Is there a tool to fill PDF forms non-interactively from the shell?
For my use case, just setting the text for each field (in whatever order the fields are defined) and saving the PDF file would be enough, but to keep the question more general, any form interaction via CLI is appreciated (tick boxes, select radio buttons, fill text, click buttons, etc.).
I know that poppler has some functions to access PDF forms, but I couldn't find a CLI utility that makes use of these, yet.
Yes, pdftk has this option. From man pdftk
fill_form <FDF data filename | XFDF data filename | - | PROMPT>Fills the single input PDF's form fields with the data from an FDF file, XFDF file or stdin. Enter the data filename af‐ ter fill_form, or use - to pass the data via stdin, like so:
pdftk form.pdf fill_form data.fdf output form.filled.pdf
How to do?
1. Extract field names
pdftk forms.pdf dump_data_fields will drop a list with e.g. entries like:
--- FieldType: Text FieldName: Name_Last FieldNameAlt: LAST FieldFlags: 0 FieldJustification: Left --- 2. Create fdf-file
fdf-files are form-fillers. We use the general notation as
<< /T(FIELD_NAME)/V(FIELD_VALUE) >> And need to add a header and footer. Source
%FDF-1.2 1 0 obj<</FDF<< /Fields[ << /T (Name_Last) /V (Smith) >> << /T (Name_First) /V (John) >> ] >> >> endobj trailer <</Root 1 0 R>> %%EOF To be saved as input.fdf
3. Create the document
pdftk forms.pdf fill_form input.fdf output filled.pdf 4. Maybe you want to check again?
pdftk filled.pdf dump_data_fields --- FieldType: Text FieldName: Name_Last FieldNameAlt: LAST FieldFlags: 0 FieldValue: Smith FieldJustification: Left --- So the last name has made it into the document.
5. Additional handling
For radio buttons, the value is one of the FieldStateOption-entries, for checkboxes use V(Yes) and /V(Off).
If you want to make it more of a line-by-line query type of thing, you might want to wrap a script around it that reads the field types and names and prompts for input, then automatically creates the fdf-file. But that is a small project of its own.
Alternatively, you might like the xfdf-format (which see) better. It is based on xml-style entries.
pdftk infile.pdf generate_fdf output foo.fdf