G

gregflurry

jakaskerl Thanks again for the help. I started to do what you suggested and I realized that my environment is likely configured wrong. I think I mentioned early in this thread that I tried getting things set up in 2022 and failed miserably. I cannot remember what I did much less why. I believe I'm running into some problems because of that history.

Some examples: 'python --version' -> 3.9.5 which is the Python that homebrew installed back in 2022; 'python3 --version' -> "3.12.5", which is the Python that homebrew installed last week; 'python3 -m ssl' -> nothing, which I've read means everything is OK; 'python -m ssl' -> "Traceback ... in <module> import _ssl import _ssl ...", in other words bad stuff; 'pyenv versions' -> "system

• 3.9.5 (set by /Users/gregflurry/.pyenv/version)"; the install appears to use 3.12.5, which might not be the right thing, but the pip WARNINGs mention Python3.9.

I fear that with all that, I suspect need to fix my Python environment; I may or may not need to use pyenv. I am not really sure what of the proper configuration (3.9.5 or 3.12.5; use pyenv or not). Advice needed.

jakaskerl This is so much fun.

I did some more GPT stuff. I used homebrew to install openssl, but it was already there. I used homebrew to reinstall python3, and set up to make sure the newly installed Python was used. Using 'python3 -m ssl' think I verified that Python has access to ssl. This time, things got further:

Gregorys-MBP:~ gregflurry$ 
bash -c "$(curl -fL https://docs.luxonis.com/install_depthai.sh)"
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100   122  100   122    0     0    830      0 --:--:-- --:--:-- --:--:--   829
100  7490  100  7490    0     0  16830      0 --:--:-- --:--:-- --:--:-- 16830
mkdir: /Users/gregflurry/Luxonis: File exists

ENTER absolute installation path for depthai or leave empty and default path: $HOME will be used.

Using default installation path: /Users/gregflurry/Luxonis
Directory: /Users/gregflurry/Luxonis is OK

