Bug: Links in PDFs do not work on files that are included via `DOCUMENT_FROM_FILE`

[CMon]

Describe the bug

I organize all my requirements in different .sdoc files and have one main.sdoc (basically as a simple TOC) and include the others via the DOCUMENT_FROM_FILE keyword. When using the html view everything works as expected, but when exported as a PDF or directly converted to a PDF the links from included files do not work anymore. links inside the main.sdoc work.

To Reproduce

Create the files needed to reproduce the above:

main.sdoc:

[DOCUMENT]
TITLE: Project Risk Management

[TEXT]
STATEMENT: Main file

[DOCUMENT_FROM_FILE]
FILE: included.sdoc

[REQUIREMENT]
UID: REQ-003
TITLE: Third one
STATEMENT: Third statement links to fourth: [LINK: REQ-004]

[REQUIREMENT]
UID: REQ-004
TITLE: fourth one
STATEMENT: The fourth statement

included.sdoc:

[DOCUMENT]
TITLE: Included doc

[REQUIREMENT]
UID: REQ-001
TITLE: First one
STATEMENT: First statement links to second: [LINK: REQ-002]

[REQUIREMENT]
UID: REQ-002
TITLE: second one
STATEMENT: The second statement

compile to pdf with:
docker run --rm -v $(pwd):/data -e HOST_UID=$(id -u) -e HOST_GID=$(id -g) -it strictdoc:latest /bin/bash -c 'cd /data && strictdoc export . --output-dir output --formats html2pdf'

I use the Dockerfile from the current repo

Expected behavior

REQ-001 should link to REQ-002, like REQ-003 does to REQ-004

Environment

strictdoc version: 0.14.0
Arch linux and official docker container from the repository to use strictdoc from within the container.

[stanislaw]

Hello @CMon,

I tried reproducing this issue using the examples you provided, but I couldn't. The important detail that is likely missing in your examples is that you need to use >>> <<< multiline fields for the LINK syntax to be recognized.

I have just added some documentation that should clarify this better: #2505.

[stanislaw]

Please reopen this issue if it still doesn't work for you 🙏

[CMon]

@stanislaw I am not allowed to reopen issues, so I hope you see the following comment.

It still does not work.

I recreated the docker image from todays main branch, changed the statements to have >>> <<< (they originally where in that state, i just shortened them for the example code, always happy to see better docs.

my strictdoc_config.py has the follwoing content (forgot that above):

from strictdoc.core.project_config import ProjectConfig

def create_config() -> ProjectConfig:
    config = ProjectConfig(
        project_title = "POC Bug",
        project_features = [
            "SEARCH",
            "HTML2PDF",
        ],
        server_host = "localhost",
        server_port = 5000,
    )
    return config

Compile everything with:
docker run --rm -v $(pwd):/data -e HOST_UID=$(id -u) -e HOST_GID=$(id -g) -p 5000:5000 -it strictdoc:latest /bin/bash -c 'cd /data && strictdoc export . --output-dir output --formats html2pdf'

output:

strictdoc/docker: running a Docker container entrypoint.
strictdoc/docker: ensuring strictdoc user with UID=1000 and GID=1000 exists
strictdoc/docker: creating new group strictdoc with GID=1000
strictdoc/docker: creating new user strictdoc with UID=1000
strictdoc/docker: show created user info:
uid=1000(strictdoc) gid=1000(strictdoc) groups=1000(strictdoc)
Parallelization: Enabled
Step 'Parse SDoc project tree' start.
Step 'Find and read SDoc files' start.
Find asset directories ...................................... 0.00s
Find SDoc files ............................................. 0.00s
Completed finding SDoc and assets ........................... 0.00s
Reading SDOC: included.sdoc ................................. 0.00s
Reading SDOC: main.sdoc ..................................... 0.00s
Completed parsing all documents ............................. 0.03s
Completed building document tree ............................ 0.03s
Step 'Find and read SDoc files' took: 0.04 sec.
Step 'Build traceability graph' start.
Step 'Build traceability graph' took: 0.01 sec.
Step 'Parse SDoc project tree' took: 0.06 sec.
Step 'Export SDoc' start.
Copying StrictDoc's assets: 51 files (51 files skipped as non-modified).
Copying HTML2PDF.js: 1 files (1 files skipped as non-modified).
Generating printable HTML document .......................... 0.21s
html2pdf4doc: version 0.0.22
html2pdf4doc: creating ChromeDriver service.
html2pdf4doc: Installed Chrome version: 141.0.7390
html2pdf4doc: OS system: Linux, OS type: linux64.
html2pdf4doc: ChromeDriver exists in the local cache: /data/output/_cache/0.14.0/html2pdf/141/chromedriver-linux64/chromedriver
html2pdf4doc: ChromeDriver available at path: /data/output/_cache/0.14.0/html2pdf/141/chromedriver-linux64/chromedriver
html2pdf4doc: creating ChromeDriver.
html2pdf4doc: opening URL with ChromeDriver: file:///data/output/html2pdf/html/data/main.html
success: HTML2PDF4Doc completed its job.
html2pdf4doc: JS logs from the print session:
"""
{'level': 'INFO', 'message': 'file:///data/output/html2pdf/html/_static/html2pdf4doc.min.js 1:92978 "[HTML2PDF4DOC] Version:" "0.2.3"', 'source': 'console-api', 'timestamp': 1760442860034}
{'level': 'INFO', 'message': 'file:///data/output/html2pdf/html/_static/html2pdf4doc.min.js 1:2956 "[HTML2PDF4DOC] Config:" Object', 'source': 'console-api', 'timestamp': 1760442860035}
{'level': 'INFO', 'message': 'file:///data/output/html2pdf/html/_static/html2pdf4doc.min.js 1:97011 "[HTML2PDF4DOC] Page count:" 4', 'source': 'console-api', 'timestamp': 1760442860078}
{'level': 'DEBUG', 'message': 'file:///data/output/html2pdf/html/_static/html2pdf4doc.min.js 1:97063 "[HTML2PDF4DOC] Total time: 44.031982421875 ms"', 'source': 'console-api', 'timestamp': 1760442860078}
"""
html2pdf4doc: executing print command with ChromeDriver.
html2pdf4doc: validating page count ......................... 0.01s
html2pdf4doc: exit handler: quitting the ChromeDriver.
PDFPrintDriver: printing HTML to PDF using HTML2PDF and Chrome Driver . 1.16s
Step 'Export SDoc' took: 1.39 sec.

Open the file in ocular and if I click on the "1.2. second one" I get the following error:

My guess would be you are not using it inside a container, and the path exists in your system => it might be able to open something

[stanislaw]

I am not allowed to reopen issues, so I hope you see the following comment.

Your comment and the provided screenshot helped me finally understand the issue 😄 Indeed, there was missing handling of the LINKs when they were contained in included documents imported into including documents and printed with HTML2PDF.

I have reflected a lot on issues like this, and I don't have a perfect way to solve them in advance. It's a great example of two features being developed relatively independently and then colliding later. I maintain a private Excel list that tracks all these collisions in an N×N matrix, with features listed both horizontally and vertically. Whenever a new feature is added, in the worst case, it could require me and other contributors to implement N integration tests to ensure all interactions with the existing features are covered. Unfortunately, this is not feasible given that StrictDoc is primarily a spare-time project.

Long story short: the current issue is fixed, but I also found that printing bundle documents with HTML2PDF still requires fixing the resolution of LINKs.

Please let me know if your original case now works by installing the alpha version of StrictDoc from Pip with:

pip install strictdoc --pre

I will reopen this ticket and keep it open until the handling of bundle documents is fully resolved 👍