhost:3000/api/v1"
+
+def validate_token(func):
+    @wraps(func)
+    async def wrapper(*args, **kwargs):
+        request = None
+        for arg in args:
+            if isinstance(arg, Request):
+                request = arg
+                break
+
+        if request is None:
+            for key, value in kwargs.items():
+                if isinstance(value, Request):
+                    request = value
+                    break
+
+        if request is None:
+            raise HTTPException(
+                status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
+                detail="Request object not found",
+            )
+
+        auth_header = request.headers.get("Authorization")
+        if not auth_header or not auth_header.startswith("Bearer "):
+            raise HTTPException(
+                status_code=status.HTTP_401_UNAUTHORIZED,
+                detail="Missing or invalid authorization header"
+            )
+        token =  auth_header.replace("Bearer ", "")
+
+        try:
+            headers = {"Authorization": f"{token}"}
+            response = requests.get(f"{GITEA_API_URL}/user", headers=headers)
+
+
+            if response.status_code == 401:
+                raise HTTPException(
+                    status_code=status.HTTP_401_UNAUTHORIZED,
+                    detail="Token expired or invalid"
+                )
+            elif response.status_code != 200:
+                raise HTTPException(
+                    status_code=status.HTTP_401_UNAUTHORIZED,
+                    detail="Authentication failed"
+                )
+
+            if "user_info" in kwargs:
+                user_info = response.json()
+                kwargs["user_info"] = user_info
+            return await func(*args, **kwargs)
+
+        except Exception as e:
+            if isinstance(e, HTTPException):
+                raise e
+            raise HTTPException(
+                status_code=status.HTTP_401_UNAUTHORIZED,
+                detail="Invalid authentication credentials"
+            )
+
+    return wrapper

BACKEND/app/crud.py

@@ -46,11 +46,15 @@ def parse_usfm_to_csv(book_name, usfm_content, project_id):
     """ Convert USFM content to CSV format and return extracted data """
     try:
         my_parser = USFMParser(usfm_content)  # Initialize parser
-        output = my_parser.to_list(include_markers=Filter.BCV + Filter.TEXT)  # Extract BCV and Text      
+        output = my_parser.to_list(include_markers=Filter.BCV + Filter.TEXT)  # Extract BCV and Text   
         processed_output = [
-            [re.sub(r"\s+", " ", value).strip() if isinstance(value, str) else value for value in row]
-            for row in output
-        ]            
+            [normalize_text(value).replace("\n", " ") if isinstance(value, str) else value for value in row]
+            for row in output]   
+        # processed_output = [
+        #     [re.sub(r"\s+", " ", value).strip() if isinstance(value, str) else value for value in row]
+        #     for row in output
+        # ]   
+
         if not processed_output:
             logging.error(f"No data extracted for {book_name}!")
         else:

BACKEND/app/router.py

@@ -1,5 +1,5 @@
 import itertools
-from fastapi import APIRouter, HTTPException,File,UploadFile,Query
+from fastapi import APIRouter, HTTPException,File,UploadFile,Query, Request, Depends
 from fastapi import Body
 from pydantic import BaseModel
 from database import SessionLocal
@@ -14,6 +14,8 @@
 import crud
 from fastapi.responses import JSONResponse
 import base64
+from auth import validate_token
+from utils import input_token
 
 
 
@@ -35,10 +37,11 @@ class USFMUploadRequest(BaseModel):
 
 
 @router.post("/add_project/")
-async def add_project(request: ProjectRequest):
+@validate_token
+async def add_project(request: Request, project_request: ProjectRequest, user_details=Depends(input_token)):
     """ Add a new project and return the project ID """
     session = SessionLocal()
-    project_name = request.project_name.strip()
+    project_name = project_request.project_name.strip()
 
     # Validate project name (Ensure it's not empty after trimming)
     if not project_name:
@@ -47,7 +50,7 @@ async def add_project(request: ProjectRequest):
 
     # Check if the project already exists
     existing_project = session.query(Project).filter_by(
-        project_name=request.project_name
+        project_name=project_request.project_name
     ).first()
 
     if existing_project:
@@ -56,7 +59,7 @@ async def add_project(request: ProjectRequest):
 