Python version: Python 3.12.5 found.
If you want to use it for installation, press ENTER key, otherwise input path to python binary.
Press ENTER key to continue
_____________________________
Calling macOS_installer.sh
_____________________________
Running macOS installer.
Installing global dependencies.
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100   127  100   127    0     0   1101      0 --:--:-- --:--:-- --:--:--  1104
100  8636  100  8636    0     0  28978      0 --:--:-- --:--:-- --:--:-- 28978
Homebrew is already installed.
Git Already installed..
Finished installing global libraries.
Demo app already downloaded. Checking out main and updating.
remote: Enumerating objects: 8, done.
remote: Counting objects: 100% (8/8), done.
remote: Compressing objects: 100% (3/3), done.
remote: Total 8 (delta 5), reused 7 (delta 5), pack-reused 0 (from 0)
Unpacking objects: 100% (8/8), 1.68 KiB | 214.00 KiB/s, done.
From https://github.com/luxonis/depthai
   6901f745..3fe8c58c  main       -> origin/main
Already on 'main'
Your branch is behind 'origin/main' by 2 commits, and can be fast-forwarded.
  (use "git pull" to update your local branch)
Updating 6901f745..3fe8c58c
Fast-forward
 depthai_sdk/src/depthai_sdk/trigger_action/trigger_action.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Creating python virtual environment in /Users/gregflurry/Luxonis/venv
/opt/homebrew/bin/python3
WARNING: pip is configured with locations that require TLS/SSL, however the ssl module in Python is not available.
Requirement already satisfied: pip in /Users/gregflurry/Luxonis/venv/lib/python3.9/site-packages (21.1.1)
WARNING: Retrying (Retry(total=4, connect=None, read=None, redirect=None, status=None)) after connection broken by 'SSLError("Can't connect to HTTPS URL because the SSL module is not available.")': /simple/pip/
WARNING: Retrying (Retry(total=3, connect=None, read=None, redirect=None, status=None)) after connection broken by 'SSLError("Can't connect to HTTPS URL because the SSL module is not available.")': /simple/pip/
WARNING: Retrying (Retry(total=2, connect=None, read=None, redirect=None, status=None)) after connection broken by 'SSLError("Can't connect to HTTPS URL because the SSL module is not available.")': /simple/pip/
WARNING: Retrying (Retry(total=1, connect=None, read=None, redirect=None, status=None)) after connection broken by 'SSLError("Can't connect to HTTPS URL because the SSL module is not available.")': /simple/pip/
WARNING: Retrying (Retry(total=0, connect=None, read=None, redirect=None, status=None)) after connection broken by 'SSLError("Can't connect to HTTPS URL because the SSL module is not available.")': /simple/pip/
Could not fetch URL https://pypi.org/simple/pip/: There was a problem confirming the ssl certificate: HTTPSConnectionPool(host='pypi.org', port=443): Max retries exceeded with url: /simple/pip/ (Caused by SSLError("Can't connect to HTTPS URL because the SSL module is not available.")) - skipping
WARNING: pip is configured with locations that require TLS/SSL, however the ssl module in Python is not available.
Could not fetch URL https://pypi.org/simple/pip/: There was a problem confirming the ssl certificate: HTTPSConnectionPool(host='pypi.org', port=443): Max retries exceeded with url: /simple/pip/ (Caused by SSLError("Can't connect to HTTPS URL because the SSL module is not available.")) - skipping
Collecting pyqt5
  Downloading PyQt5-5.15.11-cp38-abi3-macosx_11_0_arm64.whl.metadata (2.1 kB)
Collecting PyQt5-sip<13,>=12.15 (from pyqt5)
  Downloading PyQt5_sip-12.15.0-cp312-cp312-macosx_10_9_universal2.whl.metadata (421 bytes)
Collecting PyQt5-Qt5<5.16.0,>=5.15.2 (from pyqt5)
  Downloading PyQt5_Qt5-5.15.14-py3-none-macosx_11_0_arm64.whl.metadata (536 bytes)
Downloading PyQt5-5.15.11-cp38-abi3-macosx_11_0_arm64.whl (6.6 MB)
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 6.6/6.6 MB 14.6 MB/s eta 0:00:00
Downloading PyQt5_Qt5-5.15.14-py3-none-macosx_11_0_arm64.whl (36.7 MB)
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 36.7/36.7 MB 23.3 MB/s eta 0:00:00
Downloading PyQt5_sip-12.15.0-cp312-cp312-macosx_10_9_universal2.whl (124 kB)
Installing collected packages: PyQt5-Qt5, PyQt5-sip, pyqt5
Successfully installed PyQt5-Qt5-5.15.14 PyQt5-sip-12.15.0 pyqt5-5.15.11
Collecting packaging
  Downloading packaging-24.1-py3-none-any.whl.metadata (3.2 kB)
Downloading packaging-24.1-py3-none-any.whl (53 kB)
Installing collected packages: packaging
Successfully installed packaging-24.1
DepthAI has been added to your PATH in .bashrc and .zshrc (if present).
If you prefer, you can manually add the following line to your .bash_profile for it to be recognized in login shells:
export PATH=$PATH:/Users/gregflurry/Luxonis/depthai/entrypoint


:::::::::::::::: INSTALATION COMPLETE ::::::::::::::::


To run demo app write <depthai_launcher> in terminal.
Press ENTER KEY to finish and run the demo app...

STARTING DEMO APP.
Traceback (most recent call last):
  File "/Users/gregflurry/Luxonis/depthai/launcher/launcher.py", line 11, in <module>
    from splash_screen import SplashScreen
  File "/Users/gregflurry/Luxonis/depthai/launcher/splash_screen.py", line 2, in <module>
    from PyQt5 import QtCore, QtGui, QtWidgets
ModuleNotFoundError: No module named 'PyQt5'

Failed installing depthai. Could be a bug in the installer or unsupported platform. Open a bug report over at https://github.com/luxonis/depthai - exited with status 1 at line 231 

Note that the problems with pip and ssl continue. Also there is a problem with PyQt5. I have no idea what to do with the former (a.f.i.k. I have no proxy or firewall). I guess I could install PyQt5, but I'd assume the installer would install all dependencies. Basically, I remain clueless, and of course without a working OAK-D.

jakaskerl Thanks for the pointer. I tried your suggested approach and it did not work. 'bash --version' reported v3. I used GPT and it actually gave very similar instructions, but they were incorrect, possibly because I used homebrew to install bash. I tried setting the default shell in Terminal and in Apple settings. 'bash --version' reported v3. I finally found something that said put the path to the homebrew version in the shell’s configuration file. That got me to 'bash --version' reporting v5.

So then I executed the install command again, and it failed. Once again I have no clue how to proceed.

Gregorys-MBP:~ gregflurry$ bash -c "$(curl -fL https://docs.luxonis.com/install_depthai.sh)"
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100   122  100   122    0     0    990      0 --:--:-- --:--:-- --:--:--   991
100  7490  100  7490    0     0  25124      0 --:--:-- --:--:-- --:--:-- 25124
mkdir: /Users/gregflurry/Luxonis: File exists

ENTER absolute installation path for depthai or leave empty and default path: $HOME will be used.

