Im trying to get the Assistant Python library demos running. I download my client credentials file, run the auth/consent tool as follows:
(env) $ google-oauthlib-tool --scope https://www.googleapis.com/auth/assistant-sdk-prototype --scope https://www.googleapis.com/auth/gcm --save --headless --client-secrets client_secret_XXXXX.json
The tool shows a URL that I visit and click through - the subsequent web shows me an auth string that I copy and paste back into the waiting terminal session.
(the tool tells me that it updated a .config/ file)
OK I then run...
(env) $ googlesamples-assistant-hotword --project-id ${project} --device-model-id ${model-id}
which results in a segfault:
device_model_id: dawnstar-226520-dawnstar-test-rig-jcsgzj
device_id: 4BAE41228DA8890AB0DDE6FD6992F5BF
Segmentation fault (core dumped)
(env) $
I have confirmed the project ID, model ID and that the Assistant API is enabled
Any help appreciated
The problem is related to a change from Google Assistant Library to Google Assistant Service. The Google Assistant Library is deprecated.
As shown in this project page issue, the issue affect to the components still based on Google Assistant Library, like googlesamples-assistant-hotword.
Other resources, like googlesamples-assistant-pushtotalk now are working over gRPC and this is the new way to use this service.
There is also a working task to improve the system over gRPC.
Related
It seems that I'm facing a delemma. I need to run kvaser_interface ROS1 (noetic, 1.15.14) node. I'm using this repo: https://github.com/astuff/kvaser_interface
I follow this link to install kvaser CAN/USB driver on my ubuntu 20.04.4. If I don't install CANlib or linuxcan, I can have my CAN network working (can0 is listed, candump and cansend works). But kvaser_interface needs a library from linuxcan, hence gives me error:
... kvaser_can_bridge: error while loading shared libraries: libcanlib.so.1: cannot open shared object file: No such file or directory
Apparently kvaser_interface node requires CANlib. So I installed linuxcan and CANlib using:
sudo make install
kvaser_interface runs without a problem. But can0 isn't working anymore.
ip link
doesn't show can0. Also
sudo ip link set can0 type can bitrate 500000
gives
Cannot find device "can0"
This is because linuxcan blacklists socketCAN. It seems that I'm facing a delimma. Also I'm not clear what role CANlib is playing here. Any suggests will help. Thank you.
I have not tested this yet, but you can go into the
/etc/modprobe.d/kvaser.conf
and sudo edit the file to delete the blacklisted modules:
https://askubuntu.com/questions/110341/how-to-blacklist-kernel-modules
On day 28 they put an update that google-assistant-sdk now goes with the command googlesamples-assistant-pushtotalk. Previously I went with googlesamples-assistant-hotword. The previous command allowed you to say ok google and I was beginning to hear you. Now you have to give it to enter. Does anyone know how I can do to make it like before I say google google start listening to the assistant?
Install the google service like indicated here:
https://developers.google.com/assistant/sdk/guides/service/python/
Then install then in the venv google library:
source env/bin/activate
pip install google-assistant-library
Then you can run google assistant with the hotword:
googlesamples-assistant-hotword --project-id project --device-model-id model
I am completing this tutorial and am at the part where you download the code for the tutorial. The request we send to Github is:
wget https://github.com/GoogleCloudPlatform/cloudml-samples/archive/master.zip
I understand that this downloads archive to GCP, and I can see the files in the Cloud shell, but is there a way to see the files through the Google Console GUI? I would like to browse the files I have downloaded to understand their structure better.
By clicking on the pencil icon on the top right corner, the Cloud Shell Code editor will pop.
Quoting the documentation:
"The built-in code editor is based on Orion. You can use the code
editor to browse file directories as well as view and edit files, with
continued access to the Cloud Shell. The code editor is available by
default with every Cloud Shell instance."
You can find more info here: https://cloud.google.com/shell/docs/features#code_editor
If you prefer to use the command line to view files, you can install and run the tree Unix CLI command 1 and run it in Cloud Shell to list contents of directories in a tree-like format.
install tree => $ sudo apt-get install tree
run it => $ tree ./ -h --filelimit 4
-h will show human readable size of files/directories
and you can use --filelimit to set the maximum number of directories to descent within the list.
Use $ man tree to see the available parameters for the command, or check the man online documentation here: https://linux.die.net/man/1/tree
Hello Im trying to create initial flash/build for IoT development following this tutorial https://developer.android.com/things/hardware/imx7d.html#flashing_the_image
Im sorry if my questions is too broad, this is my first IoT attempt, but it seems to me like I have a wrong setup, beacuse Im constantly running into new errors.
Im stuck at step 2.4 Execute the flash-all.sh. Running
sudo ./flash-all.sh
I got this in my logs:
./flash-all.sh: line 52: ./u-boot.imx: Permission denied
If I change permissons
chmod 777 u-boot.imx
I got
./flash-all.sh: line 52: ./u-boot.imx: cannot execute binary file:
Exec format error
I already solved several other issues which werent described in tutorial, including
I have to run script as sudo, otherwise I got
< waiting for any device >
I had to rewrite fastboot command to $(which fastboot) inside flash-all.sh (same with flash and bootloader), otherwise commands are unknown even thought I added them to PATH
I am using
ubuntu 16.14,
android studio with installed sdk 26
Pico Pro Maker Kit with Pico i.MX7 Dual Development Board
What am I doing wrong?
I had to rewrite fastboot command to $(which fastboot) inside flash-all.sh (same with flash and bootloader), otherwise commands are unknown even thought I added them to PATH
This seems like it might be the root of the problem, as somehow the subsequent lines for each command are not being parsed as arguments for fastboot, but rather as their own executable commands.
You also shouldn't need to run the script with sudo. This might be why you can run which fastboot successfully (which would indicate it's in your PATH), but the script cannot see this.
So I am a complete noob and have no idea how to even navigate these forum as it all sound like a foreign language. My issue is I recent was accepted to use Google page speed service and cannot even get the first step out of the way which is installing Cloud SDK.
Here is step by step what I did:
Extracted all from zip file
Clicked to launch the Batched file (upon clicking the command prompt black screen flashed in the corner, no instructions or anything else)
When attempting to authenticate google cloud platform by clicking windows key + r. THen typing in 'gcloud auth login', I get a message that says, "cannot find 'gcloud'.....".
When you run install.bat, this is an interactive process that should not exit immediately. I'm guessing you don't have Python installed? Install Python 2.7 and try again.
After the installer runs, it will print directions on how to add gcloud to your PATH. In general, though, gcloud is a command line tool that you shouldn't really run using windows+r. It's best to open a cmd prompt so you can more easily run multiple commands.