loremfile

Getting started

How do the URLs work?

Every file lives at https://loremfile.dev/{format}/{name}. The format is the first part of the path. The name tells you what is inside.

pdf/a4-3pages.pdf is three A4 pages. csv/people-1000.csv is a thousand rows of synthetic people.

There is nothing to sign up for and no key to send. Naming and sizes explains how the names are built.

Can I use a file straight from a web page?

Yes. A page on any origin can fetch a file. CORS allows GET and HEAD from every origin. Byte-range requests work too, so a media player can seek.

Query strings are ignored. Adding ?v=2 does not bust a cache, and it does not change what you receive.

If you would rather keep a copy:

curl -O https://loremfile.dev/pdf/a4-3pages.pdf

How do I use it in a GitHub workflow?

Add the action to a step and list the files you want:

- uses: kumarprabhashanand/loremfile/action@action-v1
  with:
    paths: pdf/minimal.pdf mp4/720p-5s.mp4

The files land in loremfile-fixtures/. Each one is checked against its hash in the manifest.

To get every file of a format, add formats: pdf beside paths or in place of it.

The action runs on Linux and macOS runners. It downloads one file at a time and backs off if it is rate limited.

How do I use it from the command line?

The same small client is published for Python and for Node. Both have get, list and verify, with the same flags and the same exit codes.

Neither has a dependency. The Python package declares dependencies = [] and uses only the standard library. The npm package's package.json lists no dependencies and uses only Node's built-ins. Installing either one installs that one package.

# Python 3.11 or newer
pip install loremfile
loremfile get pdf/minimal.pdf
loremfile get --format pdf

# Node 20 or newer, no install step
npx loremfile get pdf/minimal.pdf
npx loremfile get --format pdf

Each file is checked against its hash in the manifest before it is written. A file that does not match is not written at all.

The exit codes are:

Both clients install a command named loremfile. On a machine with both, whichever comes first on PATH runs.

How do I use it in a pytest suite?

Some files are too big to commit: multi-megabyte video, audio and archives. For those, the Python client gives pytest a loremfile fixture. pip install loremfile is the whole setup.

def test_reads_a_real_mp4(loremfile):
    video = loremfile("mp4/10mb.mp4")  # a pathlib.Path, checked against the manifest

The first call downloads the file into .pytest_cache and checks its SHA-256. A file that does not match is never returned.

Published files never change at their path. So after the first call, the suite needs no network for that file. In CI, cache .pytest_cache/d/loremfile between runs.

If your tests may not touch the network, or the file you need is a 4 KB PDF, commit the file instead.

How big is a file before I download it?

curl -sI https://loremfile.dev/mp4/720p-5s.mp4 | grep -i content-length

Content-Length equals the bytes field in the manifest. Do not treat the ETag as an MD5.

How do I check what I downloaded?

sha256sums.txt lists every active file with its SHA-256 hash. It uses the format that sha256sum reads:

# Linux (coreutils)
curl -fsSL https://loremfile.dev/sha256sums.txt | sha256sum -c --ignore-missing
# macOS ships no sha256sum; shasum is preinstalled
curl -fsSL https://loremfile.dev/sha256sums.txt | shasum -a 256 -c --ignore-missing

What if I use another language?

Use it from your language has the same fetch-and-verify in shell, Python, JavaScript, Go and Java. Each one is a complete program. The page also states the contract, so you can write the same thing in any language that is not on that list.

How do I find a file without a browser?

manifest.json lists every file with its size, hash, MIME type and measured properties. Each format also has its own list at /{format}/index.json.

Keep automated traffic under 30 requests per second per IP. The action above paces its downloads for you.