Using default installation path: /Users/gregflurry/Luxonis
Directory: /Users/gregflurry/Luxonis is OK

Python version: Python 3.9.5 found.
If you want to use it for installation, press ENTER key, otherwise input path to python binary.
Press ENTER key to continue
_____________________________
Calling macOS_installer.sh
_____________________________
Running macOS installer.
Installing global dependencies.
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100   127  100   127    0     0   1020      0 --:--:-- --:--:-- --:--:--  1024
100  8636  100  8636    0     0  30815      0 --:--:-- --:--:-- --:--:-- 30815
Homebrew is already installed.
Git Already installed..
Finished installing global libraries.
Downloading demo app.
Cloning into '/Users/gregflurry/Luxonis/depthai'...
remote: Enumerating objects: 21914, done.
remote: Counting objects: 100% (62/62), done.
remote: Compressing objects: 100% (37/37), done.
remote: Total 21914 (delta 34), reused 47 (delta 25), pack-reused 21852 (from 1
Receiving objects: 100% (21914/21914), 248.50 MiB | 2.72 MiB/s, done.
Resolving deltas: 100% (14513/14513), done.
Already on 'main'
Your branch is up to date with 'origin/main'.
Already up to date.
Creating python virtual environment in /Users/gregflurry/Luxonis/venv
/Users/gregflurry/.pyenv/shims/python3
WARNING: pip is configured with locations that require TLS/SSL, however the ssl module in Python is not available.
Requirement already satisfied: pip in /Users/gregflurry/Luxonis/venv/lib/python3.9/site-packages (21.1.1)
WARNING: Retrying (Retry(total=4, connect=None, read=None, redirect=None, status=None)) after connection broken by 'SSLError("Can't connect to HTTPS URL because the SSL module is not available.")': /simple/pip/
WARNING: Retrying (Retry(total=3, connect=None, read=None, redirect=None, status=None)) after connection broken by 'SSLError("Can't connect to HTTPS URL because the SSL module is not available.")': /simple/pip/
WARNING: Retrying (Retry(total=2, connect=None, read=None, redirect=None, status=None)) after connection broken by 'SSLError("Can't connect to HTTPS URL because the SSL module is not available.")': /simple/pip/
WARNING: Retrying (Retry(total=1, connect=None, read=None, redirect=None, status=None)) after connection broken by 'SSLError("Can't connect to HTTPS URL because the SSL module is not available.")': /simple/pip/
WARNING: Retrying (Retry(total=0, connect=None, read=None, redirect=None, status=None)) after connection broken by 'SSLError("Can't connect to HTTPS URL because the SSL module is not available.")': /simple/pip/
Could not fetch URL https://pypi.org/simple/pip/: There was a problem confirming the ssl certificate: HTTPSConnectionPool(host='pypi.org', port=443): Max retries exceeded with url: /simple/pip/ (Caused by SSLError("Can't connect to HTTPS URL because the SSL module is not available.")) - skipping
WARNING: pip is configured with locations that require TLS/SSL, however the ssl module in Python is not available.
Could not fetch URL https://pypi.org/simple/pip/: There was a problem confirming the ssl certificate: HTTPSConnectionPool(host='pypi.org', port=443): Max retries exceeded with url: /simple/pip/ (Caused by SSLError("Can't connect to HTTPS URL because the SSL module is not available.")) - skipping
WARNING: pip is configured with locations that require TLS/SSL, however the ssl module in Python is not available.
WARNING: Retrying (Retry(total=4, connect=None, read=None, redirect=None, status=None)) after connection broken by 'SSLError("Can't connect to HTTPS URL because the SSL module is not available.")': /simple/pyqt5/
WARNING: Retrying (Retry(total=3, connect=None, read=None, redirect=None, status=None)) after connection broken by 'SSLError("Can't connect to HTTPS URL because the SSL module is not available.")': /simple/pyqt5/
WARNING: Retrying (Retry(total=2, connect=None, read=None, redirect=None, status=None)) after connection broken by 'SSLError("Can't connect to HTTPS URL because the SSL module is not available.")': /simple/pyqt5/
WARNING: Retrying (Retry(total=1, connect=None, read=None, redirect=None, status=None)) after connection broken by 'SSLError("Can't connect to HTTPS URL because the SSL module is not available.")': /simple/pyqt5/
WARNING: Retrying (Retry(total=0, connect=None, read=None, redirect=None, status=None)) after connection broken by 'SSLError("Can't connect to HTTPS URL because the SSL module is not available.")': /simple/pyqt5/
Could not fetch URL https://pypi.org/simple/pyqt5/: There was a problem confirming the ssl certificate: HTTPSConnectionPool(host='pypi.org', port=443): Max retries exceeded with url: /simple/pyqt5/ (Caused by SSLError("Can't connect to HTTPS URL because the SSL module is not available.")) - skipping
ERROR: Could not find a version that satisfies the requirement pyqt5 (from versions: none)
ERROR: No matching distribution found for pyqt5
WARNING: pip is configured with locations that require TLS/SSL, however the ssl module in Python is not available.
Could not fetch URL https://pypi.org/simple/pip/: There was a problem confirming the ssl certificate: HTTPSConnectionPool(host='pypi.org', port=443): Max retries exceeded with url: /simple/pip/ (Caused by SSLError("Can't connect to HTTPS URL because the SSL module is not available.")) - skipping

