From 3334f8b00fa1b400354f212b12c4e6bd546ce5f1 Mon Sep 17 00:00:00 2001
From: dylan <dylan@bitmind.ai>
Date: Mon, 24 Aug 2026 15:36:19 -0500
Subject: [PATCH] feat(cli): stream model uploads with progress

---
 gas/protocol/miner_requests.py       | 111 +++++++++++++++++++++++----
 tests/test_miner_upload_streaming.py |  46 +++++++++++
 2 files changed, 141 insertions(+), 16 deletions(-)
 create mode 100644 tests/test_miner_upload_streaming.py

diff --git a/gas/protocol/miner_requests.py b/gas/protocol/miner_requests.py
index e6058d1b..dd9d3ab8 100644
--- a/gas/protocol/miner_requests.py
+++ b/gas/protocol/miner_requests.py
@@ -2,8 +2,9 @@
 import json
 import os
 import re
+import time
 from pathlib import Path
-from typing import Dict, Optional
+from typing import BinaryIO, Dict, Optional
 
 import bittensor as bt
 import httpx
@@ -18,6 +19,73 @@ def calculate_sha256(data: bytes) -> str:
     return hashlib.sha256(data).hexdigest()
 
 
+def calculate_file_sha256(file_path: Path, chunk_size: int = 1024 * 1024) -> str:
+    """Hash a file incrementally without loading it into memory."""
+    digest = hashlib.sha256()
+    with file_path.open("rb") as file:
+        while chunk := file.read(chunk_size):
+            digest.update(chunk)
+    return digest.hexdigest()
+
+
+class UploadProgressReader:
+    """File-like request body that reports upload progress and throughput."""
+
+    def __init__(self, file_path: Path, report_interval: float = 0.5):
+        self._file: BinaryIO = file_path.open("rb")
+        self._total = file_path.stat().st_size
+        self._uploaded = 0
+        self._started_at = time.monotonic()
+        self._last_report = 0.0
+        self._report_interval = report_interval
+
+    def __len__(self) -> int:
+        return self._total
+
+    def read(self, size: int = -1) -> bytes:
+        chunk = self._file.read(size)
+        self._uploaded += len(chunk)
+        now = time.monotonic()
+        if (
+            self._uploaded == self._total
+            or now - self._last_report >= self._report_interval
+        ):
+            self._report(now)
+        return chunk
+
+    def tell(self) -> int:
+        return self._file.tell()
+
+    def seek(self, offset: int, whence: int = os.SEEK_SET) -> int:
+        position = self._file.seek(offset, whence)
+        self._uploaded = position
+        return position
+
+    def close(self) -> None:
+        self._file.close()
+
+    def __enter__(self) -> "UploadProgressReader":
+        return self
+
+    def __exit__(self, *_args) -> None:
+        self.close()
+
+    def _report(self, now: float) -> None:
+        elapsed = max(now - self._started_at, 1e-9)
+        rate = self._uploaded / elapsed
+        percent = 100.0 if not self._total else self._uploaded / self._total * 100
+        remaining = max(self._total - self._uploaded, 0)
+        eta = remaining / rate if rate else 0.0
+        print(
+            f"\r      {percent:5.1f}% "
+            f"({self._uploaded / 1024 / 1024:.1f}/{self._total / 1024 / 1024:.1f} MB) "
+            f"{rate / 1024 / 1024:.1f} MB/s ETA {eta:.0f}s",
+            end="",
+            flush=True,
+        )
+        self._last_report = now
+
+
 def generate_presigned_url(
     wallet: bt.Wallet, 
     upload_endpoint: str, 
@@ -76,15 +144,24 @@ def generate_presigned_url(
         }
 
