Import a Word file

brevier import mahnung.docx -o templates/mahnung

The import writes three files: mahnung.html, mahnung.css and mahnung.schema.json. It writes no sample.json. Its input is a Word file in which a template library has marked where the data goes. The data code of that library stays as it is.

The fields that the import reads

An example: the Mahnung

The Mahnung is a dunning letter with a customer address, a table of open positions and a header and footer. All names and numbers are made up. This is the Velocity version, as XDocReport reads it. The right column is what the import wrote into mahnung.html.

The table in the Word file has a header row and one row that repeats. The row starts with the merge field @before-row#foreach($p in $mahnung.posten) and ends with @after-row#end. Its cells hold $p.bezeichnung, $p.menge and $p.betrag. The import writes:

<table class="MahnTabelle">
<thead>
<tr><th><p><strong>Bezeichnung</strong></p></th><th><p><strong>Menge</strong></p></th><th><p><strong>Betrag</strong></p></th></tr>
</thead>
<tbody>
{{#each mahnung.posten}}<tr><td><p>{{ bezeichnung }}</p></td><td><p>{{ menge }}</p></td><td><p>{{ betrag }}</p></td></tr>{{/each}}
</tbody>
</table>

The same letter written for docxtpl, docxtemplater or DocxTemplater for .NET gives the same mahnung.html.

Styles, header and footer

• Each Word style that the file uses becomes a CSS rule with the name of the style, for example .Heading1 (font-size: 16pt, bold, colour #1F3864) and .Betont (bold, colour #C00000).
• The page size and the margins of the document become @page.
• The header and the footer of the page become the blocks docx-header and docx-footer. The CSS puts them in the page margin with @top-center { content: element(docx-header); } and the same for the footer.
• A font that is embedded in the Word file is not copied. Install the font on the machine that renders, or load it with @font-face.

The schema

The schema lists every field of the template. Each value is a string, and a repeating row becomes an array of objects:

{
  "type": "object",
  "properties": {
    "mahnung": {
      "type": "object",
      "properties": {
        "nummer": { "type": "string", "title": "nummer" },
        "posten": {
          "type": "array",
          "items": {
            "type": "object",
            "properties": {
              "bezeichnung": { "type": "string", "title": "bezeichnung" },
              "menge": { "type": "string", "title": "menge" },
              "betrag": { "type": "string", "title": "betrag" }
            }
          }
        }
      }
    }
  }
}

The real file lists all fields of mahnung and kunde, and each object has required and "additionalProperties": false. The engine rejects a field that the schema does not list.

Change the type of a field where the Word file could not say it, for example "type": "number" for menge.

If your Java code binds objects by the name of their class (step 2), move each object into $defs, with the name of the class, and point its root at the type with $ref:

{
  "properties": { "kunde": { "$ref": "#/$defs/Kunde" } },
  "$defs": {
    "Kunde": {
      "type": "object",
      "properties": { "name": { "type": "string" } }
    }
  }
}

Check the result

Open mahnung.html and mahnung.css and compare them with the Word file. The import writes what it can read. Then render the letter with data:

brevier pdf templates/mahnung/mahnung.html --data mahnung.json -o mahnung.pdf

Next: Replace the call. For a worked example from start to end, see Example.