Failed installing depthai. Could be a bug in the installer or unsupported platform. Open a bug report over at https://github.com/luxonis/depthai - exited with status 1 at line 159

Some context: I have had an OAK-D since they first came out. I was successful using it on an Intel-based MacBook, and a Raspberry Pi. Sadly, I had to replace the MacBook with an M1-based MacBook in early 2022. Despite days of trying, with a lot of help from Luxonis, I was never able to get DepthAI to work on the M1 MacBook. I am sure some of the reasons have to do with my modest familiarity with some tools like homebrew and git.

Today, I tried again to install the latest DepthAI on my M1 MacBook, as I was told several months ago it should should succeed. I used the instructions here. I entered the bash command, accepted the default install folder and Python 3.9.5. I got the following response:

This script requires Bash 4.0 or higher. You are using Bash 3.2.57(1)-release. Please upgrade your Bash version.

A bit of searching suggested that makes sense because macOS only comes with 3.2.57(1).

I next searched for ways to upgrade bash. One way suggesting using homebrew,and starting with the command "brew doctor". I got the following response, which confuses me, and I am reluctant to proceed without advice:

Please note that these warnings are just used to help the Homebrew maintainers
with debugging if you file an issue. If everything you use Homebrew for is
working fine: please don't worry or file an issue; just ignore this. Thanks!

Warning: "config" scripts exist outside your system or Homebrew directories.
`./configure` scripts often look for *-config scripts to determine if
software packages are installed, and which additional flags to use when
compiling and linking.

Having additional scripts in your path can confuse software installed via
Homebrew if the config script overrides a system or Homebrew-provided
script of the same name. We found the following "config" scripts:
  /Users/gregflurry/.pyenv/shims/python3.9-config
  /Users/gregflurry/.pyenv/shims/python-config
  /Users/gregflurry/.pyenv/shims/python3-config

Warning: You are using macOS 14.
We do not provide support for this pre-release version.
You will encounter build failures with some formulae.
Please create pull requests instead of asking for help on Homebrew's GitHub,
Twitter or any other official channels. You are responsible for resolving
any issues you experience while you are running this
pre-release version.


Warning: Ruby version 2.6.10 is unsupported on macOS 14. Homebrew
is developed and tested on Ruby 2.6.9, and may not work correctly
on other Rubies. Patches are accepted as long as they don't cause breakage
on supported Rubies.

Warning: Your Xcode is configured with an invalid path.
You should change it to the correct path:
  sudo xcode-select --switch /Applications/Xcode.app

I am not sure what to do about the Python stuff. I suspect the multiple scripts are an artifact of my futile attempts at installation in 2022.

I am on macOS 14.6.1, but macOS is certainly not pre-release, so I suspect the warning is bogus. Perhaps I need to update homebrew.

I don't have a clue what to do about the Ruby warning or Xcode warning.

Thanks for any assistance.

In a post a few weeks ago someone asked if "Is RAE dead?" The answer, to me, was ambiguous. I believe I have evidence suggesting that RAE is not supported, and in fact may be dead. I hope someone from Luxonis can clarify the company's expectations with respect to RAE.

What follows is my admittedly circumstantial evidence.

I experimented with an OAK-D during almost all of 2022 and into early 2023. I felt the documentation, sample programs, etc., bordered on great. During that time, I found support in most cases to be timely and close to superb. There was lots of activity from a community of users. That experience gave me confidence I'd see the same with RAE.