 
-def upload_to_r2(presigned_url: str, file_content: bytes, content_type: str = 'application/octet-stream') -> dict:
-    """Upload file directly to R2 using presigned URL."""
+def upload_to_r2(
+    presigned_url: str,
+    file_path: Path,
+    content_type: str = 'application/octet-stream',
+) -> dict:
+    """Stream a file directly to R2 using a presigned URL."""
     try:
-        response = requests.put(
-            presigned_url,
-            data=file_content,
-            headers={'Content-Type': content_type},
-            timeout=300  # 5 minutes for large files
-        )
+        with UploadProgressReader(file_path) as upload:
+            response = requests.put(
+                presigned_url,
+                data=upload,
+                headers={
+                    'Content-Type': content_type,
+                    'Content-Length': str(len(upload)),
+                },
+                timeout=300,  # 5 minutes for large files
+            )
+        print()
         
         error_detail = None
         if response.status_code != 200:
@@ -103,6 +180,7 @@ def upload_to_r2(presigned_url: str, file_content: bytes, content_type: str = 'a
         }
         
     except requests.exceptions.RequestException as e:
+        print()
         return {
             "status_code": 0,
             "success": False,
@@ -164,11 +242,8 @@ def upload_single_modality(
     if not file_path_obj.exists():
         raise FileNotFoundError(f"File not found: {file_path}")
 
-    with open(file_path_obj, 'rb') as f:
-        file_content = f.read()
-
-    file_hash = calculate_sha256(file_content)
-    file_size = len(file_content)
+    file_hash = calculate_file_sha256(file_path_obj)
+    file_size = file_path_obj.stat().st_size
     filename = file_path_obj.name
 
     print(f"  File: {filename} ({file_size / 1024 / 1024:.2f} MB)")
@@ -222,7 +297,11 @@ def extract_error(result: dict) -> str:
     submissions_max = presigned_data.get('submissions_max')
 
     print(f"  [2/3] Uploading to R2...", end=' ', flush=True)
-    upload_result = upload_to_r2(presigned_url, file_content, 'application/octet-stream')
+    upload_result = upload_to_r2(
+        presigned_url,
+        file_path_obj,
+        'application/octet-stream',
+    )
 
     if not upload_result['success']:
         print("FAILED")
@@ -369,4 +448,4 @@ def fetch_generator_performance(
             "error": f"API error {resp.status_code}: {resp.text}",
         }
 
-    return {"success": True, "data": resp.json()}
\ No newline at end of file
+    return {"success": True, "data": resp.json()}
diff --git a/tests/test_miner_upload_streaming.py b/tests/test_miner_upload_streaming.py
new file mode 100644
index 00000000..7ea8c47e
--- /dev/null
+++ b/tests/test_miner_upload_streaming.py
@@ -0,0 +1,46 @@
+"""Behavioral tests for discriminator model uploads."""
+
+import hashlib
+from types import SimpleNamespace
+
+from gas.protocol import miner_requests
+
+
+def test_file_hash_is_computed_incrementally(tmp_path):
+    model = tmp_path / "model.zip"
+    payload = b"streamed-model-data" * 100
+    model.write_bytes(payload)
+
+    assert miner_requests.calculate_file_sha256(
+        model,
+        chunk_size=7,
+    ) == hashlib.sha256(payload).hexdigest()
+
+
+def test_r2_upload_streams_file_and_reports_progress(tmp_path, monkeypatch, capsys):
+    model = tmp_path / "model.zip"
+    payload = b"0123456789" * 100
+    model.write_bytes(payload)
+    captured = {}
+
+    def consume_upload(url, data, headers, timeout):
+        captured["url"] = url
+        captured["body_type"] = type(data)
+        captured["headers"] = headers
+        captured["timeout"] = timeout
+        chunks = []
+        while chunk := data.read(17):
+            chunks.append(chunk)
+        captured["payload"] = b"".join(chunks)
+        return SimpleNamespace(status_code=200, headers={"ETag": "etag"}, text="")
+
+    monkeypatch.setattr(miner_requests.requests, "put", consume_upload)
+
+    result = miner_requests.upload_to_r2("https://r2.example/upload", model)
+
+    assert result["success"] is True
+    assert captured["payload"] == payload
+    assert captured["body_type"] is miner_requests.UploadProgressReader
+    assert captured["headers"]["Content-Length"] == str(len(payload))
+    assert captured["timeout"] == 300
+    assert "100.0%" in capsys.readouterr().out