mstransfer

Transfer mass spectrometry files (mzML / MSZ) between machines over HTTP. Files are always transported in compressed MSZ format using mscompress — mzML sources are compressed on-the-fly without temp files, and the receiving end can optionally decompress back to mzML.

Install

# Recommended — installs as an isolated CLI tool
uv tool install mstransfer

# Or with pipx
pipx install mstransfer

# Or as a project dependency
uv pip install mstransfer
pip install mstransfer

Quick start

Start a server on the receiving machine:

mstransfer serve --port 1319 --store-as msz

Upload files from the source machine:

mstransfer upload /data/experiment1.mzML /data/batch/ remote-host:1319

Usage

mstransfer serve

Start the receiver server.

mstransfer serve [--host 0.0.0.0] [--port 1319] [--output-dir ./received] [--store-as msz|mzml]

mstransfer upload

Upload files to a server. Accepts any mix of files and directories. The last positional argument is the target.

mstransfer upload <paths...> <host[:port]> [--recursive] [--parallel 4]

Supported file types: .mzML, .msz, .mszx.

• mzML files are compressed to MSZ on-the-fly via compress_stream() and streamed directly into the HTTP request — no temp files on the sender.
• MSZ / MSZX files are streamed as-is.

Programmatic access

You can use mstransfer as a library to send files from your own Python code:

from pathlib import Path
from mstransfer.client import send_file, send_batch

# Send a single file (Path)
result = send_file(Path("experiment.mzML"), "http://remote-host:1319")

# Send an already-opened mscompress object
from mscompress import MZMLFile, MSZFile
from mscompress.mszx import MSZXFile

mzml = MZMLFile(b"/data/experiment.mzML")
send_file(mzml, "http://remote-host:1319")

msz = MSZFile(b"/data/experiment.msz")
send_file(msz, "http://remote-host:1319")

mszx = MSZXFile.open("/data/experiment.mszx")
send_file(mszx, "http://remote-host:1319")

# Send multiple files in parallel
send_batch(
    [Path("a.mzML"), Path("b.msz"), mszx],
    "http://remote-host:1319",
    parallel=4,
)

send_file and send_batch accept any mix of Path, MZMLFile, MSZFile, and MSZXFile inputs.

API

The server exposes a REST API under /v1/:

Embedding in another app

The server is built as a FastAPI app factory:

from mstransfer import create_app

# Standalone
app = create_app(output_dir="/data/ms", store_as="mzml")

# Mount in an existing FastAPI app
from fastapi import FastAPI
main_app = FastAPI()
main_app.mount("/transfer", create_app())

How it works

flowchart LR
    subgraph Sender
        mzML[".mzML"] --> compress["compress_stream()"]
        msz[".msz / .mszx"] --> read["read file"]
    end

    compress --> upload["POST /v1/upload\n(MSZ bytes)"]
    read --> upload

    subgraph Server
        upload --> mode{store-as?}
        mode -- msz --> write["write to disk"]
        mode -- mzml --> temp["write temp .msz"]
        temp --> decompress["decompress"]
        decompress --> cleanup["cleanup temp"]
    end

Development

uv sync --dev
uv run pytest
uv run ruff check src/ tests/

Pre-commit hooks

Install the pre-commit hooks to automatically run linting, type checking, and tests before each commit:

uv run pre-commit install