I started experimenting with my RAE in October 2023. At the time, I found little documentation, and most of what I found was incomplete. I could get my RAE to do next to nothing. Support was present, but delayed and sometimes unhelpful. Due to other pressing projects, I put RAE on a shelf hoping for maturity.

I started with RAE again in March 2024. Not much had changed overall, but I got a little support and I was able to get enough working that I could see camera output and actually add some logging statements to an app. There seemed to be only handful of RAE users. Somewhere I read that Luxonis had concluded the ROS2 interfaces were too hard for users and was defining a Python SDK to make programming much easier. That sounded attractive, and with other projects waiting, I again put RAE on a shelf hoping for maturity.

I started with RAE the third time in August 2024. Not much had changed overall, but there was a bit of documentation on the SDK, and I was able to get enough help from two folks (one of whom may be from Luxonis) that I got the Car app to work, even in follow-me mode (which as pretty nice). That said, RAE was unreliable (sometimes the app would "hang"), and the SDK documentation was incomplete, including no sample apps and almost nothing about how you actually code, install, run, or debug an SDK-based app.

So, I am to the point of putting RAE on the shelf again. But unless Luxonis plans to make dramatic improvements to documentation and support, it feels meaningless to ever take it off the shelf again.

Can someone from Luxonis address my concerns about support?

Thanks.

IvoGermann IvoGermann I've continued to work on this in between other projects. I've made some progress thanks to you and @MartinMotycka. I now have I think achieved step 5 in your list; however, I've done it with the new Hub version, which seems to allow me to install/develop a somewhat different set of code that is called "RAE Starter"; here is what I see after installing it (I gave it the name 'Maybe RAE Starter'), assigning rae to it, opening up Studio, then starting it, and poking around in the source:

There are a couple of things worth noting. The code is different; for example, I see no images in the source. I note that the start_ros.sh contains the code in your step 6 and your step 7.

At this point I am still quite confused. The rae-sdk page mentions "ghcr.io/luxonis/rae-ros:v0.2.2-humble" as needed in a "robotapp.toml". It is not in that file for the app (see the right side) of the screenshot). Further, in steps 8 and 9 you are using ros2 commands to run stuff. I know next to nothing about ros(2), but I'd expect to run the rae Python SDK using a python3 command.

Any elaboration would be much appreciated.

MartinMotycka Thanks. Your description has helped me make progress!

First, let me describe my experiences with Run installations. I already had the Car App installed and rae assigned to it. I started it and then clicked Local Frontend. I got a warning Safari message about potentially malicious website. I pushed through to open the page and got a "frontend" with some controls. I was able to move the rae and the LEDs and the speaker. I was even able to get it to "follow me"! [as an amusing aside, while showing my wife, rae started to follow her :-)]. Progress! That said, apparently the wifi connection was poor or something, as the frontend lost connection a few times.

I then installed in Run mode the Default App. It too ran, but apparently the same front end. This in retrospect makes sense, but I realize now that I'd seen the front end before, I just did not realize it was associated with multiple apps. Oh well.

Second, I next tried to install a Develop app. I was successful in installing a Streaming template, getting it started, and viewing the video streams. I also opened up the code editor. Great! All thanks to your guidance. That said, I now remember I was at this point back in March, but activities seem to be a little bit different, and the overall system seems a bit faster and a bit more stable.

That said, I think the code I'm looking at is based on ROS2. I really want to build apps on the Python SDK. I seem to be no closer to that. Can you give me any guidance on how to do that?

MartinMotycka I have in the past created a new app using the <> Develop option. That did give me access to the code editor. It did me little good due to the little documentation on what the code means.

Using the latest Hub version, I installed an app via the Run option. I could find no magic incantation that allowed me to indicate I wanted to develop it, use the code editor, or anything of the sort either during or after installation. Do you have any example screen shots?

Further, I really want to develop using the new Python SDK. I can find nothing that shows how to do that. I think there are example apps on GitHub, but I have no idea how to get them onto the rae, run them, or leverage the web UI that some of them seem to include.

Relative to a UI, I have the Car App installed. I cannot find any way to use it. There must be some UI somewhere, likely a web based. Do you have any hints about where to look for the UI?

Thanks for the advice.

IvoGermann I started the old Hub version, and it reports that I am running OS 1.14. The mystery continues.

IvoGermann Thank you so very much for your help!

