-
Notifications
You must be signed in to change notification settings - Fork 2
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Initial draft of linux instructions and removal of internal instructions
- Loading branch information
1 parent
c01d1f5
commit 3d118b8
Showing
1 changed file
with
38 additions
and
85 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -1,4 +1,7 @@ | ||
| ## Setup | ||
| ## What is Gooey Server? | ||
| Gooey Server is a no-code/low-code AI orchestration engine. It allows users to discover, customize, and deploy AI "recipes" using the best of private and open source AI, all using a single API with a single auth token. Recipes are workflows that incorporate various models to accomplish a task; they are designed to be highly customizable and shareable. | ||
|
|
||
| ## Setup (Mac) | ||
|
|
||
| * Install [pyenv](https://github.com/pyenv/pyenv) & install the same python version as in our [Dockerfile](Dockerfile) | ||
| * Install [poetry](https://python-poetry.org/docs/) | ||
|
|
@@ -14,13 +17,46 @@ | |
| * Run `./manage.py migrate` | ||
| * Install the zbar library (`brew install zbar`) | ||
|
|
||
| ## Setup (Linux) | ||
| * Install [pyenv](https://github.com/pyenv/pyenv) & install the same python version as in our [Dockerfile](Dockerfile) (currently Python 3.10) | ||
| - `curl https://pyenv.run | bash` | ||
| * Install [poetry](https://python-poetry.org/docs/) | ||
| - This is likely available in your distro's package repos. | ||
| * Clone the gooey-server repository: | ||
| - `git clone https://github.com/GooeyAI/gooey-server.git` | ||
| * If you want to use the web application frontend, you must clone that repo as well, in the same directory as gooey-server: | ||
| - `git clone https://github.com/GooeyAI/gooey-ui` | ||
| * Create and activate a virtualenv using `poetry shell` | ||
| * Install dependencies using `poetry install --with dev` | ||
| - Note: you may have to remove `package-mode=false` on line 7 of `pyproject.toml` | ||
| * Install redis, rabbitmq-server, and postgresql 15 using your distro's package manager. | ||
| * Enable these services as background services using `sudo systemctl enable --now redis rabbitmq-server postgresql` | ||
| * Configure Postgres to ensure that password authentication is enabled for the gooey user | ||
| - open the pg_hba.conf file in a text editor. On Linux, by default, it is usually located either at ```/etc/postgresql/<version>/main/``` or ```/var/lib/pgsql/<version>/data/``` | ||
| - add/edit the file so that there are lines at the bottom that looks like this: | ||
| ``` | ||
| local all gooey md5 | ||
| host all gooey md5 | ||
| ``` | ||
| - restart postgresql using ```sudo systemctl restart postgresql``` | ||
| * Use the manage.py script to set up the Postgres database: | ||
| - To create the user and database for gooey: `./manage.py sqlcreate | sudo -u postgres psql postgres ` | ||
| - Test your setup to ensure that gooey-server can access the database by running `psql -W -U gooey gooey` and supplying "gooey" as the password | ||
| * Create a .env file from `.env.example` | ||
| * Install the zbar library using your distro's package manager. | ||
| ## Run | ||
| Note: The gooey-server project is not currently set up to be run without support from Gooey. This software requires access to a Google Cloud instance as well as business data loaded in the database. If you are interested in running this software totally independently, reach out to [email protected] to communicate with our enterprise team. | ||
| You can start all required processes in one command with Honcho: | ||
| ```shell | ||
| $ poetry run honcho start | ||
| ``` | ||
| This will spin up the API server at `http://localhost:8080`. To view the autogenerated API documentation, navigate to `http://localhost:8080/docs` | ||
|
|
||
| If you installed the gooey-ui server, you can navigate to 'http://localhost:3000' to access the web application. | ||
|
|
||
| The processes that it starts are defined in [`Procfile`](Procfile). | ||
| Currently they are these: | ||
|
|
@@ -45,7 +81,7 @@ can do this by stopping and starting Honcho. | |
|
|
||
| ## To run any recipe | ||
|
|
||
| * Save `serviceAccountKey.json` to project root | ||
| * In order to run recipes, you will need API keys. To connect to Gooey's Google Cloud instance, you need a personal key, stored in `serviceAccountKey.json` in the project root. | ||
|
|
||
