CLI tool for Word DOCX templating and analysis.

• Commands
  • Unzip DOCX: All files, or only media files, format XML
  • Zip files into DOCX
  • Output DOCX contents
    • List files: All, filtered, images only
    • List meta data
    • List referenced fonts
    • List fields
    • Output XML
    • Output document as plaintext
  • Compare DOCX documents
  • Modify document
    • Modify meta data
    • Replace image
    • Replace text
    • Insert text into existing table
    • Replace by markup
      • Insert heading
      • Insert text
      • Insert paragraph containing text
      • Insert hyperlink
      • Insert image
      • Insert list
      • Insert new table
    • Remove content between text
    • Set field value: Merge fields, generic fields
    • Randomize document text
  • Batch Templating
    • Replacement Pre/Post-Markers
  • Arbitrary manual and scripted anlysis / modification
  • Output docxBox help or version number
• Configuration
• Build Instructions
• Running Tests
• Code Convention
• Changelog
• Roadmap
• Bug Reporting and Feature Requests
• Third Party References
• License

Unzip all files: docxbox uz foo.docx

Unzip only media files:
docxbox uzm foo.docx or docxbox uz foo.docx -m
or docxbox uz foo.docx --media

Unzip all files and indent XML files:
docxbox uzi foo.docx
or docxbox uz foo.docx -i
or docxbox uz foo.docx --indent

docxbox zp path/to/directory out.docx

Compress XML, than zip files into DOCX:

When having indented XML (i.e. via uzi command) for manual manipulation, the zpc command compresses (= unindents) all XML files before zipping them into a new DOCX:

docxbox zpc path/to/directory out.docx

Lists files contained within a given DOCX, and their attributes:

docxbox ls foo.docx

To output as JSON:

docxbox lsj foo.docx
or docxbox ls foo.docx -j
or docxbox ls foo.docx --json

docxbox ls foo.docx *.xml Lists only files ending w/ .xml

docxbox lsl foo.docx foo Lists all files containing the string foo
or docxbox ls foo.docx -l foo
or docxbox ls foo.docx --locate foo

This command is a shorthand to the grep tool (must be installed on your system when using this command).
The search-string therefor can also be given as a regular expression:

docxbox lsl foo.docx '/[0-9A-Z]\{8\}/'
Lists all files containing 8-digit IDs, e.g. word recent session IDs (ISO/IEC 29500-1).

docxbox lslj foo.docx foo
or docxbox lsl foo.docx -j foo
or docxbox ls foo.docx -lj foo
or docxbox lsl foo.docx --json foo
or docxbox ls foo.docx --locate -j foo
or docxbox ls foo.docx --locate --json foo

Output list of contained images and their media attributes (like width, height, encoding, compression, etc.)

docxbox lsi foo.docx
or docxbox ls foo.docx -i
or docxbox ls foo.docx --images

To output as JSON:

docxbox lsij foo.docx
or docxbox lsi foo.docx -j
or docxbox ls foo.docx -ij
or docxbox lsi foo.docx --json
or docxbox ls foo.docx --images --json

Note: Media attributes are read using the file command, which must be installed on your system (but usually should be already) when using docxBox's lsi command.

docxBox displays only attributes that are contained within the current DOCX file (the attributes can vary by DOCX version and word processor used for creation), also if given empty.

Output meta data of given DOCX:

docxbox lsm foo.docx
or docxbox ls foo.docx -m
or docxbox ls foo.docx --meta

To output as JSON:

docxbox lsmj foo.docx
or docxbox lsm foo.docx -j
or docxbox ls foo.docx -mj
or docxbox lsm foo.docx --json
or docxbox ls foo.docx --meta --json