I confirmed that my device is running the agent 23.223.1855, per steps 1 and 2. I added a screenshot of what I think is the equivalent of your screenshot for step 2. I am apparently using an updated Hub, and it looks different. First, I don't see the error message, and second, my device is named 'rae' rather than 'keembay', though I may have done something 6 months ago to change that. .

When I click on the device I get a a very different view, as shown in the screenshot below. There is no tab called "Perception Apps"!

Using the URL for the legacy Hub in step 2, I get something quite similar to your 3 screenshot. However, the only app I see is one I downloaded two days ago. There is no RaeStarterApp pre-installed.

When I click on Install App I get the following popup.

When I click Run I see the following popup. Again, no RaeStarterApp.

So, once again, not sure what is going on, nor is there documentation enough to begin to understand the situation.

In any case, I went back to the basic Perception Apps tab and clicked the Car App. The following screenshot shows the results. I can find nothing that lets me open Studio to view/edit the code! Just to make sure I'd not missed something, I tried the same thing in the new Hub, and again, no way that I can see to open Studio.

Once again, I'm stymied. That said, again, I very much appreciate your help.

@IvoGermann I am trying to achieve what you appear to have achieved. despite being a software developer for 40 years, I don't apparently have enough skills/context/something to do so. I apologize upfront, but I hope you can help me. If not, no worries.

What is the RaeStarterApp? Where did you find it? How did you install it on the device?

Where did you find robotapp.toml? I don't see one on my device.

When you say "I pressed the button to apply this change", what button do you mean? Maybe the Restart in the Hub?

When you say "opened python3 and tried 'import rae_*sdk -> works fine" does that mean you ssh'ed into your device and issued the command python3?

You say "it is important to note that the RaeStarter app has to be started first". Do you think that any app can be running? So far I've only installed and a couple of pre-built apps using the Hub. In these cases, it seems that only one app can be using the device at a time, but I confess I've not tested that assertion.

Any other advice you can give me would greatly help. I've become very frustrated trying to do anything with my rae.

Thanks.

<RANT>I started working with my rae again yesterday. I was disappointed to find that documentation is no better than 6 months ago, terse and mostly incomplete. What I did find is that to achieve my goal (app development) I needed to install the latest firmware. I was able to do that, but of course, it took three attempts to do so successfully (possibly because of wifi issues, but not clear). Even then, the default app that was supposed to be uninstalling (for around 2 hours) was still uninstalling after the firmware update (23.223.1855). So, it was time to reset the device (for the 6th or 7th time). I did so (with the small switch and hole it took 5 minutes just to get a reset started). The reset completed and the device was initializing ... and the battery died. After charging overnight, today I powered up the device and reconnected to the hub. I then installed the RAE - Car app; it took at least 10 minutes to install.</RANT>

Once the app was installed, I assigned the device to it. I started the app and once it was running, I went to the Live View and was able to see video from both the front and rear cameras. But, I can't do anything else! I'd expect a "car app" to have some sort of controls to drive it. I can't find how to engage such controls. Help!

So ... what I really want to do is to use the Python SDK. The SDK documentation says "Specify correct image in robotapp.toml file. As of the moment of writing this guide, latest image is ghcr.io/luxonis/rae-ros:v0.2.2-humble". That statement means nothing to me; I clearly don't have enough context to begin to interpret, much less implement it. Questions:

• where do I find robotapp.toml? a bit of searching says in root, but my rae had no such file in root; also, searching suggests the robotapp.toml file is somewhat complex; am I supposed to be able to create one from scratch, or is there a sample relevant to rae somewhere?

• what is the "image"? where does it go in the robotapp.toml file? what elso is needed for a custom app?

• where to I write the Python script? in the Hub or in a coding tool? how do I get the app downloaded to the device?

• is there anything anywhere that describes the complete process of creating a custom app based on the SDK?

• are there any examples of apps that use the SDK?

IvoGermann By some miracle I could get to Google Drive and check. I think the file is as public as I ca make it. Here is the like Let me know what happens.

IvoGermann I thought I made the second link public. I am traveling with limited access so it could be Tuesday before I can investigate.

Luxonis-Vlad Thanks for the information. Reading between the lines, perhaps over-pessimistically, I see large changes in architecture, significant development, significant testing, and a significant documentation effort. Based on what I've experienced in my career as a SW and HW developer (tho I'm long retired), it could be many months before "the new RAE" is ready for those of us who want an underlying platform that is stable, works reliably with reasonable performance, facilitates fast iterative development, and is well-documented, so we can do what we want to do. Luxonis, in my opinion, achieved that quite well with the original OAK-D, which is why I was excited to get a RAE. Unfortunately, my experience with RAE so far has been quite frustrating. It is unstable, not well documented, and doing the simplest, mundane things is like paddling through molasses.

