tex.beijaflor.io

TeX PDF generation web service.

Problem

A TeX installation is very large and it's nice to separate the role of running tex to a dedicated service.

Usage (CLI)

curl -XPOST -F "file=@./input.tex" > output.pdf

The request to tex.beijaflor.io responds with a redirect to the Amazon S3 location of the file. It's your choice to follow the redirect to display the PDF or to manually read response headers for more information.

We set headers:

• x-sts-pdf - The S3 PDF url
• x-sts-log - The S3 TeX log url
• x-sts-log - The S3 TeX input file url

The files are always at s3://bucket/<MD5 of the Contents> so services plugging into this can generate the PDF URL based on the contents, or regenerate PDFs at no cost by re-POSTing the same contents.

Usage (JavaScript)

Additionally to accepting a file multipart form data parameter with an input file, tex.beijaflor.io accepts a text form parameter with the TeX file as an URLEncoded string.

import fetch from 'isomorphic-fetch';
const latex = '...';
fetch('https://tex.beijaflor.io', {
  method: 'post',
  headers: {
    'content-type': 'application/x-www-form-urlencoded',
  },
  mode: 'cors',
  body: querystring.stringify({ text: latex }),
}).then(res => {
  const s3Location = res.url;
  // ... show PDF in iframe or show link for the users to download the result
});

Installation

Distributed as a docker image with tex installed plus the haskell server binary.

docker pull beijaflorio/tex.beijaflor.io
docker run -it \
  -p 80 \
  -e AWS_ACCESS_KEY_ID=... \
  -e AWS_SECRET_ACCESS_KEY=... \
  -e AWS_DEFAULT_REGION=... \
  # The name of the bucket to use to serve and cache PDF and output files
  -e STS_BUCKET_NAME=... \
  # The public URL for the service
  -e STS_HOST=... \
  # The public URL for the S3 bucket (say you want to support https & CORS)
  -e STS_S3_PUBLIC_URL=... \
  beijaflorio/tex.beijaflor.io

Other

• https://github.com/tectonic-typesetting/tectonic/