Access files in a volume without spinning up a sandbox. This is useful for browsing data, downloading results, or uploading input files.
List Files
List files and directories in a volume:
from moru import Volume
vol = Volume.get("my-workspace")
# List root directory
files = vol.list_files("/")
for f in files:
if f.type == "directory":
print(f"[DIR] {f.name}")
else:
print(f"[FILE] {f.name} ({f.size} bytes)")
Download Files
Download file content from a volume:
from moru import Volume
vol = Volume.get("my-workspace")
# Download as bytes
content = vol.download("/README.md")
print(content.decode())
# Save to local file
with open("local_readme.md", "wb") as f:
f.write(vol.download("/README.md"))
Upload Files
Upload files to a volume:
from moru import Volume
vol = Volume.get("my-workspace")
# Upload from bytes
vol.upload("/data/input.txt", b"Hello, World!")
# Upload from file
with open("local_file.txt", "rb") as f:
vol.upload("/data/input.txt", f.read())
Upload creates parent directories automatically. You don’t need to create /data/ before uploading to /data/input.txt.
Delete Files
Delete files or directories from a volume:
from moru import Volume
vol = Volume.get("my-workspace")
# Delete a file
vol.delete("/data/old_file.txt")
# Delete a directory recursively
vol.delete("/temp", recursive=True)
File Info
File entries include:
| Property | Description |
|---|
name | File or directory name |
path | Full path within volume |
type | "file" or "directory" |
size | Size in bytes (files only) |
modified_at / modifiedAt | Last modification time |
from moru import Volume
vol = Volume.get("my-workspace")
# Get first page
files, next_token = vol.list_files("/", limit=100)
# Get next page
if next_token:
more_files, next_token = vol.list_files("/", limit=100, next_token=next_token)
Working with Binary Files
Upload and download binary files like images, archives, and executables:
from moru import Volume
vol = Volume.get("my-workspace")
# Upload binary file
with open("model.bin", "rb") as f:
vol.upload("/models/model.bin", f.read())
# Download and verify
content = vol.download("/models/model.bin")
with open("downloaded_model.bin", "wb") as f:
f.write(content)
Common Patterns
Pre-populate Volume Before Sandbox
from moru import Volume, Sandbox
# Create volume and upload data
vol = Volume.create(name="agent-input")
vol.upload("/config.json", b'{"model": "gpt-4"}')
vol.upload("/prompt.txt", b"You are a helpful assistant...")
# Create sandbox with pre-populated volume
sbx = Sandbox.create(
volume_id=vol.volume_id,
volume_mount_path="/input"
)
# Config is ready to use
sbx.commands.run("cat /input/config.json")
Download Results After Sandbox
from moru import Volume, Sandbox
vol = Volume.create(name="agent-output")
sbx = Sandbox.create(
volume_id=vol.volume_id,
volume_mount_path="/output"
)
# Agent writes results
sbx.commands.run("python generate_report.py > /output/report.txt")
sbx.kill()
# Download results without sandbox
report = vol.download("/output/report.txt")
print(report.decode())
Next Steps