• Authors: Creator, lastModifiedBy (<dc:creator> and <cp:lastModifiedBy> of docProps/core.xml)
• Dates (ISO 8601): Creation-, modification and print-date
  (<dcterms:created> and <cp:modified> and <cp:lastPrinted> of docProps/core.xml)
• Descriptions: Description, Keywords, Subject, Title
  (<dc:description>, <dc:keywords>, <dc:subject>, <dc:title> of docProps/core.xml)
• Language (<dc:language> of docProps/core.xml)
• Revision (<cp:revision> of docProps/core.xml)
• Application created with and its version, name of used template, company, XML schema of document (<Application>, <AppVersion>, <Template>, <Properties xmlns ... and <Company> of docProps/app.xml)

docxbox lsf foo.docx
or docxbox ls foo.docx -f
or docxbox ls foo.docx --fonts

To output as JSON:

docxbox lsfj foo.docx
or docxbox lsf foo.docx -j or docxbox ls foo.docx -fj
or docxbox lsf foo.docx --json
or docxbox ls foo.docx --fonts --json

docxbox lsd foo.docx
or docxbox ls foo.docx -d
or docxbox ls foo.docx --fields

To output as JSON:

docxbox lsdj foo.docx
or docxbox ls foo.docx -dj
or docxbox lsd foo.docx --json
or docxbox ls foo.docx --fields --json

docxbox cat foo.docx word/_rels/document.xml.rels
outputs the given file's XML, indented for better readability.

Hint: For viewing or editing complex XML, e.g. with syntax highlightning, you can use your favorite text editor via the cmd command

docxbox txt foo.docx outputs the given document's plaintext (ATM: w/o header and footer)

Output plaintext segments:
docxbox txt foo.docx -s
or docxbox txt foo.docx --segments

Outputs the plaintext from document, with markup sections separated by newlines. This can be helpful to identify "segmented" sentences: Texts which visually appear as a unit, but are declared within multiple separate XML elements (due to formatting or change-tracking purposes).

docxBox helps tracing changes to the files contained within DOCX archives, made when manipulating documents in word processor applications.

When given two DOCX files, the ls command lists all files of both DOCX documents side-by-side. docxBox compares all files and highlights files w/ different attributes or (identical attributes but) different content.

docxbox ls foo_v1.docx foo_v2.docx

Note: Comparisons are always output as plaintext, JSON output is not supported.

Files that have changed between versions of a document, can be inspected using the diff tool (which must be installed on your system).

Display side-by-side comparison of the formatted XML of given file (word/settings.xml), with differences indicated:
docxbox diff foo_v1.docx foo_v2.docx word/settings.xml

Display unified diff: docxbox diff foo_v1.docx foo_v2.docx word/settings.xml -u
or: docxbox diff foo_v1.docx foo_v2.docx word/settings.xml --unified

docxBox allows to modify existing attributes, or adds attributes if not present.

• Set creation-date: docxbox mm foo.docx created "2020-01-29T09:21:00Z"
• Set creator attribute: docxbox mm foo.docx creator "docxBox v0.0.1"
• Set description attribute: docxbox mm foo.docx description "Foo bar baz"
• Set keywords attribute: docxbox mm foo.docx keywords "Foo bar baz"
• Set language attribute: docxbox mm foo.docx language "en-US"
• Set lastModifiedBy attribute: docxbox mm foo.docx lastModifiedBy "docxBox v0.0.1"
• Set lastPrinted attribute: docxbox mm foo.docx lastPrinted "2020-01-10T10:31:00Z"
• Set modification-date: docxbox mm foo.docx modified "2020-01-29T09:21:00Z"
• Set revision attribute: docxbox mm foo.docx revision 2
• Set subject attribute: docxbox mm foo.docx subject "Foo bar"
• Set title attribute: docxbox mm foo.docx title "Foo bar, baz"

Notes:

1. Altering meta data does NOT automatically update preview texts of generic fields, which display respective meta data.
   For updating field values, use the sfv command.
2. All modifications automatically update the modification-date attribute to the current timestamp, unless explicitly setting a different one.
3. During Batch Templating the modification-date is not updated automatically.

To alter/insert an attribute and save the modified document to a new file:
docxbox mm foo.docx <attribute> <value> new.docx

To update multiple meta attributes with one mm command, tuples of attribute-keys and -values can be given as JSON: docxbox mm foo.docx "{\"<attribute>\":\"<value>\",\"<attribute>\":\"<value>\", ...}" new.docx

docxbox rpi foo.docx image1.jpeg /home/replacement.jpeg overwrites the DOCX w/ the modified document.

