Shell utility to configure Cognicept tools.
Project description
COGNICEPT SHELL
This is a shell utility to manage Cognicept tools.
Installation
Dependencies
You need:
- Python 3
- Python 3 PIP
openssh-server
: (Optional) in order to enable terminal function for Kriya
Install:
sudo apt-get install python3 python3-pip
Package installation
To install the package locally, run:
pip3 install -e <path-to-the-repo>
To install from Python Package Index (PyPI), run:
pip3 install cognicept-shell
To verify installation, try to run
cognicept -h
If you get cognicept: command not found
error, make sure that ~/.local/bin/
is in your $PATH
. You could run this:
export PATH=$PATH:/home/$USER/.local/bin/
and add it to your .bashrc
file.
Autocomplete
To set-up autcomplete when using the cognicept shell, run:
cognicept config --autocomplete
Restart the terminal and autocomplete should be functioning
This will setup python argcomplete globally and may affect other packages. If argcomplete has been setup globally for other project, this setup is not required.
This command essentially calls the function activate-global-python-argcomplete
More information about this can be found in https://pypi.org/project/argcomplete/#activating-global-completion
Usage
For details on usage, use
cognicept -h
Commands
config
: Configure Cognicept tools
cognicept-shell
and Cognicept agents are configured in runtime.env
file typically placed in ~/.cognicept/runtime.env
. The file defines the docker environment used by the Cognicept agents.
This command allows to inspect and modify the configuration file. You can use parameter --path
to modify the path to the Cognicept config directory.
To print full configuration, run:
cognicept config --read
To add new configuration parameter (or modify single value), run:
cognicept config --add
Variables used by cognicept-shell
:
COGNICEPT_ACCESS_KEY
COGNICEPT_API_URI
COGNICEPT_USER_API_URI
COG_AGENT_CONTAINERS
COG_AGENT_IMAGES
COG_EXTRA_IMAGES
: images to pull from Cognicept and general docker image repositoriesCOG_ORBITTY_ENABLED
COG_ORBITTY_IMAGE
AWS_ACCESS_KEY_ID
,AWS_SECRET_ACCESS_KEY
,AWS_SESSION_TOKEN
: temporary AWS credentials modified bykeyrotate
and needed forupdate
COG_ENABLE_SSH
COG_ENABLE_SSH_KEY_AUTH
COG_ENABLE_AUTOMATIC_SSH
COG_SSH_DEFAULT_USER
To setup ssh
access from remote_intervention_agent
to the host machine, run:
cognicept config --ssh
It will generate ssh keys in ~/.cognicept/ssh/
and configures COG_ENABLE_SSH
, COG_ENABLE_SSH_KEY_AUTH
, COG_ENABLE_AUTOMATIC_SSH
, COG_SSH_DEFAULT_USER
according to the user preferences. If COG_ENABLE_AUTOMATIC_SSH
is enabled, public key is copied into ~/.ssh/authorized_hosts
; sudo
access is requested to perform this action.
status
: Get status of Cognicept agents
cognicept status
Prints status of agents and other containers managed by cognicept-shell
. Possible values:
- "ONLINE": Everything fine
- "CONTAINER NOT FOUND": Agent/container is not initiated
- "OFFLINE": Container stopped
- "Error": Error accessing API
- "ERROR": Detected error, needs inspection
- "NOT INITIALIZED": Agent is in init state
- "STALE": Agent didn't update within last minute
update
: Update Cognicept tools
Updates images for agents and tools specified in COG_AGENT_IMAGES
and COG_EXTRA_IMAGES
config variables. It requires temporary credentials to be valid. For updates to take effect, containers need to be restarted with restart
command.
To run update in detached mode:
cognicept update -d
In detached mode, printing of update statuses in muted. Update is run in a seperate process. Update will continue to run even if terminal session is closed.
lastevent
: Display last event log reported by Cognicept agent
Displays last event saved by cgs_diagnostics_agent
from ~/.cognicept/logs
.
start
/stop
/restart
: start/stop/restart cognicept agents
These commands are used to start/stop/restart containers specified in COG_AGENT_CONTAINERS
/COG_AGENT_IMAGES
. Certain container names are reserved for Cognicept agents and are preconfigured:
cgs_diagnostics_agent
remote_intervention_agent
kriya_watchdog
cgs_diagnostics_ecs_api
cgs_diagnostics_streamer_api
cgs_bagger_server
Any agent name or image type can be put in the list as long as default command for the image is specified. All containers are started in host
network mode.
Following are examples for using start
. stop
follows same API as start
. restart
first calls stop
and then stop
.
To start all listed agents COG_AGENT_CONTAINERS
don't specify any argument:
cognicept start
To start all agents:
cognicept start --agents
To start specific agents:
cognicept start remote_intervention_agent cgs_diagnostics_agent
To run restart in detached mode:
cognicept restart -d
Detached mode is particularly useful for restarting a cognicept agent remotely. In detached mode, progress of restart will not printed.
keyrotate
: Rotate Cognicept cloud keys
Updates temporary AWS credentials (AWS_ACCESS_KEY_ID
, AWS_SECRET_ACCESS_KEY
, AWS_SESSION_TOKEN
) using COGNICEPT_ACCESS_KEY
from COGNICEPT_API_URI
. The validity of the credentials is 12 hours. Internet access with HTTPS allowed is needed.