I first started trying to use RAE 5 months ago, and managed to get it into a state from which I recently found could be fixed only with a factory reset. Other tasks caused me to put RAE on the shelf. Now, 5 months later, I've spent over a week and have made almost no progress in terms of actually developing my own application. It seems to me I might be better off putting RAE back on the shelf for another 5 months, hoping the new architecture will be usable at that time. Does that make sense?

MartinMotycka 1. What is the proper version of the Default app to install; the candidates are 1.2.1 and 2.0.0; I have had 1.2.1 work, but I think I've tried twice to download 2.0.0 and it has failed both times

• I need a lot more help with the USB-C cable stuff. I connected the RAE to my macBook via USB-C and used ssh root@192.168.197.55 with no password and that worked. I can look at the file system But now what? How would I download an app? How would I run an app? I can see various processes running, but I would not know which to kill if the need arose.

• Regarding downloads, I have zero experience with Docker. How do I know if I have to download it? Would it be possible for the Hub to tell the user how big the download is going to be? Also, the Hub indicates my RAE only has 0.4 GB of "disk" available. There appears to be no way I could install an app several GB in size. Am I missing something? One thing I have found with RAE is patience is a must; everything seems to take a lot longer than one might expect.

• Regarding the RAE agent, if you notice in the document I posted, you will find that I again experienced the Hub trying to update the agent. Why is that happening?

• Regarding lag time, I started informally timing the lag, and it seems to be less than 2 seconds, so I guess my experience is expected.

• And finally, the Default app. I have spent the last 2 hours trying to achieve what you suggest is possible and so far I am not even close. I first had to install the Default app; I had to use 1.2.1 because I could not ever get 2.0.0 to install or even download. Using the Hub, I was able to assign the device to the Default app, and was able to successfully run the app. I noticed in your post that the app was stopped, so I stopped the app; I think, anyway, because the LEDs stayed on. I fired up the agent to https://192.168.1.69:9010 (the WIFI IP address of my RAE) and when I clicked "Perception Apps" I saw something similar to you image, though my Test app was on the left and the Default app was on the right. I clicked "Open Frontend" and got a black page; I waited for a few minutes and concluded it did not work. I shutdown the device and started again and the Default app started. I tried the agent again, with the app running, and got the same results -- a blank page for a few minutes. I tried again and again, but something changed and all I got was a "twirly circle pair". This happened over several power cycles, so I gave up. Can you give me the magic incantation to get to point where I can use the Default app to manipulate the motors, LEDs, etc?

Thanks for the help.

MartinMotycka Once again, I much appreciate the help. I do have some followup questions, as I still have many points of confusion.

Before I get to the followup questions, I tried something new today, creating an app. I was moderately successful, but it was yet another rather painful experience. I thought it would be helpful for me, and perhaps others, to document what I did, discovered, failed at, etc. I humbly suggest that the level of detail I've tried to include could be helpful to those, like myself, that are just getting started. The document includes some more questions/mysteries, and also includes some helpful info (an appendix) on resetting and reconnection to the hub. I'd be willing to update/polish that if it would help "the cause". Sadly it is too big to attach; this is the Google drive link.

MartinMotycka I mistakenly posted my attempt to execute your suggestions to the general thread, so I'm replying to you to make sure you see it. Thanks.

Here is a report at my attempt to do what you suggested. I don't know if this matters, but I am running on a MacBook Pro with Apple silicon, using macOS, and Safari as the browser.

I attempted to upgrade the OS. This was a bit tricky, as I had to do a bit of searching to find the user ID/PW to use to SSH into rae (root and no password). I executed the command shown in your link. Based on the feedback from the command execution, it was not obvious if it had failed or succeeded. I must add that I did this over wifi, because I could not determine how to do it with the USB cable attached to my MacBook. Is that a better way?

I subsequently opened the Hub and deleted the two outage devices. I peeked at the online device, and it indicated OS 1.14, so I guess that worked. I then deleted the online device from the Hub.

I then attempted to reset rae. I found it very difficult to press the reset button, and I could not detect any obvious action caused by reset. I then attempted to connect to the Hub again, but failed. I pinged the device at its original IP address and that succeeded, so I figured I actually did not reset rae.