Note: The original and replacement image must be of the same format (bmp, gif, jpg, etc.).

docxbox rpi foo.docx image1.jpeg /home/replacement.jpeg new.docx
This creates a new file: new.docx

Replace all (case-sensitive) occurrences of given string in DOCX text:

docxbox rpt foo.docx old new updates foo.docx
docxbox rpt foo.docx old new new.docx creates a new file new.docx

stv inserts values (and cells if needed) into an existing table, starting at 1st cell of given row. If there are less columns in the row than values given, more rows are added after the row.

This is useful for maintaining a specific table style (borders, coloring, font, etc.) when rendering dynamic documents from DOCX templates.

Example: Fill/Insert four cells starting w/ second row of first table in document:
docxbox stv foo.docx "{\"table\":1,\"row\":2,\"values\":[\"foo\",\"bar\",\"baz\",\"qux\"]}

Note: Table and rows are indexed starting w/ 1 (not 0).

The table and row to start inserting data into can also be identified by text (distinct within the document) contained within a cell of that table and row:

docxbox stv foo.docx "{\"cell\":\"insert-data-here\",\"values\":[\"foo\",\"bar\",\"baz\",\"qux\"]}

Moreover replacing text and fields, docxBox supports rendering and inserting the following Office Open XML elements:

• Heading 1, 2, 3
• Text
• Paragraph containing text
• Image (formats: bmp, emg, gif, jpeg, jpg, png, tif, tiff, wmf)
• Table
• Unordered list

Markup specification for such elements must be given as JSON, following these rules:

• JSON must be wrapped within {...}
• The first item must be a type identifier (h1, h2, h3, img / image, ol, table, ul)
• All attributes are given associative (as JSON object related to the type)
• The order of attributes within the config of the type is arbitrary

Example: Replace string search by a Heading 1 with the text Foo:
docxbox rpt foo.docx search "{\"h1\":{\"text\":\"Foo\"}}"

docxBox supports rendering of Header 1, 2 and 3 (h1, h2, h3).

Example: Replace string search (by a new run) with the text Foo:
docxbox rpt foo.docx search "{\"text\":{\"text\":\"Foo\"}}"

Example: Replace string search (by a new paragraph containing a run) with the text Foo:
docxbox rpt foo.docx search "{\"p\":{\"text\":\"Foo\"}}"

Example: Replace string search by a hyperlink:
docxbox rpt foo.docx search "{\"link\":{\"text\":\"docxBox\",\"url\":\"https://github.com/gyselroth/docxbox\"}}"

Replace string search by an unordered list:
docxbox rpt foo.docx search "{\"ul\":{\"items\":[\"item-1\",\"item-2\",\"item-3\"]}}"

Image markup specification example:

{
    "img":{
        "name":"example.jpg",
        "offset":[0,0],
        "size":[2438400,1828800]
    }
}

Specification rules:

• The name parameter is optional
• The offset argument is optional
• Image size is per default expected to be given in EMUs (= English Metric Unit, being: pixels * 9525), but can also be specified in Pixels like: "size\":[\"256px\",\"192px\"]

When inserting a new image file, it must be given as additional argument:
docxbox rpt foo.docx search "{\"image\":{\"size\":[2438400,1828800]}}" images/ex1.jpg

To replace text by a newly rendered table like:

the table specification as JSON looks like:

{
    "table":{
        "columns":3,
        "rows":3,
        "header":["A","B","C"],
        "content":[
            ["a1","b1","c1"],
            ["a2","b2","c2"],
            ["a3","b3","c3"]
        ]
    }
}

• header is optional, when given: columns is optional
• content is optional, when given: rows is optional

Replace search by table:
docxbox rpt foo.docx search "{\"table\":{\"header\":[\"A\",\"B\",\"C\"],\"content\":[[\"a1\",\"a2\",\"a3\"],[\"b1\",\"b2\",\"b3\"],[\"c1\",\"c2\",\"c3\"]]}}"

Remove content between (and including) given strings (left and right):

docxbox rmt foo.docx left right updates foo.docx
docxbox rmt foo.docx left right new.docx creates a new file new.docx

When setting the value (text) of a merge field, the merge field is reduced to its textual component (maintaining its visual style).

