Hi
I’m trying to update my uploading process to make use of the new ‘zip’ option, and I can’t quite figure out how it’s supposed to work. I’ve updated to v0.10.0.
I’m using “mapillary_tools process PIXPATH/FOLDER”, which is happy, and generates the description .json file.
Then “mapillary_tools zip PIXPATH/FOLDER ZIPPATH/FOLDER”, which generates a zip file called ZIPPATH/FOLDER/mly_tools_etc.zip - which contains all the jpg files, but not the .json file.
Then “mapillary_tools upload ZIPPATH/FOLDER”, as per the docs at GitHub - mapillary/mapillary_tools: Command line tools for processing and uploading Mapillary imagery
This fails with MapillaryFileNotFoundError: Description file ZIPPATH/FOLDER/mapillary_image_description.json not found. Has the directory been processed yet?
If I try giving it zip file itself, it fails with MapillaryBadParameterError: The description path must be specified (with --desc_path) when uploading a single file
.
If I add --desc_path PIXPATH/FOLDER/mapillary_image_description.json, it fails with Nothing uploaded. Bye.
What’s the correct sequence of commands to process a folder, generate the zip, and upload the zip, please?
PS: Why don’t I just keep using process/upload? Because the upload command does the prepare-as-a-zip stage, which takes around 10 mins for each sequence I upload, which means for 10 mins nothing is uploading. If I can pre-zip while the previous sequence is uploading, I can close that gap. Why do I care? Because I’ve a fairly slow upload speed, and I gather photos faster than I can upload them. I’ve just (as in, yesterday) finished uploading sequences I took last summer, with the uploader running 24/7 since then.