chunk upload

app/api/middleware.py

@@ -15,7 +15,7 @@ from datetime import datetime, timedelta
 def attach_headers(response):
     response.headers["Access-Control-Allow-Origin"] = "*"
     response.headers["Access-Control-Allow-Methods"] = "GET, POST, OPTIONS"
-    response.headers["Access-Control-Allow-Headers"] = "Origin, Content-Type, Accept, Authorization, Referer, User-Agent, Sec-Fetch-Dest, Sec-Fetch-Mode, Sec-Fetch-Site, x-file-name, x-content-sha256"
+    response.headers["Access-Control-Allow-Headers"] = "Origin, Content-Type, Accept, Authorization, Referer, User-Agent, Sec-Fetch-Dest, Sec-Fetch-Mode, Sec-Fetch-Site, x-file-name, x-content-sha256, x-chunk-start, x-upload-id"
     response.headers["Access-Control-Allow-Credentials"] = "true"
     return response
 

app/api/routes/progressive_storage.py

@@ -4,10 +4,10 @@ import asyncio
 from uuid import uuid4
 from datetime import datetime
 from mimetypes import guess_type
+from base64 import b64decode
 
 import aiofiles
 from base58 import b58encode
-from base64 import b64decode
 from sanic import response
 
 from app.core.logger import make_log
@@ -18,8 +18,8 @@ from app.core._utils.resolve_content import resolve_content
 
 # POST /api/v1.5/storage
 async def s_api_v1_5_storage_post(request):
-    # Log the start of the file upload process
+    # Log the receipt of a chunk upload request
-    make_log("uploader_v1.5", "Received file upload request", level="INFO")
+    make_log("uploader_v1.5", "Received chunk upload request", level="INFO")
     
     # Get the provided file hash from header (hex format)
     provided_hash_hex = request.headers.get("X-Content-SHA256")
@@ -34,38 +34,80 @@ async def s_api_v1_5_storage_post(request):
         make_log("uploader_v1.5", f"Invalid X-Content-SHA256 header format: {e}", level="ERROR")
         return response.json({"error": "Invalid X-Content-SHA256 header format"}, status=400)
     
