mapillary_download

Simple code to download images in one or several mapillary sequences. The images will be geotagged and oriented.

How to use

Setup env variables

copy the secret variables files to setup your maipllary dev token, and your panoramax OpenStreetMap token.

cp secrets_variables_examples.sh secrets_variables.sh
editor secrets_variables.sh

Get ALL the sequences of SEVERAL users

To avoid to download everything at once, this is a 2 steps process:

• 1/ We set a list of users for which we want to get the list of sequences to download, and get a bash script for each user. Change the list of users in the batch file batch_get_username.sh. change the access token with your access token and the sequence ids with the ids of the sequences you want to download

  usernames=( "someone_having_nice_pictures" "someone_else" "oh_look_a_these_usernames" ) # use real accounts, it might work better
  

  This will give you a few files that the other scripts will use to generate a script to download all sequences of each user. One user, one bash script generated.

• 2/ We can launch the download of all the sequences of a user for every user specified in the batch script.

  bash script_bash_get_sequences_for_user_SOMEONE.sh
  

  Or all of the generated batch scripts at once.

  bash script_bash_get_sequences_for_user_*
  

Don't worry if you rerun the same user script_bash_get_sequences_for_user_someone.sh twice, the pictures will not be downloaded twice. Every run of mapillary_download.py checks that pictures are not existing before willing to download them.

Get only one sequence with its ID

To find a sequence id you can use the batch script, or click on a sequence on the Mapillary website, click on a picture, click on advanced data, and copy the sequence ID.

python mapillary_download.py "MLY|xxxx|xxxxxxx" --sequence_ids xxxxxxxxxxx xxxxxxxxxxx

Available arguments

python mapillary_download.py -h
usage: mapillary_download.py [-h] [--sequence_ids [SEQUENCE_IDS ...]] [--image_ids [IMAGE_IDS ...]] [--destination DESTINATION]
                   [--image_limit IMAGE_LIMIT] [--overwrite]
                   access_token

positional arguments:
  access_token          Your mapillary access token

optional arguments:
  -h, --help            show this help message and exit
  --sequence_ids [SEQUENCE_IDS ...]
                        The mapillary sequence id(s) to download
  --image_ids [IMAGE_IDS ...]
                        The mapillary image id(s) to get their sequence id(s)
  --destination DESTINATION
                        Path destination for the images
  --image_limit IMAGE_LIMIT
                        How many images you want to download
  --overwrite           overwrite existing images
  -v, --version         show program's version number and exit

Sort pictures depending of a bounding box in their Exif GPS Data in a separate folder

This is used to restrict zones for national instances of Panoramax to upload to reuse the sequences of multi national users.

Edit the bounding boxes in find_france_photos_and_move.py and change the destination folder. Running the script will check in the /data folder for every user sequences and move the files being in the bouding boxes in the destionation folder path of the script.

For this script to read the exif of pictures, you will need the python package exifread It was tested with pictures taken with the App Open Camera.

How to get my Mapillary access token

• Go to https://www.mapillary.com/dashboard/developers

• Click on "Registrer Application", enter the needed informations, enable the application to "Read" data, then click on register :

  

• When this registration is done, click on "view" in the token column. This is you access token :

  

How to get my sequence id (or sequence key)

• Go to https://mapillary.com/app

• Click on one of the picture of the sequence you want to download

• Click on the "image option" button (right panel)

• Click on "advanced" then click on the sequence key to copy it in the clipboard

  

How to install the script with a virtual environnement

cd ~
git clone https://github.com/Stefal/mapillary_download.git
cd mapillary_download
python3 -m venv mly_venv
source mly_venv/bin/activate
python -m pip install -r requirements.txt

Then you can run python mapillary_download "MLY|xxxx|xxxxxxx" --sequence_ids xxxxxxxxxxx When you're done with the script, simply run deactivate to exit the virtual environnement.

On windows you can create a virtual environnement too, or use the prebuilt mapillary_download.exe available on the release page.

How to batch the upload to panoramax

After having downloaded the sequences of a user you will find them in the folder data. go in your data folder, and to the user subfolder, and assuming you have already sent pictures with geovisio_cli, you can run this command to send all the sequences of the current folder you're in:

cd data/some_user

for dir in */; do   dir=${dir%?}  ;   geovisio upload --api-url https://panoramax.openstreetmap.fr "$dir" --token=BLAH_BLAH ; done

Have fun!