I tried again to reset. Again, it was very difficult (bad eye/hand coordination?). While the LEDs on the bottom of the device never changed, I did eventually notice that the device display went blank and then eventually said rae. I pinged again and got nothing.

I again attempted to connect to the Hub. It took a bit, but then seemed to connect. I attempted to install the default app (2.0.0). The Hub said the download was happening, then eventually showed a warning about taking a long time and network speed. I could ping the device on its new IP address.

The warning also said something about restarting the download. I could find no way to do that in any UI in the Hub, so I simply turned off the device. Eventually the Hub showed the device status as Outage.

After a couple of hours, I turned on the device. Unlike what I remember seeing earlier, the lights did not ever come on. Meanwhile, I went to the Hub. More anomalous behavior. In the main area, a dialog box said "Upgrade in Progress...". It also said "Waiting for device to connect...". A progress bar indicated something was happening, for moving from 23-something to 24-something. Shortly after, I noticed the Hub indicating the device was online (but still no indication of that from the device). Then in the dialog, a sub-box said "Upgrade is taking longer than usual...". The progress bar reached what I'd interpret as done. I waited maybe 10 minutes and nothing changed. The Hub indicated online, but also waiting to connect; progress bar complete, but still upgrade in progress.

I refreshed the Hub, and got the upgrade dialog again. This time the Hub indicated the device was "Outage". I noticed this time the dialog "Upgrade is taking longer than usual..." popped up just as the progress bar reached done. Having lost my patience, I closed the browser window and shut down the device.

I turned on the device again; a blue light flashed when I pressed the power button. But after several minutes, all I see is rae in the display, no lights. I opened a new browser window and started the Hub. Nothing on the device changed and the Hub reported 1 device online. In the Hub devices, the rae showed up as online. I could also ping the device, but the response times were very long (I'm on a wifi network of at least 100 MB fed by a 1 GB internet connection). I clicked the device and saw an indication that an app install had failed.

I uninstalled the failed app; I found how to do that by accident. I then tried to install v1.2.1. The Hub indicated the app got downloaded, and then said the app was initializing -- and the LEDs came on! And the Hub then indicated the app was running. I could see video from both the front and back cameras! So far so good -- but it was very painful to get to this point! What the heck was all that "upgrade" stuff?

I stopped the app. The lights went out. I guess that is correct. I shutdown the device. I noticed that the Hub, even after 3-4 minutes still indicates the device is online, but finally did indicate Outage. I also noticed that in the device information, a column named Agent said "23.223.1855 / RAE". That looks like the number that was trying to get upgraded. So clearly my device is still not all the way up to date. Why was I finally able to get past the attempt to upgrade? Why did the upgrade fail?

I turned the device on again. I waited a few minutes for the Hub to see it and indicate Online. When it did come Online, the LEDs turned on. When I looked at the device details there was no attempt to upgrade the agent. Why not? Is there a way to force that? Should I do so?

I noticed the app was running (I guess the LEDs indicate that). I was able once again to look at video from both the front and back cameras. As I'm sure has been reported, there was a significant lag from real time, and a slow frame rate.

So I guess your suggestions have solved my immediate problem. I can now use some version of the default app somewhat reliably. Clearly my experience indicates there are still some bugs or coin-flips in the device and maybe Hub. My experience has raised a number of questions in the way-too-long dialog above. But there are many more:

• why is everything so slow? could there be more status indication, either in the the Hub or on the device display?

• what version of the default app should I be using? from above, 2.0.0 would not even download for me, at least on the first attempt

• what of the prebuilt apps should work? how do I find out what they actually do?

• can the device only have on app at a time? this is suggested by the fact that a device has to be assigned to an app at some point, tho the only time that may work is when the app gets installed

• if multiple apps can be installed is there a proper way to reassign the device?

• what is the proper shutdown procedure? should I first stop any running app? should I shutdown using the power button on the device? Or should I use the shutdown button I discovered in the Hub?

• in the Finishing Up section, there is a mention of "direct control". what does that mean? what can I control?

• when I enter the URL, I see a web page saying "Robot Hub Agent". is that the same agent that was trying to get upgraded? I see an icon for the Default app, and for "Login to manage". is the latter the same as direct control? I am a bit hesitant to click anything for fear of getting myself messed up again.

Thanks again for the help. I really do appreciate it. But I have to say, that for something that was supposed to be for everyone/anyone, I am not sure that has been achieved. But is does seem better now than 5 months ago, even with all the remaining mysteries.