How to generate pdf with typst syntax using python?

Questions
1

Hello,
I have a typst file on the webapp, it uses a csv file as an argument:

#import "ABC_lib.typ": *
#show: ABC_format

#import "@preview/cetz:0.3.2": canvas, draw
#import "@preview/cetz-plot:0.1.1": plot




// Load CSV and separate first row from others
#let (names, ..data) = csv("ABC_data.csv",
row-type: array)

//here the rest of the typst code

I want to generate the same PDF file for different Csv files, my idea is to use python to :

  1. loop over the csv files in python
  2. feed the template Csv_i
  3. output a temporary typst file_i
  4. use subprocess and convert it to PDF_i using typst compile.
  5. delete the temporary file_i
  6. repeat

I know how to do this in latex, but not in typst since it uses scripting, so it this optimal ? can it be done another way ?

requirements:
I need to do it locally since I have thousands of CSV files.
if you noticed i am importing a ABC_lib.typ which contains all the formatting and rules i usually use to generate pdfs, how do I include that file in the python process?

2

With Typst you can skip step 3 and 5 with the temporary file and use in step 4 with the Typst compile command --input key=value to provide the csv file name (and read the data with csv(...)) or provide the csv data directly to your Typst file. Inside the Typst file you can access the argument with sys.inputs.yourkey.

Edit: The full command you would use in step 4 with the Python subprocess is:

typst c main.typ --input yourkey=somefilename.csv

With the Typst compile command, #import ABC_lib.typ ... would be automatically used, no additional actions needed.

1 Like
3

Hi, thank you for your response,
In the typst file how do I reference the csv file to take the “yourkey=somefilename.csv” as an argument fed to him by the subprocess ?

4

I think you’ll be interested in the other questions tagged data-loading (in fact, I will add that tag to your question after I finish writing this :wink:). Particularly this one, which sounds very close to what you’re trying to do:

2 Likes
5

Yes, that helped me indeed, but I have hit a road block:
As I understand it, i must either feed typst a dictionary of all csv files or one csv file at a time.
The optimal way as I see it, is to loop over the files in python, feed them to typst, generate the pdf, but now I have multiple pdf files, I need to merge them into one report :/
Is there a way around this?

6

You can pass all files as a JSON-object, for example. If you have have a folder csvs/, you can load all of them in python as a json string:

import subprocess
import glob
import os
import json

folder_path = "csvs"

files = glob.glob(
    os.path.join(folder_path, "*.csv")
)

json_files = json.dumps(files)

subprocess.run([
    "typst",
    "c",
    "main.typ",
    "--input",
    f"files={json_files}"
], capture_output=True, text=True)

and then read it in typst:

#let files = json.decode(sys.inputs.files)

#files


#for f in files {
  let data = csv(f)
  // do what you want with the files
}
2 Likes