Note: A particular merge field can NOT be merged repeatedly: merging turns the former field into a text element (the field subsequently does not exist as such any more).

docxbox sfv foo.docx "MERGEFIELD foo" bar (Updates foo.docx)
Changes all merge fields, whose identifier begins with foo, into the text bar.

docxbox sfv foo.docx "MERGEFIELD foo" bar new.docx Saves the resulting DOCX to a new DOCX file: new.docx

Hint: To find out field identifiers use docxBox's lsd command.

Setting field values includes also preview texts of otherwise generic fields, which in some word processing applications have to be updated explicitly.

docxbox sfv foo.docx "PRINTDATE" "10.01.2020"
Updates the shown text of all print-date fields to 10.01.2020.

Replace all text of an existing document by similarly structured random "Lorem Ipsum" dummy text, helpful for generating DOCX documents for testing purposes:

docxbox lorem foo.docx updates foo.docx
docxbox lorem foo.docx new.docx creates a new file new.docx

docxBox's batch templating mode allows to perform an arbitrary sequence of operations (supporting all docxBox commands for document manipulation) upon a given DOCX. It thereby facilitates a more extensive range of templating options than the commands directly (= without batch templating) available.

Example: docxBox does not directly support replacing merge fields by other than plain textual content. Via batch templating, merge fields can be transformed into text in one step of a sequence, which can completely or in part, in a later step be replaced by generic content like for example a table, which can later be filled with more content.

Batch templating can make use of "markers": optional text elements containing a distinct identifier string. Markers can temporarily be inserted and can subsequently be replaced again at a later step of the batch sequence by other generic content.

Rules:

• Markers can be added before (key: pre) and after (key: post) the actual generic replacement content
• Markers can either be of the type text or paragraph (or p) to insert surrounding line-breaks
• Markers contain a textual identifier, which can use any text (but should be distinct within the document)

Sequences of templating steps to be batch-processed must be given like:

{
 "<STEP_ID>": {"<COMMAND>": [("<ARGUMENT_1>",)(,"<ARGUMENT_2>",...)]},
 "<STEP_ID>": {"<COMMAND>": [("<ARGUMENT_1>",)(,"<ARGUMENT_2>",...)]},
 ...
}

Example:

{
 "1": {"mm": ["description", "foo"]},
 "2": {"rpt": ["bar", "baz"]},
 "3": {"rpt": [
    "qux", 
    {"h1": {"text": "Quux"}}
 ]}
}

Rules:

• Every step must be given as a tuple of step-ID and -parameters
• <STEP_ID> is an arbitrary string, must be distinct within the sequence
• Parameters must be given as a tuple of a command and its respective arguments
• <COMMAND> accepts any of docxBox's commands for DOCX manipulation (rmt, rpi, rpt, lorem, mm and sfv)
• <ARGUMENT>: Argument(s) for respective command, same as in non-batch mode
• When a command has no arguments (e.g. lorem), an empty array must be given though (E.g.: {"lorem":[]})
• Arguments for markup-configuration of generic document elements can be given as nested JSON

Templating sequence:

1. Step "1": Replace string foo by heading-1 with the text: Foobar (followed by a temporary marker my-marker-1)
2. Step "2": Replace the marker my-marker-1 by table containing 2x2 cells
3. Steps "3" to "6": Replace (the placeholder texts within the) table cells by images
4. Add new image files into docx document

Batch config:

{
 "1": {"rpt": [
    "foo",
    {
     "h1": {
        "text": "Foobar",
        "post": {"text": "my-marker-1"}
     }
    }
 ]},
 "2": {"rpt": [
   "my-marker-1",
   {
     "table": {
         "columns": 2,
         "rows": 2,
         "header": ["A","B"],
         "content": [
             ["img-a1", "img-b1"],
             ["img-a2", "img-b2"]
         ]
     }
   }  
 ]},
 "3": {"rpt": [
    "img-a1",
    {
      "img": {
          "name": "blue.png",
          "size": [2438400, 1828800]
      }
    }  
 ]},
 "4": {"rpt": [
    "img-b1",
    {
      "img": {
          "name": "green.png",
          "size": [2438400, 1828800]
      }
    }
 ]},
 "5": {
    "rpt": [
      "img-a2",
      {
        "img":{
          "name": "orange.png",
          "size": [2438400, 1828800]
        }
      }
 ]},
 "6": {"rpt": [
    "img-b2",
    {
      "img": {
          "name": "red.png",
          "size": [2438400,1828800]
      }
    }
 ]}  
}