-    provided_filename = request.headers.get("X-File-Name")
+    # Get the provided file name from header and decode it from base64
-    provided_filename = b64decode(provided_filename).decode("utf-8")
+    provided_filename_b64 = request.headers.get("X-File-Name")
-
+    if not provided_filename_b64:
-    # Check if the file already exists in the database
+        make_log("uploader_v1.5", "Missing X-File-Name header", level="ERROR")
-    db_session = request.ctx.db_session
+        return response.json({"error": "Missing X-File-Name header"}, status=400)
-    existing = db_session.query(StoredContent).filter_by(hash=provided_hash_b58).first()
-    if existing:
-        make_log("uploader_v1.5", f"File with hash {provided_hash_b58} already exists in DB", level="INFO")
-        serialized_v2 = existing.serialize_v2()  # returns a string
-        serialized_v1 = existing.serialize_v1()  # returns a string
-        return response.json({
-            "content_sha256": provided_hash_b58,
-            "content_id": serialized_v2,
-            "content_id_v1": serialized_v1,
-            "content_url": f"dmy://storage?cid={serialized_v2}",
-        })
-
-    # Save uploaded file to a temporary location using streaming
-    temp_filename = f"v1.5_upload_{uuid4()}"
-    temp_path = os.path.join(UPLOADS_DIR, temp_filename)
-    make_log("uploader_v1.5", f"Saving file to temporary location: {temp_path}", level="INFO")
     try:
-        async with aiofiles.open(temp_path, 'wb') as out_file:
+        provided_filename = b64decode(provided_filename_b64).decode("utf-8")
+    except Exception as e:
+        make_log("uploader_v1.5", f"Invalid X-File-Name header: {e}", level="ERROR")
+        return response.json({"error": "Invalid X-File-Name header"}, status=400)
+    
+    # Get X-Chunk-Start header (must be provided) and parse it as integer
+    chunk_start_header = request.headers.get("X-Chunk-Start")
+    if chunk_start_header is None:
+        make_log("uploader_v1.5", "Missing X-Chunk-Start header", level="ERROR")
+        return response.json({"error": "Missing X-Chunk-Start header"}, status=400)
+    try:
+        chunk_start = int(chunk_start_header)
+    except Exception as e:
+        make_log("uploader_v1.5", f"Invalid X-Chunk-Start header: {e}", level="ERROR")
+        return response.json({"error": "Invalid X-Chunk-Start header"}, status=400)
+    
+    # Enforce maximum chunk size (80 MB) using Content-Length header if provided
+    max_chunk_size = 80 * 1024 * 1024  # 80 MB
+    content_length = request.headers.get("Content-Length")
+    if content_length is not None:
+        try:
+            content_length = int(content_length)
+            if content_length > max_chunk_size:
+                make_log("uploader_v1.5", f"Chunk size {content_length} exceeds maximum allowed", level="ERROR")
+                return response.json({"error": "Chunk size exceeds maximum allowed (80 MB)"}, status=400)
+        except:
+            pass
+
+    # Determine if this is a new upload or a continuation (resume)
+    upload_id = request.headers.get("X-Upload-ID")
+    is_new_upload = False
+    if not upload_id:
+        # New upload session: generate a new uuid
+        upload_id = str(uuid4())
+        is_new_upload = True
+        make_log("uploader_v1.5", f"Starting new upload session with ID: {upload_id}", level="INFO")
+    else:
+        make_log("uploader_v1.5", f"Resuming upload session with ID: {upload_id}", level="INFO")
+    
+    # Determine the temporary file path based on upload_id
+    temp_path = os.path.join(UPLOADS_DIR, f"temp_{upload_id}")
+    
+    # Check current size of the temporary file (if it exists)
+    current_size = 0
+    if os.path.exists(temp_path):
+        current_size = os.path.getsize(temp_path)
+    
+    # If the provided chunk_start is less than current_size, the chunk is already received
+    if chunk_start < current_size:
+        make_log("uploader_v1.5", f"Chunk starting at {chunk_start} already received, current size: {current_size}", level="INFO")
+        return response.json({"upload_id": upload_id, "current_size": current_size})
+    elif chunk_start > current_size:
+        make_log("uploader_v1.5", f"Chunk start {chunk_start} does not match current file size {current_size}", level="ERROR")
+        return response.json({"error": "Chunk start does not match current file size"}, status=400)
+    
+    # Append the received chunk to the temporary file
+    try:
+        mode = 'wb' if is_new_upload else 'ab'
+        async with aiofiles.open(temp_path, mode) as out_file:
             async for chunk in request.stream:
                 await out_file.write(chunk)
-        make_log("uploader_v1.5", f"Finished saving file to temporary location: {temp_path}", level="INFO")
+        new_size = os.path.getsize(temp_path)
+        make_log("uploader_v1.5", f"Appended chunk. New file size: {new_size}", level="INFO")
     except Exception as e:
-        make_log("uploader_v1.5", f"Error saving uploaded file: {e}", level="ERROR")
+        make_log("uploader_v1.5", f"Error saving chunk: {e}", level="ERROR")
-        return response.json({"error": "Failed to save uploaded file"}, status=500)
+        return response.json({"error": "Failed to save chunk"}, status=500)
     