| ## To run vespa (used for vector search) | ||
|
|
||
|
|
@@ -68,24 +104,6 @@ docker run \ | |
| ./manage.py runscript setup_vespa_db | ||
| ``` | ||
|
|
||
| ## To connect to our GPU cluster | ||
|
|
||
| * Connect to k8s cluster - | ||
| ```bash | ||
| gcloud container clusters get-credentials cluster-5 --zone us-central1-a | ||
| ``` | ||
|
|
||
| * Port-forward the rabbitmq and redis services - | ||
| ```bash | ||
| kubectl port-forward rabbitmq-1-rabbitmq-0 15674:15672 5674:5672 & kubectl port-forward redis-ha-1-server-0 6374:6379 | ||
| ``` | ||
|
|
||
| * Add the following to `.env` file - | ||
| ``` | ||
| GPU_CELERY_BROKER_URL="amqp://rabbit:<password>@localhost:5674" | ||
| GPU_CELERY_RESULT_BACKEND="redis://:<password>@localhost:6374" | ||
| ``` | ||
|
|
||
| ### Install imagemagick | ||
|
|
||
| Needed for HEIC image support - https://docs.wand-py.org/en/0.5.7/guide/install.html | ||
|
|
@@ -102,60 +120,8 @@ Use black - https://pypi.org/project/black | |
| **Recommended**: Black IDE integration Guide: [Pycharm](https://black.readthedocs.io/en/stable/integrations/editors.html#pycharm-intellij-idea) | ||
|
|
||
|
|
||
| ## Running test Whatsapp bot | ||
|
|
||
| We use the following facebook app for testing - | ||
| ``` | ||
| gooey.ai (dev) | ||
| App ID: 228027632918921 | ||
| ``` | ||
|
|
||
| Create a [meta developer account](https://developers.facebook.com/docs/development/register/) & send admin your **facebook ID** to add you to the test app [here](https://developers.facebook.com/apps/228027632918921/roles/roles/?business_id=549319917267066) | ||
|
|
||
| 1. start ngrok | ||
|
|
||
| ``` | ||
| ngrok http 8080 | ||
| ``` | ||
|
|
||
| 2. set env var `FB_WEBHOOK_TOKEN = asdf1234` | ||
|
|
||
|
|
||
| 3. Open [WhatsApp Configuration](https://developers.facebook.com/apps/228027632918921/whatsapp-business/wa-settings/?business_id=549319917267066), set the Callback URL and Verify Token | ||
| <img width="500" alt="image" src="https://github.com/GooeyAI/gooey-server/assets/19492893/95bb3a87-ae4f-4f6b-a04e-583ee51b85de"> | ||
|
|
||
| 4. Open [WhatsApp API Setup](https://developers.facebook.com/apps/228027632918921/whatsapp-business/wa-dev-console/?business_id=549319917267066), send yourself a message from the test number. | ||
| <img width="500" alt="image" src="https://github.com/GooeyAI/gooey-server/assets/19492893/f9417723-77c0-4be5-9814-778662215d9c"> | ||
|
|
||
| 5. Copy the temporary access token there and set env var `WHATSAPP_ACCESS_TOKEN = XXXX` | ||
|
|
||
|
|
||
| **(Optional) Use the test script to send yourself messages** | ||
|
|
||
| ```bash | ||
| python manage.py runscript test_wa_msg_send --script-args 104696745926402 +918764022384 | ||
| ``` | ||
| Replace `+918764022384` with your number and `104696745926402` with the test number ID | ||
|
|
||
| ## Dangerous postgres commands | ||
|
|
||
| ### backup & restore postgres db | ||
|
|
||
| **on server** | ||
| ```bash | ||
| # select a running container | ||
| cid=$(docker ps | grep gooey-api-prod | cut -d " " -f 1 | head -1) | ||
| # give it a nice name | ||
| fname=gooey_db_$(date +"%Y-%m-%d_%I-%M-%S_%p").dump | ||
| # exec the script to create the fixture | ||
| docker exec -it $cid pg_dump --dbname $PGDATABASE --format c -f "$fname" | ||
| # copy the fixture outside container | ||
| docker cp $cid:/app/$fname . | ||
| # print the absolute path | ||
| echo $PWD/$fname | ||
| ``` | ||
|
|
||
| **on local** | ||
| ```bash | ||
| # reset the database | ||
| ./manage.py reset_db -c | ||
|
|
@@ -167,18 +133,6 @@ pg_restore --no-privileges --no-owner -d $PGDATABASE $fname | |
|
|
||
| ### create & load fixtures | ||
|
|
||
| **on server** | ||
| ```bash | ||
| # select a running container | ||
| cid=$(docker ps | grep gooey-api-prod | cut -d " " -f 1 | head -1) | ||
| # exec the script to create the fixture | ||
| docker exec -it $cid poetry run ./manage.py runscript create_fixture | ||
| # upload the fixture | ||
| docker exec -it $cid poetry run ./manage.py runscript upload_fixture | ||
| ``` | ||
|
|
||
| Save the new fixture url in `scripts/run-tests.sh` and run the tests - | ||
|
|
||
| ```bash | ||
| ./scripts/run-tests.sh | ||
| ``` | ||
|
|
@@ -200,7 +154,6 @@ To load the fixture on local db - | |
|
|
||
| ### copy one postgres db to another | ||
|
|
||
| **on server** | ||
| ```bash | ||
| ./manage.py reset_db | ||
| createdb -T template0 $PGDATABASE | ||
|
|
||