Uploads

api/endpoints/upload_route.py

@@ -0,0 +1,117 @@
+"""Contains routes for uploading files and accessing uploaded files."""
+
+from __future__ import annotations
+
+from fastapi import APIRouter, File, Form, HTTPException, UploadFile
+
+from api.utils.upload_helper import (
+    calculate_md5_checksum,
+    get_all_filenames,
+    reassemble_file,
+    save_chunk,
+    save_file,
+)
+
+router = APIRouter(prefix="/upload", tags=["upload"])
+
+
+@router.post("/chunk")
+async def upload_chunk(
+    file: UploadFile = None,
+    chunk_number: int = Form(...),
+    total_chunks: int = Form(...),
+    chunk_hash: str = Form(...),
+):
+    """
+    Allow individual chunks to be uploaded and later reassembled.
+
+    Parameters
+    ----------
+    file : UploadFile
+        The chunk file to be uploaded.
+    chunk_number : int
+        The number of the chunk being uploaded.
+    total_chunks : int
+        The total number of chunks for the file.
+    chunk_hash : str
+        The MD5 hash of the chunk.
+
+    Returns
+    -------
+    dict
+        A dictionary containing a message and the path where the chunk was saved.
+
+    Raises
+    ------
+    HTTPException
+        If there is an error during the upload process.
+    """
+    if file is None:
+        file = File(...)
+    print(file.filename)
+    print(f"Received chunk {chunk_number} of {total_chunks}")
+    print(f"Received Chunk MD5 Hash:   {chunk_hash}")
+    try:
+        file_content = await file.read()
+        calculated_hash = calculate_md5_checksum(file_content)
+        print(f"Calculated Chunk MD5 Hash: {calculated_hash}")
+        print(f"Hash matches: {calculated_hash == chunk_hash}")
+
+        chunk_path = save_chunk(file_content, chunk_number, file.filename)
+        print(f"Chunk saved at: {chunk_path}")
+
+        if chunk_number == total_chunks - 1:
+            reassemble_file(total_chunks, file.filename)
+        return {"message": "Chunk uploaded successfully", "chunk_path": chunk_path}
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=str(e)) from e
+
+
+@router.post("/single")
+async def upload_single(file: UploadFile = None, file_hash: str = Form(...)):
+    """
+    Upload a single file.
+
+    Parameters
+    ----------
+    file : UploadFile
+        The file to be uploaded.
+    file_hash : str
+        The MD5 hash of the file.
+
+    Returns
+    -------
+    dict
+        A dictionary containing a message and the path where the file was saved.
+
+    Raises
+    ------
+    HTTPException
+        If there is an error during the upload process.
+    """
+    if file is None:
+        file = File(...)
+    print(f"Received File MD5 Hash:   {file_hash}")
+    try:
+        file_content = await file.read()
+        calculated_hash = calculate_md5_checksum(file_content)
+        print(f"Calculated File MD5 Hash: {calculated_hash}")
+        print(f"Hash matches: {calculated_hash == file_hash}")
+
+        file_path = save_file(file)
+        return {"message": "File uploaded successfully", "file_path": file_path}
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=str(e)) from e
+
+
+@router.get("/files")
+async def get_files() -> list[str]:
+    """
+    Get a list of all uploaded files.
+
+    Returns
+    -------
+    list of str
+        A list of filenames of all uploaded files.
+    """
+    return get_all_filenames()

api/utils/upload_helper.py

@@ -0,0 +1,122 @@
+"""Helper functions for uploading files and accessing uploaded files."""
+
+from __future__ import annotations
+
+import hashlib
+import os
+
+DATA_DIR = "/home/ubuntu/janus-api/janus-web/data"
+
+
+def save_chunk(file, chunk_number, original_filename, directory=DATA_DIR):
+    """
+    Save a chunk of a file to the specified directory.
+
+    Parameters
+    ----------
+    file : bytes
+        The content of the chunk to be saved.
+    chunk_number : int
+        The number of the chunk being saved.
+    original_filename : str
+        The original filename of the file being chunked.
+    directory : str, optional
+        The directory where the chunk will be saved (default is DATA_DIR).
+
+    Returns
+    -------
+    str
+        The path where the chunk was saved.
+    """
+    os.makedirs(directory, exist_ok=True)
+    chunk_path = os.path.join(directory, f"{original_filename}_chunk_{chunk_number}")
+    with open(chunk_path, "wb") as chunk_file:
+        chunk_file.write(file)
+    return chunk_path
+
+
+def reassemble_file(total_chunks, original_filename, directory=DATA_DIR):
+    """
+    Reassemble a file from its chunks.
+
+    Parameters
+    ----------
+    total_chunks : int
+        The total number of chunks.
+    original_filename : str
+        The original filename of the file being reassembled.
+    directory : str, optional
+        The directory where the chunks are stored (default is DATA_DIR).
+
+    Returns
+    -------
+    str
+        The path where the reassembled file was saved.
+    """
+    output_path = os.path.join(directory, original_filename)
+    with open(output_path, "wb") as complete_file:
+        for i in range(total_chunks):
+            chunk_path = os.path.join(directory, f"{original_filename}_chunk_{i}")
+            with open(chunk_path, "rb") as chunk_file:
+                complete_file.write(chunk_file.read())
+            os.remove(chunk_path)
+    return output_path
+
+
+def save_file(file, directory=DATA_DIR):
+    """
+    Save a file to the specified directory.
+
+    Parameters
+    ----------
+    file : UploadFile
+        The file to be saved.
+    directory : str, optional
+        The directory where the file will be saved (default is DATA_DIR).
+
+    Returns
+    -------
+    str
+        The path where the file was saved.
+    """
+    os.makedirs(directory, exist_ok=True)
+    file_path = os.path.join(directory, file.filename)
+    with open(file_path, "wb") as buffer:
+        buffer.write(file.file.read())
+    return file_path
+
+
+def calculate_md5_checksum(file_chunk):
+    """
+    Calculate the MD5 checksum of a file chunk.
+
+    Parameters
+    ----------
+    file_chunk : bytes
+        The content of the file chunk.
+
+    Returns
+    -------
+    str
+        The MD5 checksum of the file chunk.
+    """
+    md5 = hashlib.md5()
+    md5.update(file_chunk)
+    return md5.hexdigest()
+
+
+def get_all_filenames() -> list[str]:
+    """
+    Get a list of all filenames in the data directory.
+
+    Returns
+    -------
+    list of str
+        A list of filenames in the data directory.
+    """
+    filenames = [
+        filename
+        for filename in os.listdir(DATA_DIR)
+        if os.path.isfile(os.path.join(DATA_DIR, filename))
+    ]
+    return filenames if filenames else ["No files found"]

main.py

@@ -0,0 +1,17 @@
+"""Main module for the application."""
+
+from __future__ import annotations
+
+from fastapi import FastAPI
+
+from api.endpoints import upload_route
+
+app = FastAPI()
+
+app.include_router(upload_route.router)
+
+
+if __name__ == "__main__":
+    import uvicorn
+
+    uvicorn.run(app, host="0.0.0.0", port=8000)