-    # Compute file SHA256 using subprocess to avoid loading the file into memory
+    # Compute the SHA256 hash of the temporary file using subprocess
-    make_log("uploader_v1.5", f"Computing file hash using subprocess for file: {temp_path}", level="INFO")
     try:
         proc = await asyncio.create_subprocess_exec(
             'sha256sum', temp_path,
@@ -77,7 +119,6 @@ async def s_api_v1_5_storage_post(request):
             error_msg = stderr.decode().strip()
             make_log("uploader_v1.5", f"sha256sum error: {error_msg}", level="ERROR")
             return response.json({"error": "Failed to compute file hash"}, status=500)
-        # Parse output: "<hash>  <filename>"
         computed_hash_hex = stdout.decode().split()[0].strip()
         computed_hash_bytes = bytes.fromhex(computed_hash_hex)
         computed_hash_b58 = b58encode(computed_hash_bytes).decode()
@@ -86,49 +127,35 @@ async def s_api_v1_5_storage_post(request):
         make_log("uploader_v1.5", f"Error computing file hash: {e}", level="ERROR")
         return response.json({"error": "Error computing file hash"}, status=500)
     
-    # Verify that the computed hash matches the provided hash
+    # If computed hash matches the provided one, the final chunk has been received
-    if computed_hash_b58 != provided_hash_b58:
+    if computed_hash_b58 == provided_hash_b58:
-        make_log("uploader_v1.5", f"Hash mismatch: provided {provided_hash_b58} vs computed {computed_hash_b58}", level="ERROR")
-        try:
-            os.remove(temp_path)
-            make_log("uploader_v1.5", f"Temporary file removed due to hash mismatch: {temp_path}", level="INFO")
-        except Exception as e:
-            make_log("uploader_v1.5", f"Error removing temp file: {e}", level="ERROR")
-        return response.json({"error": "Hash mismatch"}, status=400)
-
-    # Determine the final file path
         final_path = os.path.join(UPLOADS_DIR, f"v1.5_{computed_hash_b58}")
-    if os.path.exists(final_path):
-        make_log("uploader_v1.5", f"File already exists on disk: {final_path}", level="INFO")
         try:
-            os.remove(temp_path)
+            os.rename(temp_path, final_path)
-            make_log("uploader_v1.5", f"Temporary file removed: {temp_path}", level="INFO")
+            make_log("uploader_v1.5", f"Final chunk received. File renamed to: {final_path}", level="INFO")
         except Exception as e:
-            make_log("uploader_v1.5", f"Error removing temp file: {e}", level="ERROR")
+            make_log("uploader_v1.5", f"Error renaming file: {e}", level="ERROR")
+            return response.json({"error": "Failed to finalize file storage"}, status=500)
+        
+        db_session = request.ctx.db_session
         existing = db_session.query(StoredContent).filter_by(hash=computed_hash_b58).first()
         if existing:
+            make_log("uploader_v1.5", f"File with hash {computed_hash_b58} already exists in DB", level="INFO")
             serialized_v2 = existing.serialize_v2()
             serialized_v1 = existing.serialize_v1()
             return response.json({
+                "upload_id": upload_id,
                 "content_sha256": computed_hash_b58,
                 "content_id": serialized_v2,
                 "content_id_v1": serialized_v1,
                 "content_url": f"dmy://storage?cid={serialized_v2}",
             })
-    else:
-        try:
-            os.rename(temp_path, final_path)
-            make_log("uploader_v1.5", f"Renamed temporary file to final location: {final_path}", level="INFO")
-        except Exception as e:
-            make_log("uploader_v1.5", f"Error renaming file: {e}", level="ERROR")
-            return response.json({"error": "Failed to finalize file storage"}, status=500)
         
-    # Create a new StoredContent record with user_id from request.ctx.user and commit to DB
         try:
             new_content = StoredContent(
                 type='local/content_bin',
                 hash=computed_hash_b58,
-            user_id=request.ctx.user.id,  # 'user_id' is added to StoredContent
+                user_id=request.ctx.user.id,
                 filename=provided_filename,
                 key_id=None,
                 meta={},
@@ -144,11 +171,15 @@ async def s_api_v1_5_storage_post(request):
         serialized_v2 = new_content.serialize_v2()
         serialized_v1 = new_content.serialize_v1()
         return response.json({
+            "upload_id": upload_id,
             "content_sha256": computed_hash_b58,
             "content_id": serialized_v2,
             "content_id_v1": serialized_v1,
             "content_url": f"dmy://storage?cid={serialized_v2}",
         })
+    else:
+        # Not the final chunk yet – return current upload status
+        return response.json({"upload_id": upload_id, "current_size": os.path.getsize(temp_path)})
 
