downlaod_lexicon.py

[HHAOi]

Hello, sorry to bother you. I have some questions. I've downloaded the signsuisse dataset online, but I don't know how to change download_lexicon.py. Hope you can give me a hint, I would appreciate it!

[AmitMY]

What are you trying to do? download_lexicon allows you to download the signsuisse lexicon.
Once you have it, you can for example run:

text_to_gloss_to_pose \
  --text "Kleine Kinder essen Pizza in Zürich." \
  --glosser "simple" \
  --lexicon "<directory to lexicon>" \
  --spoken-language "de" \
  --signed-language "sgg" \
  --pose "quick_test.pose"

[HHAOi]

What I mean is that when I run download_lexicon.py, a network connection error is reported, which causes the dataset to fail to download (because I am in China), so I downloaded the signsuisse dataset on Google Datasets. Now I want to run download_lexicon.py to get index.csv.

[AmitMY]

Ah! China is not a use case I thought about.

For now, I assume you downloaded from https://www.swissubase.ch/en/catalogue/studies/20452/latest/datasets/2327/2685/files

You'll have to unpack the mediapipe files, and then use metadata_train.csv to create a csv smilar to the dummy lexicon https://github.com/sign-language-processing/spoken-to-signed-translation/blob/main/assets/dummy_lexicon/index.csv

Sorry you are facing these issues, I will consider hosting the database on Cloudflare R2 instead of GCP in order to make it available from china. (or you could use a VPN)

[HHAOi]

Thanks，I'll try it.