orbitty
: Run Orbitty
TODO
record
: Manage rosbag recording session to start/stop/pause/resume/get status/record all topics
These commands are used to start/stop/pause/resume/get status/record all topics in a rosbag. All the operations are done by running an exec
command inside the cgs_bagger_server
container. All bags are typically placed in the bags
folder of the Cognicept configuration directory: ~/.cognicept/bags/
To start start a recording session list all topics to be recorded separated by a space and with their /
prefix:
cognicept record --start /odom /cmd_vel /rosout_agg /tf /tf_static
To record ALL topics use --all
:
NOTE: rosbags tend to become quite big quite fast especially if there are camera topics. If there is less than 1 GB disk space, the recording self terminates.
cognicept record --all
To stop a recording session:
cognicept record --stop
Alternatively, the latest bag recording can be stopped and automatically pushed to the Cognicept cloud with a single command by providing autopush
value:
cognicept record --stop autopush
To pause a recording session:
cognicept record --pause
To resume a recording session:
cognicept record --resume
To get the status of the current recording:
cognicept record --status
Status can be Ready
, Started
, Stopped
, Paused
and Resumed
.
NOTE: A note on the retry mechanisms built in to improve reliability. If the retry mechanism fails, both cognicept record --start
and cognicept record --all
will automatically restart the cgs_bagger_server
. A new recording session needs to be started if this happens. All other commands will just inform that the retry has failed since these are not essential operations for bag recording.
push
: Push data to Cognicept cloud
These commands are used to push data on to the Cognicept cloud.
To push a bag file, specify the name of the bag file:
cognicept push --bag sample_recording.bag
NOTE: Only bags in the dedicated Cognicept configuration bags
directory, usually ~/.cognicept/bags
, will be detected and uploaded. Bag files in any other location need to be moved to this directory before upload. Bag files generated using the cognicept record
feature will be automatically placed in the Cognicept configuration bags
directory.
As a shorthand, you can push the latest bag file recording by not specifying any argument. This will automatically look for the latest bag file by create time and upload it:
cognicept push --bag
init
: Initiates the runtime.env file with values from the COGNICEPT_USER_API_URI
To initiate a runtime.env file, specify the robot id and organisation id and then provide credentials when prompted
cognicept init --robot_id sample_robot --org_id sample_org
Username: sample_username
Password: sample_password (masked)
Building
Tests
cognicept-shell
is using pytest
as the test framework. Make sure you install manually:
pip3 install pytest pytest-cov cli_test_helpers mock
To run tests, execute:
pytest --cov=cogniceptshell tests
Output will look like this:
user@computer:~/cognicept-shell$ pytest --cov=cogniceptshell tests
============================= test session starts ==============================
platform linux -- Python 3.8.5, pytest-5.4.3, py-1.9.0, pluggy-0.13.1
rootdir: /home/jakub/code/cognicept-shell
plugins: cov-2.10.0
collected 50 items
tests/functional/test_config.py ........s [ 18%]
tests/functional/test_push.py .......... [ 38%]
tests/functional/test_record.py .......... [ 58%]
tests/unit/test_config.py .... [ 66%]
tests/unit/test_keyrotate.py ... [ 72%]
tests/unit/test_lifecycle.py .............. [100%]
----------- coverage: platform linux, python 3.8.5-final-0 -----------
Name Stmts Miss Cover
--------------------------------------------------------
cogniceptshell/__init__.py 2 0 100%
cogniceptshell/agent_life_cycle.py 319 117 63%
cogniceptshell/common.py 15 5 67%
cogniceptshell/configuration.py 189 53 72%
cogniceptshell/interface.py 79 2 97%
cogniceptshell/pusher.py 104 11 89%
cogniceptshell/rosbag_record.py 90 13 86%
--------------------------------------------------------
TOTAL 798 201 75%
======================== 49 passed, 1 skipped in 28.59s ========================
Build
To build the PyPI package, run:
python3 setup.py sdist bdist_wheel
This will generate the build files.
Upload
To upload the dev package, run:
python3 -m twine upload --repository testpypi dist/* --verbose
To upload the prod package, run:
python3 -m twine upload dist/* --verbose
Contribution
Please follow the successful branching model. The naming of branches follows:
- Feature branch: /feature/name-of-the-feature
- Bug fix branch: /fix/name-of-the-bug
- Release branch: /release/name-of-the-release
Version history
-
1.3 [21/10/2021]
- Autocomplete for basic commands and agent names
-
1.2 [29/6/2021]
- Revamp credentials for AWS
- Added robot init command
- Revamped CI/CD
- Added detached update/restart
- Removed datadog integration
- Added kriya watchdog run configuration
- Added handling of the docker permission error
- Added run configuration to support bunch reporting
-
1.1 []
- Limit agent logs to 5MB
-
1.0.3 []
- Change for cognicept update command to fetch cloud credentials from cognicept backend
-
1.0 [15/12/2020]
- Added
start
command - Start/Stop/Restart agents separately as parameter of command
- Migrated to new credential management using
COGNICEPT_ACCESS_KEY
- Added
keyrotate
command - Added unit and functional tests
- Added
record
andpush
commands for management of rosbags - Added support for Python 3.5
- Added
lastevent
command to read last event - Added ssh configuration for
remote_intervention_agent
- Added
-
0.1 [10/6/2020]
- First version of the CLI utility able to configure, restart, and update agents
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for cognicept_shell-1.3.0-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 819db7d573142c4f67f7444b6b8a8d420897d2c1902269fb3c6ad251509198ac |
|
MD5 | a46b6c51a88521d299af455c03283b01 |
|
BLAKE2b-256 | f91817eb384d066bc4ba767538bcc905d4c79977ac3b156f0f3a06651ca7e5c7 |