 
 # GET /api/v1.5/storage/<file_hash>
@@ -156,22 +187,18 @@ async def s_api_v1_5_storage_get(request, file_hash):
     # Log the file retrieval request
     make_log("uploader_v1.5", f"Received file retrieval request for hash: {file_hash}", level="INFO")
     
-    # Determine the file path based on the provided file_hash
     final_path = os.path.join(UPLOADS_DIR, f"v1.5_{file_hash}")
     if not os.path.exists(final_path):
         make_log("uploader_v1.5", f"File not found: {final_path}", level="ERROR")
         return response.json({"error": "File not found"}, status=404)
     
-    # Retrieve the StoredContent record from the database
     db_session = request.ctx.db_session
     stored = db_session.query(StoredContent).filter_by(hash=file_hash).first()
     if stored and stored.filename:
         filename_for_mime = stored.filename
     else:
-        # If the record is not found or filename is not set, fallback to the file path
         filename_for_mime = final_path
     
-    # Determine MIME type using filename from StoredContent
     mime_type, _ = guess_type(filename_for_mime)
     if not mime_type:
         mime_type = "application/octet-stream"
@@ -186,7 +213,6 @@ async def s_api_v1_5_storage_get(request, file_hash):
             make_log("uploader_v1.5", f"Invalid Range header: {range_header}", level="ERROR")
             return response.json({"error": "Invalid Range header"}, status=400)
         range_spec = range_spec[len("bytes="):]
-        # Split by comma to handle multiple ranges
         range_parts = [part.strip() for part in range_spec.split(',')]
         parsed_ranges = []
         try:
@@ -195,7 +221,6 @@ async def s_api_v1_5_storage_get(request, file_hash):
                     raise ValueError("Invalid range format")
                 start_str, end_str = part.split('-', 1)
                 if start_str == "":
-                    # Suffix byte range: last N bytes
                     suffix_length = int(end_str)
                     if suffix_length > file_size:
                         start = 0
@@ -215,7 +240,6 @@ async def s_api_v1_5_storage_get(request, file_hash):
             make_log("uploader_v1.5", f"Invalid Range header: {range_header} - {e}", level="ERROR")
             return response.json({"error": "Invalid Range header"}, status=400)
         
-        # If only one range is requested, use single range response
         if len(parsed_ranges) == 1:
             start, end = parsed_ranges[0]
             content_length = end - start + 1
@@ -226,12 +250,11 @@ async def s_api_v1_5_storage_get(request, file_hash):
                 "Content-Type": mime_type,
             }
             async def stream_file_range():
-                # Stream single range content
                 make_log("uploader_v1.5", f"Starting to stream file from byte {start} to {end}", level="INFO")
                 async with aiofiles.open(final_path, mode='rb') as f:
                     await f.seek(start)
                     remaining = content_length
-                    chunk_size = 1024 * 1024  # 1MB chunks
+                    chunk_size = 1024 * 1024
                     while remaining > 0:
                         read_size = min(chunk_size, remaining)
                         data = await f.read(read_size)
@@ -242,14 +265,12 @@ async def s_api_v1_5_storage_get(request, file_hash):
                 make_log("uploader_v1.5", f"Finished streaming file: {final_path}", level="INFO")
             return response.stream(stream_file_range, status=206, headers=headers)
         else:
-            # Multiple ranges requested: create a multipart/byteranges response
+            boundary = uuid4().hex
-            boundary = uuid4().hex  # Generate a random boundary string
             headers = {
                 "Content-Type": f"multipart/byteranges; boundary={boundary}",
                 "Accept-Ranges": "bytes",
             }
             async def stream_multipart():
-                # For each range, yield the boundary, part headers, and the file content
                 for start, end in parsed_ranges:
                     part_header = (
                         f"--{boundary}\r\n"
@@ -262,7 +283,7 @@ async def s_api_v1_5_storage_get(request, file_hash):
                     async with aiofiles.open(final_path, mode='rb') as f:
                         await f.seek(start)
                         remaining = part_length
-                        chunk_size = 1024 * 1024  # 1MB chunks
+                        chunk_size = 1024 * 1024
                         while remaining > 0:
                             read_size = min(chunk_size, remaining)
                             data = await f.read(read_size)
@@ -271,10 +292,8 @@ async def s_api_v1_5_storage_get(request, file_hash):
                             remaining -= len(data)
                             yield data
                     yield b"\r\n"
-                # Final boundary marker
                 yield f"--{boundary}--\r\n".encode()
             return response.stream(stream_multipart, status=206, headers=headers)
     else:
-        # No Range header: return full file
         make_log("uploader_v1.5", f"Returning full file for video/audio: {final_path}", level="INFO")
         return await response.file(final_path, mime_type=mime_type)