The full batch command:

Note: As when inserting new images in non-batch mode (via rpt or rpi), also during batch templating, image files to be added into the document must be given as trailing arguments.

docxbox batch foo.docx "{\"1\":{\"rpt\":[\"foo\",{\"h1\":{\"text\":\"Foobar\",\"post\":{\"text\":\"my-marker-1\"}}}]},\"2\":{\"rpt\":[\"my-marker-1\",{\"table\":{\"columns\":2,\"rows\":2,\"header\":[\"A\",\"B\"],\"content\":[[\"img-a1\",\"img-b1\"],[\"img-a2\",\"img-b2\"]]}}]},\"3\":{\"rpt\":[\"img-a1\",{\"img\":{\"name\":\"blue.png\",\"size\":[2438400,1828800]}}]},\"4\":{\"rpt\":[\"img-b1\",{\"img\":{\"name\":\"green.png\",\"size\":[2438400,1828800]}}]},\"5\":{\"rpt\":[\"img-a2\",{\"img\":{\"name\":\"orange.png\",\"size\":[2438400,1828800]}}]},\"6\":{\"rpt\":[\"img-b2\",{\"img\":{\"name\":\"red.png\",\"size\":[2438400,1828800]}}]}}" blue.png green.png orange.png red.png

To save the resulting document of batch processed manipulations to a new file, instead of overwriting the source document, the destination filename can optionally be given as the very last argument (also trailing other optional arguments like image files):

docxbox batch foo.docx "{\"1\":{\"mm\":[\"description\",\"foo\"]},\"2\":{\"rpt\":[\"bar\",\"baz\"]},\"3\":{\"rpt\":[\"qux\",{\"h1\":{\"text\":\"Quux\"}}]}}" new.docx

docxBox eases conducting arbitrary modifications on files contained within a DOCX, manually and scripted.
All steps besides the actual modification are automated via docxBox, with the respective user-defined modification inserted.

Example - Edit XML file manually:

docxbox cmd foo.docx "nano *DOCX*/word/document.xml"

docxBox in the above example does:

1. Unzip foo.docx
2. Indent all extracted XML files
3. Render (= replace *DOCX* w/ the resp. extraction path)
   and execute the command: nano *DOCX*/word/document.xml, thereby opening document.xml for editing in nano, halting docxBox until exiting the editor.
4. Unindent all extracted XML files
5. Zip the extracted files back into foo.docx

docxbox
or docxbox h
Outputs docxBox's help text.

docxbox h <command> Outputs more help on a given command.

docxbox v Outputs the installed docxBox's version number.

docxBox can optionally be configured using the following environment variables:

Example:
Export variable to the environment docxBox runs in: export docxBox_verbose=1

cmake CMakeLists.txt; make

In order to run functional tests, Bats must be installed.

Run all tests: ./test.sh

Run specific test suite:
./test.sh <suite>
E.g.: ./test.sh ls - Filenames in test/functional/ correspond to test suite names.

Check all tests for memory-leaks via Valgrind:
./test.sh valgrind
In order to check for memory-leaks, Valgrind must be installed on your computer.

The source code of docxBox follows the Google C++ Style Guide.
The source code of functional tests follows the Google Shell Style Guide

See Changelog

• v1.0.0: Ensure all templating options work and output is microsoft word compatible
• v1.0.0: Add HTTP/s server mode (make usable as local web service)
• v1.1.0: Libre-Office compatible appending of two DOCX files into a single one (by XML appending, instead of adding sub-documents)

If you find a bug or have an enhancement request, please file an issue on the github repository.

Microsoft Office and Word are registered trademarks of Microsoft Corporation.

docxBox was built using the following third party libraries and tools:

Thanks a lot!

docxBox is licensed under The MIT License (MIT)