I did a conversion program to change the object structure of a DWFx file, and it works fine. What I did was to open the DWFx file as a zip archive, parse the internal XML files, and reorganize them, creating new parent nodes when needed.
But what doesn't work is changing the names of these nodes. When I open the file in any Autodesk viewer (the offline Design Review program and the online Viewer are the ones I tested), the tree structure is changed as it should, but the parent node names are not. In fact, the nodes that already existed keep their old names, and the new ones are called Object XXXX. The child nodes (actual objects) have their names changed correctly.
I tried to search in every readable (text) file inside the DWFx, but none of them hold any other reference to these nodes. I didn't open binary files, like W3D files, which probably hold the geometry.
Does anyone have any experience in creating or altering DWFx files? Do I need to change anything else besides the 'label' tag in the Presentation XML file?
Instead of manipulating the contents of DWFx files manually, consider using one of the Autodesk Forge services: Design Automation. The APIs allow you to run AutoCAD "in the cloud". You could theoretically load your DWFx file there, update the structure/names, and generate an updated DWFx file. Here's an example of how the service can be used to generate PDFs out of DWG files: https://forge.autodesk.com/en/docs/design-automation/v2/tutorials/convert-dwg-to-pdf.
Related
In certain circumstances, BIM360 will serve a zip file of a Revit document along with its links, such as explained here: https://forums.autodesk.com/t5/bim-360-document-management/linked-revit-files-in-bim-360-docs/td-p/8774004
In this circumstances, however, when interacting with GET projects/:project_id/folders/:folder_id/contents the file still is shown as a regular file (potentially the isCompositeDesign attribute distinguishes it) with a .rvt file extension. In addition, the filesize shown in storageSize of the object is the sum of the main Revit file and all of its links. Checking the details in GET buckets/:bucketKey/objects/:objectName/details equally show the size object size attribute to be the sum of the main Revit file and all of its links.
I cannot seem to find functionality in Forge that:
Distinguishes a zip file from a lone file (potentially the isCompositeDesign attribute does this)
Provides a list of which other files are linked into the main file, or a list of the zip file contents and their URNs.
Provides a true filesize of the main revit file itself, not just the sum of all linked files in the zip.
Ideas?
Revit 4 worksharing, publishes a file to BIM360.
This file is named as a .rvt file (ie. 'mybigrevitproject.rvt'), but in fact, it's really a zip file in disguise. If you rename it to zip, download it, and unzip it, you'll find lots of .RVT inside the zip.
There's a neat trick to figuring this out, without downloading the entire file.
Use a range GET on the first 16 bytes, and check for the magic header.
For full details, check out this repo: https://github.com/wallabyway
Here's a snippet of the code that will help:
https://github.com/wallabyway/bim360-zip-extract/blob/master/server.js#L167
I think it's related to this question: Forge Data management returns zip file
I've implemented offline viewing based on the tutorial and github here. The problem I'm having is forge is looking for the db jsons in an odd location that makes url pathing awkward. The files in particular are 'objects_attrs.json.gz', 'objects_vals.json.gz', 'objects_offs.json.gz', 'objects_ids.json.gz', 'objects_avs.json.gz'. For some reason the forge viewer strips two layers of directories off the url then looks for the files there. Afterword Forge looks for the files in the original location but, looks for the straight json instead of the gziped jsons.
This can be handled in a few hacky ways like creating two arbitrary parent directories in the url and accepting files at the higher url as well. Or unzipping the gzips and saving them in location. But, these kind of hacks could easily be unstable if I'm not exactly right about the forge implementation.
Is there any reason these files are searched for in a different location? For example are these files also used by the 2d files not just the 3d files? Are there other files like this?
Look for the .svf file, say it's Design.svf. Rename it to Design.zip, and then upzip the file. This will get you the manifest.json.
Inside the manifest.json file, you'll file the URI's of all the files, including the one's you mentioned (the property database files).
Modify the URI as you wish, save the manifest.json file and re-zip it back up and rename it back to Design.svf.
If you got the URI path matching your file-system, then those property files will now load correctly.
Good luck!
I need to generate many parent-children related HTML files representing multilevel content (from a graph of nodes in this case), but I don't know the best output structure to create.
I can imagine the next options:
Generate all nodes (HTML files) into a single MHTML file. Problem: I've read that this format is not very "standard", requiring extension to browsers.
Generate a directory structure, where each dir will containt its node HTML file plus dirs with the children nodes. Problem: Maybe a lot of directories will be created and could become a mess for the user.
Generate a directory structure, like in point 2, but inside a .ZIP file. Problem: Not directly viewable by a web browser.
Generate a single directory with all the HTML files at the same level, but renamed to avoid same-name clashes (like "node15atlevel1.node38atlevel2.node74atlevel3.html"). Problem: File names could become too long due to deep nested containment.
Generate files in the same directory, like in point 4, but using short coded file names (like "n015038074.html" instead of "node15atlevel1.node38atlevel2.node74atlevel3.html".
So, what do you think? what is the most practical and viable way to output nested content as HTML?
Thanks!
I need to visit a folder and all of its children with SSIS (SQL Server Integration Services). At the moment by setting the folder path into a variable after reading it, I able to loop through all the .txt files of the current folder and fill a pre-generated (with head info) xml file.
What I would need now is to be able to create one per each accessed folder, a new xml file (the beginning content will be always the same). Once I would be able to create it, as first action once a new folder is accessed, I can then simply apply the logic I developed so far.
However I am blocked at the moment, since within the loop where i read the files (with their full path) I cannot find a way to express "create the xml file if the accessed folder is new".
Assuming I understand the problem, you need to walk the entirety of a directory structure and for each folder you find, you need to create a base XML file. Then for each text file you find in that folder, you will perform some operation on the XML file. The trick being how do you only create the XML file once.
I would envision a process like this.
A script task that makes use of the System.IO.GetDirectories to populate a variable (directoryXML> that contains the folder structure, something like
<Dir>
<D>C:\ssisdata</D>
<D>C:\ssisdata\a</D>
<D>C:\ssidata\a\b</D>
</Dir>
Use a Foreach Nodelist Enumerator to shred that XML out into a variable (currentDirecotry).
You'd perform your one-time task of creating the XML file in currentDirectory.
Further using the currentDirectory variable as an expression on the Foreach File Enumerator (assign to Directory with a FileSpec of *.txt) you can then perform your task on all the files meeting that specification. Do not check the traverse subfolder option as that will not give the desired results.
This is a fairly high level approach to the problem as I'm assuming you have some familiarity with SSIS but the approach should be sound. Let me know if you have any particular sticking points.
Is it possible, via Adobe Air, to save multiple types of data in a single file? For example, an application would allow the user to load in external images, position them on stage and label them. This data would be then be stored in a ByteArray (I guess) using BitmapData for the images and probably XML for the metadata.
I would then like to write this to a single file, with a bespoke file extension that could be associated with said Air app.
I've asked this on various forums and never received a single reply.
You can add everything to a byte array and write it to file - but defining boundaries and extracting individual entities back from the file would take some effort. How about writing them to normal files, zipping them to a single file and deleting the originals? This way you can still have a single file and deal with the individual items more easily.
This article describes some ActionScript zip libraries. I've used nochump in the past and it was easy - this page has some sample code
If you want some individuality for your files, you can rename the zipped file to whatever extension you want - that's what Firefox extensions do, they have .xpi extension, but they're plain zip files renamed.