uploader_test.html

@@ -43,7 +43,7 @@
     // Base URL for endpoints
     const BASE_URL = "https://my-public-node-1.projscale.dev/api/v1.5/storage";
     
-    // Function to append messages to log div
+    // Append log message to log div
     function appendLog(message) {
       const logDiv = document.getElementById('log');
       const p = document.createElement('p');
@@ -51,7 +51,7 @@
       logDiv.appendChild(p);
     }
     
-    // Function to compute SHA-256 hash of a file in hex using jsSHA library with incremental reading
+    // Compute SHA-256 hash of a file using jsSHA library (incremental reading)
     function computeSHA256(file) {
       return new Promise((resolve, reject) => {
         const chunkSize = 2097152; // 2MB per chunk
@@ -60,7 +60,7 @@
         const shaObj = new jsSHA("SHA-256", "ARRAYBUFFER");
   
         reader.onload = function(e) {
-          // Update the hash object with the current chunk data
+          // Update hash with current chunk data
           shaObj.update(e.target.result);
           offset += chunkSize;
           appendLog(`Processed ${Math.min(offset, file.size)} из ${file.size} байт`);
@@ -89,6 +89,70 @@
       });
     }
     
+    // Upload file in chunks (max 80 MB per chunk)
+    async function uploadFileInChunks(file) {
+      const maxChunkSize = 80 * 1024 * 1024; // 80 MB
+      let offset = 0;
+      let uploadId = null;
+      try {
+        appendLog("Starting hash computation...");
+        const hashHex = await computeSHA256(file);
+        appendLog(`Computed SHA-256 hash: ${hashHex}`);
+        
+        while (offset < file.size) {
+          const chunk = file.slice(offset, Math.min(offset + maxChunkSize, file.size));
+          appendLog(`Uploading chunk starting at byte ${offset}`);
+          
+          // Prepare headers for the chunk upload
+          const headers = {
+            "X-Content-SHA256": hashHex,
+            "X-File-Name": btoa(unescape(encodeURIComponent(file.name))), // File name in base64
+            "X-Chunk-Start": offset.toString(),
+            "Content-Type": file.type || "application/octet-stream"
+          };
+          if (uploadId) {
+            headers["X-Upload-ID"] = uploadId;
+          }
+          
+          const response = await fetch(BASE_URL, {
+            method: "POST",
+            headers: headers,
+            body: chunk
+          });
+          
+          if (!response.ok) {
+            const errorData = await response.json();
+            appendLog(`Chunk upload failed: ${errorData.error}`);
+            throw new Error(`Upload failed at offset ${offset}: ${errorData.error}`);
+          }
+          
+          const resultData = await response.json();
+          // Save uploadId from first response if not set
+          if (!uploadId && resultData.upload_id) {
+            uploadId = resultData.upload_id;
+          }
+          
+          // If final response contains content_id, upload is complete
+          if (resultData.content_id) {
+            appendLog(`Upload complete. File ID: ${resultData.content_id}`);
+            return resultData;
+          }
+          
+          // Update offset based on server-reported current size
+          if (resultData.current_size !== undefined) {
+            offset = resultData.current_size;
+            appendLog(`Server reports current_size: ${offset}`);
+          } else {
+            appendLog("Unexpected response from server, missing current_size.");
+            throw new Error("Missing current_size in response");
+          }
+        }
+      } catch (err) {
+        appendLog(`Error during upload: ${err}`);
+        throw err;
+      }
+    }
+    
     // Upload button event listener
     document.getElementById('uploadBtn').addEventListener('click', async () => {
       const fileInput = document.getElementById('uploadFile');
@@ -101,42 +165,15 @@
       }
       