     # Insert new project
     new_project = Project(
-        project_name=request.project_name
+        project_name=project_request.project_name
     )
     session.add(new_project)
     session.commit()
@@ -69,13 +72,19 @@ async def add_project(request: ProjectRequest):
 
 
 @router.get("/list_projects/")
-async def list_projects(project_name: str = Query(None)):
+@validate_token
+async def list_projects(request: Request, project_name: str = Query(None),
+                         user_details=Depends(input_token)):
     """
     List all projects or fetch a specific project by project_name.
     If project_name is provided, returns the matching project or null if not found.
     """
     session = SessionLocal()
     try:
+
+        # username = user_info.get("login")
+        # print(f"User {username} is fetching projects")
+
         if project_name:
             project = session.query(Project).filter(Project.project_name == project_name).first()
             return {"project": project if project else None}
@@ -95,8 +104,9 @@ async def list_projects(project_name: str = Query(None)):
 
 
 @router.post("/upload_usfm/")
-async def upload_usfm(
-    request: USFMUploadRequest
+@validate_token
+async def upload_usfm(request: Request,
+    project_request: USFMUploadRequest, user_details=Depends(input_token)
 ):
     """ 
     Upload a USFM content as a string, process it, and store data in DB.
@@ -106,9 +116,9 @@ async def upload_usfm(
 
     try:
         # Get project_id from project_name
-        project_name = request.project_name
-        usfm_sha = request.usfm_sha
-        encoded_usfm = request.encoded_usfm
+        project_name = project_request.project_name
+        usfm_sha = project_request.usfm_sha
+        encoded_usfm = project_request.encoded_usfm
         project_id = crud.get_project_id(session,project_name)
         logging.info(f"Processing USFM file for project: {project_name} (Project ID: {project_id})")
 
@@ -200,25 +210,27 @@ async def upload_usfm(
 
 
 @router.put("/update_usfm/")
-async def update_usfm(
-    request: USFMUploadRequest
+@validate_token
+async def update_usfm(request: Request,
+    project_request: USFMUploadRequest, user_details=Depends(input_token)
 ):
     """ Update an existing USFM file, reprocess it, and update both book and verse tables properly. """
     session = SessionLocal()
 
     try:
         # Extract values from request body
-        project_name = request.project_name
-        usfm_sha = request.usfm_sha
-        encoded_usfm = request.encoded_usfm
+        project_name = project_request.project_name
+        usfm_sha = project_request.usfm_sha
+        encoded_usfm = project_request.encoded_usfm
         # Extract book name from USFM
         project_id = crud.get_project_id(session,project_name)
 
         logging.info(f"Updating USFM file for project: {project_name} (Project ID: {project_id})")
         try:
             usfm_bytes = base64.b64decode(encoded_usfm)
             usfm = usfm_bytes.decode("utf-8")  # Convert bytes to string
-            usfm=crud.normalize_text(usfm)
+            #errors are not displaying properly
+            # usfm=crud.normalize_text(usfm)
         except Exception as e:
             logging.error(f"Failed to decode USFM content: {str(e)}")
             raise HTTPException(status_code=400, detail="Invalid encoded USFM content")
@@ -289,7 +301,9 @@ async def update_usfm(
 
 
 @router.get("/list_books/")
-async def list_books(project_name: str = Query(None)):
+@validate_token
+async def list_books(request: Request,project_name: str = Query(None),
+                      user_details=Depends(input_token)):
     """ Retrieve all Bibles (projects) along with their books and their status, optionally filtering by project name """
     session = SessionLocal()
     try:
@@ -328,7 +342,9 @@ async def list_books(project_name: str = Query(None)):
 
 
 @router.get("/find_missing_verses/")
-async def find_missing_verses(book_name: str, project_name: str):
+@validate_token
+async def find_missing_verses(request: Request,book_name: str, project_name: str,
+                              user_details=Depends(input_token)):
     """Find missing verses for a given book_id and project_id by comparing with versification.json."""
 
