I tracked the time and takes ~200ms everytime that we send the "Executing <NODE NAME> n%".
So this means that if you have 10 custom nodes we are adding 2 extra seconds to the execution.
200 * 10 = 2,000.
Some workflows are more complext and have more custom nodes, so this only keeps increasing.
* feat(text): create node for external text list
This is to send a list of texts to other nodes
* refactor: remove prints and rename variable
* style: update comment
* refactor: remove unused optional inputs
* fix(image upload): skip when using the CD_BYPASS_UPLOAD env var
* Revert "fix(image upload): skip when using the CD_BYPASS_UPLOAD env var"
This reverts commit 384eda63e6.
* fix(upload outputs): skip images/gifs/files/mesh when env var is true
The env var is `CD_BYPASS_UPLOAD`.
When that variables is `True`, we don't upload the media to our comfy
deploy s3 bucket.
There are 2 steps.
1. save the file into our s3 bucket
2. save the saving into our database.
When `CD_BYPASS_UPLOAD` is True:
1. Skip the save file into our s3 bucket
2. Skip the save into our database
Previously we were skipping the step 1, but not the step 2. So that is
the reason of why we keep seeing the comfy deploy URL when fetching the
run details:
```
outputs: [
{
data:{
gifs: [
{
url: "https://comfy-deploy-output.s3.amazonaws.com/video.mp4"
}
],
text: [
"A text that you displayed with show text node"
]
}
}
]
```
With the new changes we don't save that into our database, and fetching
the details of a run will look like this:
```
outputs: [
{
data:{
text: [
"A text that you displayed with show text node"
]
}
}
]
```
* feat(example page): add new examples section in the navbar
* feat(example page): initial layout of the new page
* feat(example page): create initial 4 workflows
* style(example page): center title
* style(example page): remove comment
* fix(example page): add hyperlink for txt2img card
Emmanuel Morales