       const file = fileInput.files[0];
-      appendLog("Starting hash computation...");
       try {
-        // Compute SHA-256 hash of the file in hex format
+        const resultData = await uploadFileInChunks(file);
-        const hashHex = await computeSHA256(file);
-        appendLog(`Computed SHA-256 hash: ${hashHex}`);
-  
-              // Prepare the POST request with file as body and additional header for filename
-        const encodedFileName = btoa(unescape(encodeURIComponent(file.name)));
-        const response = await fetch(BASE_URL, {
-          method: "POST",
-          headers: {
-            "X-Content-SHA256": hashHex,
-            "X-File-Name": encodedFileName, // NEW: передаём имя файла в base64
-            "Content-Type": file.type || "application/octet-stream"
-          },
-          body: file
-        });
-
-  
-        if (!response.ok) {
-          const errorData = await response.json();
-          uploadResult.textContent = `Ошибка: ${errorData.error}`;
-          appendLog(`Upload failed: ${errorData.error}`);
-          return;
-        }
-  
-        const resultData = await response.json();
         uploadResult.textContent = `Файл загружен успешно. content_sha256: ${resultData.content_sha256}`;
-        appendLog(`Upload successful. Response: ${JSON.stringify(resultData)}`);
       } catch (err) {
         uploadResult.textContent = "Ошибка при загрузке файла.";
-        appendLog(`Error during upload: ${err}`);
       }
     });
     
-    // Load file button event listener for streaming
+    // Load file for streaming (remains unchanged)
     document.getElementById('loadFileBtn').addEventListener('click', async () => {
       const fileHash = document.getElementById('fileHashInput').value.trim();
       const mediaContainer = document.getElementById('mediaContainer');
@@ -147,12 +184,10 @@
         return;
       }
       
-      // Construct file URL
       const fileUrl = `${BASE_URL}/${fileHash}`;
       appendLog(`Fetching file info for hash: ${fileHash}`);
   
       try {
-        // Perform a HEAD request to determine Content-Type
         const headResponse = await fetch(fileUrl, { method: "HEAD" });
         if (!headResponse.ok) {
           mediaContainer.textContent = "Файл не найден.";
@@ -164,7 +199,6 @@
         appendLog(`Content-Type: ${contentType}`);
   
         let mediaElement;
-        // Create appropriate element based on Content-Type
         if (contentType.startsWith("image/")) {
           mediaElement = document.createElement("img");
           mediaElement.style.maxWidth = "100%";
@@ -176,12 +210,10 @@
           mediaElement = document.createElement("audio");
           mediaElement.controls = true;
         } else {
-          // For other types, create a download link
           mediaElement = document.createElement("a");
           mediaElement.textContent = "Скачать файл";
         }
   
-        // Set the src or href attribute to stream the file
         if (mediaElement.tagName === "A") {
           mediaElement.href = fileUrl;
           mediaElement.download = "";