     session = SessionLocal()
@@ -400,7 +416,9 @@ async def find_missing_verses(book_name: str, project_name: str):
 
 
 @router.get("/book/usfm/")
-async def get_book_usfm(project_name: str, book_name: str):
+@validate_token
+async def get_book_usfm(request: Request,project_name: str, book_name: str,
+                              user_details=Depends(input_token)):
     """
     Get the USFM content of a book from the database.
     """
@@ -434,7 +452,9 @@ async def get_book_usfm(project_name: str, book_name: str):
 
 
 @router.get("/book/json/")
-async def get_book_json(project_name: str, book_name: str):
+@validate_token
+async def get_book_json(request: Request,project_name: str, book_name: str,
+                              user_details=Depends(input_token)):
     """
     Get the book's content in JSON format.
     """
@@ -476,7 +496,9 @@ async def get_book_json(project_name: str, book_name: str):
 
 
 @router.get("/chapter/json/")
-async def get_chapter_json(project_name: str, book_name: str, chapter: int):
+@validate_token
+async def get_chapter_json(request: Request,project_name: str, book_name: str,
+                            chapter: int, user_details=Depends(input_token)):
     """
     Get the chapter's content in JSON format.
     """
@@ -518,8 +540,10 @@ async def get_chapter_json(project_name: str, book_name: str, chapter: int):
 
 
 @router.get("/book/chapters/")
+@validate_token
 # async def get_book_chapters(book_id: int):
-async def get_book_chapters(project_name: str, book_name: str):
+async def get_book_chapters(request: Request, project_name: str, book_name: str,
+                            user_details=Depends(input_token)):
     """
     Get the list of chapters available in a book.
     """
@@ -554,10 +578,13 @@ async def get_book_chapters(project_name: str, book_name: str):
 
 
 @router.get("/parallel_corpora/withbcv/")
+@validate_token
 async def get_parallel_corpora_withbcv(
+    request: Request,
     project_name_1: str, 
     project_name_2: str, 
-    response_type: str = Query("csv", description="Set 'json' for JSON response, 'csv' for file download")
+    response_type: str = Query("csv", description="Set 'json' for JSON response, 'csv' for file download"),
+    user_details=Depends(input_token)
 ):
     """
     Generate and return the parallel corpus between two projects (two languages) in CSV or JSON format.
@@ -675,8 +702,12 @@ async def get_parallel_corpora_withbcv(
 
 
 @router.get("/parallel_corpora/withoutbcv/")
-async def get_parallel_corpora_texts(project_name_1: str, project_name_2: str,
-                                         response_type: str = Query("csv", description="Set 'json' for JSON response, 'csv' for file download")):
+@validate_token
+async def get_parallel_corpora_texts(request: Request,
+                                     project_name_1: str, project_name_2: str,
+                                         response_type: str = Query("csv",
+                                 description="Set 'json' for JSON response, 'csv' for file download"),
+                                     user_details=Depends(input_token)):
     """
     Generate and return the parallel corpus between two projects in CSV format with only Text_1 and Text_2.
     """

BACKEND/app/utils.py

@@ -0,0 +1,9 @@
+from fastapi import  Security
+from fastapi.security import HTTPBearer
+
+oauth2_scheme = HTTPBearer()
+
+async def input_token(token: str = Security(oauth2_scheme)):
+    '''Function to validate API key, currently only for swagger to have a token input'''
+    user_token = {"token": token}
+    return user_token

BACKEND/requirements.txt

@@ -4,4 +4,5 @@ SQLAlchemy
 psycopg2
 uvicorn
 python-multipart
-sacremoses
+sacremoses
+requests

UI/.dockerignore

@@ -0,0 +1,2 @@
+node_modules
+.env.local

UI/Dockerfile

@@ -0,0 +1,15 @@
+FROM node:18-alpine
+
+WORKDIR /app
+
+COPY package*.json ./
+
+RUN npm install
+
+COPY . .
+
+RUN npm run build
+
+EXPOSE 3001
+
+CMD ["npm", "run", "dev", "--", "--host", "0.0.0.0"]